A B S T R A C T
It is frequently assumed that migrant sta-
tus constitutes a health risk because migra-
tion is inevitably associated with a period of
significant adjustment and stress. This paper
describes the role of acculturation in under-
standing the relationship between migration
and low birthweight (LBW). Psychosocial
and behavioural risk factors for LBW were
explored using semi-structured interviews
with 17 pregnant Southeast Asian women
who represented different levels of accultura-
tion. Findings suggested that acculturation
had negative consequences for immigrant
women. Higher levels of acculturation were
associated with dieting during pregnancy,
inadequate social support and stressful life
experiences.
A B R É G É
On présume souvent que l’immigration
pose un risque pour la santé en raison du
stress intense et de l’adaptation énorme
qu’elle implique. Cet article examine le rôle
de l’acculturation dans l’incidence des cas
d’insuffisance de poids à la naissance chez les
immigrants. À l’aide d’entrevues semi-
structurées auprès de 17 femmes enceintes
ayant immigré de l’Asie du Sud-Est et ayant
atteint divers niveaux d’acculturation, on y
explore les facteurs de risques psychosociaux
et comportementaux associés à l’insuffisance
de poids à la naissance. Les conclusions sug-
gèrent que l’acculturation nuit à la situation
des femmes immigrantes. On a observé un
lien entre les niveaux élevés d’acculturation et
l=alimentation durant la grossesse, le manque
de soutien social et les expÈriences stressantes
de la vie quotidienne.
SEPTEMBER – OCTOBER 2000
Negative Consequences of
Acculturation on Health Behaviour,
Social Support and Stress among
Pregnant Southeast Asian Immigrant
Women in Montreal: An Exploratory
Study
Ilene Hyman, PhD,1 Gilles Dussault, PhD2
Acculturation, the process of incorporat-
ing new values, attitudes and behaviours,
provides a conceptual bridge for under-
standing the relationship between migra-
tion and changes in health.1 Acculturation
has been implicated in immigrant adop-
tion of ‘bad’ North American health
habits, such as smoking,2-4 high fat diets,5-9
and substance abuse.10,11 In the area of peri-
natal health, several studies have reported
that more acculturated women experience
higher rates of low birthweight than less
acculturated counterparts.3,12,13 However,
explanations for these occurrences have
been inadequately researched.14,15
Low birthweight (LBW) can be due to a
short gestation (prematurity), an intra-uterine
growth retardation, or a combination of
both. The causes of term low birthweight are
multifactorial and include race/ethnicity,
maternal height and weight, general morbid-
ity, and health behaviours such as gestational
1. University Health Network Women’s Health
Program, Culture, Community and Health
Studies Program – Centre for Addiction and
Mental Health, Department of Public Health
Sciences, University of Toronto
2. Département d’administration de la santé,
Université de Montréal
Correspondence and reprint requests: Dr. Ilene
Hyman, University Health Network Women’s
Health Program, Toronto General Hospital, 657
University Avenue, Mulock/Larkin 2-008A,
Toronto, ON, M5G 2N2, Tel: 416-340-4800,
ext. 6783, Fax: 416-340-4185, E-mail:
ilene_hyman@camh.net
Ilene Hyman was the recipient of National Health
Research & Development Program Doctoral and
Post-Doctoral Training Fellowships from Health
Canada. This study received funding from the
Conseil québécois de la recherche sociale.
CANADIAN JOURNAL OF PUBLIC HEALTH 357
weight gain and caloric intake, cigarette
smoking and alcohol consumption. 16-21
Other researchers have examined the role of
social support22-25 and stress26-29 to explain
differences in pregnancy outcome.
Few studies have examined the conse-
quences of acculturation in terms of psy-
chosocial and behavioural risk factors that
impact on term LBW. The purpose of the
current study is to explore health behav-
iours (e.g., smoking, alcohol, diet), social
support and stress, in a group of pregnant
Southeast Asian immigrant women dis-
playing different levels of acculturation.
METHOD
The study population consisted of a
group of 17 pregnant Southeast Asian
women (Vietnamese, Cambodian, Laotian)
living in Montreal, Canada. Southeast
Asian women were selected because this
group experienced a significantly higher
rate of term LBW than native Quebecers
(4.0% vs. 3.4%), particularly among the
more acculturated members of this group.3
Subjects were identified from an interpre-
tation and orientation agency serving the
Southeast Asian community of Montreal
(SIARI), community health departments
(CLSCs), obstetricians, Southeast Asian
cultural and religious organizations,
Southeast Asian health professionals and
word of mouth. Informed consent was
obtained prior to the interview sessions.
Qualitative techniques were used to
identify the range of health and migration
ACCULTURATION AND HEALTH BEHAVIOUR
experiences encountered by the study pop-
ulation. Data collection, in the form of
semi-structured interviews, lasted between
one and two and a half hours. Vietnamese,
Cambodian and Laotian interpreters were
used in cases where the subject’s compre-
hension of English or French was poor.
During each interview, subjects were asked
about their migration history and resettle-
ment experiences in Canada as well as
about their current pregnancy.
Acculturation was assessed using two proxy
variables, length of stay30,31 and host coun-
try language fluency.32,33 Pregnancy-related
questions focussed on health behaviours,
social support and stress. All of the inter-
views were recorded with the permission of
the participants. The transcripts were sub-
sequently reviewed and coded into pre-
determined study themes: acculturation,
health behaviour, social support and stress.
RESULTS
Among the 17 women who were inter-
viewed, two distinct patterns of migration
and acculturation were observed. These
corresponded to the two major waves of
Southeast Asian refugee migration to
Canada. After the fall of Saigon to North
Vietnam in 1975, the first wave of
refugees, a predominantly urban, middle
class, and well-educated group, were relo-
cated to the United States, and about
9,000 resettled in Canada, primarily in
Montreal and Quebec City. During the
height of the ‘Vietnamese boat people cri-
sis’ which erupted in 1978, Canada pro-
vided haven to 60,000 second wave SEA
refugees. Compared to the first wave
group, second wave refugees came from a
wider socioeconomic spectrum, were less
well-educated and had less previous expo-
sure to the West.34
Subjects in our study population were
equally split between the two waves, corre-
sponding to our pre-defined acculturation
criteria. Compared to the more acculturat-
ed group who were members of the first
wave, the less acculturated group consisted
of more recent arrivals and individuals who
were not fluent in English and/or French.
There were also sharp contrasts between
the two groups with respect to socioeco-
nomic status. All spouses and some of the
358 REVUE CANADIENNE DE SANTÉ PUBLIQUE
subjects who arrived during the first wave
were employed, whereas most of the fami-
lies of the second wave (six out of the
eight) were receiving social assistance. The
age range of respondents (from 26 to 37)
and the high proportion of women who
were married were similar in both groups.
Health behaviours during pregnancy
Subjects were questioned about cigarette
smoking, alcohol consumption and diet.
None of the women in our study popula-
tion smoked, either before or during their
pregnancies, or used alcohol during preg-
nancy. Most of the women reported that
their diets in Canada were similar to the
diets they followed in Southeast Asia.
However, several subjects had observed
that the longer SEA women were in
Canada, the more preoccupied they were
with thinness, even during pregnancy. As
one 27-year-old Vietnamese subject pro-
claimed,
In my country, women ate well to have
a healthy baby, but here, many women
I know are on diets, even when they are
pregnant, they don’t eat a lot.
A 35-year-old Vietnamese subject com-
mented,
Women who have been here a long time,
from wealthy classes, are very obsessed
with their weight. Perhaps they ate less
during their pregnancy because they
wanted to maintain their figures.
Social support during pregnancy
During the in-depth interviews, each
subject was asked about her social support
networks including the availability and ade-
quacy of different types of support (infor-
mational, instrumental and emotional). In
the more acculturated group, we found that
most of the women included many family
members in their social support networks.
In spite of this, many subjects confided that
they had no one to share their problems or
worries with. Sometimes this appeared to
be due to an individual’s nature (“not inter-
ested in making friends”, “doesn’t like to
discuss worries with people”) but in other
cases, women explained that they didn’t
have enough time to see or talk with friends
and family.
In the less acculturated group, we found
surprising differences between the avail-
ability of family member support and per-
ceived levels of social support. For exam-
ple, although at least six subjects had no
close family in Montreal, they named
friends and neighbours upon whom they
called for assistance. Furthermore, the vast
majority expressed satisfaction with the
availability and adequacy of social support
they received.
Stress during pregnancy
Subjects were questioned about sources
of stress, symptoms and methods of cop-
ing. We discovered that the more accultur-
ated group of women were much more
likely than their counterparts to report that
they were experiencing a great deal of
stress. Financial pressures were the most
frequently cited source of stress, even
though this group was much better off eco-
nomically than the second group. As one
34-year-old Vietnamese subject explained,
If somebody lives here a long time, more
is necessary. When I came here I didn’t
know about fashion and hair. I lived
like I did in Vietnam. Now, when you
know about that, you want to buy, you
need a lot of money. People have more
problems and worries when they want
everything.
It also became evident during the inter-
views that the subjects shared the belief
that Southeast Asian women in Canada
were obligated to work and often contin-
ued to perform stressful and strenuous
work, even during pregnancy. As one 34-
year-old Laotian subject described her sis-
ter,
She works very hard for a manufacturer.
She is always standing. I encouraged her
to apply for maternity leave but she
worked until the end of her pregnancy.
She was afraid to ask, but also didn’t
want to go to the CLSC, even if she was
sick.
Another 34-year-old Laotian subject
summarized the experiences of immigrant
women in this way,
The women who are here longer are
more preoccupied with work. Their lives
are more stressful.
Not surprisingly, the second most fre-
quently mentioned source of stress was
inadequate social support. Women felt that
they were forced to assume more than their
VOLUME 91, NO. 5

share of responsibility for looking after
children, household chores and decision
making. They also worried about not hav-
ing enough assistance at home after the
baby was born, especially since they did
not feel as though there were many family
members or friends upon whom they
could rely for help.
Among the less acculturated group of
women, the inability to speak and compre-
hend French was the most frequently
reported source of stress. However, many
women felt that they were able to rely on
their partners, friends or SIARI to act as
interpreters. Surprisingly, financial con-
cerns were only expressed by two subjects.
More commonly, subjects shared the belief
that their incomes were adequate, if only
they could save a little money or economize
(“On arrive juste, juste.”). Pregnancy did
not appear to represent a great deal of stress
to this group as women expressed great
confidence in the Western medical system.
DISCUSSION
The findings of this study suggest that
acculturation had negative consequences
for immigrant women. Study respondents
reported that it was associated with
unhealthy behaviours and with different
types of acculturative stress believed to
impact on term LBW. Although the study
population did not smoke or consume alco-
hol during pregnancy, the respondents sug-
gested that more acculturated women were
more likely to be concerned with thinness
and to limit their weight gain. The impor-
tance of behavioural changes following
migration to explain differences in LBW
has been proposed by other authors.35-38
In our study women in the less accultur-
ated group expressed fewer psychosocial
concerns about social support and stress
than women in the more acculturated
group. This may have been because these
women felt secure within their established
social networks. Unlike the women in the
more acculturated group, all of the less
acculturated women lived in immigrant
neighbourhoods, in close proximity to
other members of their ethnic community.
Other studies have examined the role of
neighbourhood social environment and
LBW.39 For example, Kieffer et al.40 attrib-
SEPTEMBER – OCTOBER 2000
ACCULTURATION AND HEALTH BEHAVIOUR
uted the lower rates of LBW observed in
geographical areas of high ethnic homo-
geneity in Hawaii to the wider accessibility
of social support.
The finding that life may become more
stressful with increasing length of stay in a
host country is supported by other litera-
ture describing the acculturation experi-
ence of refugees.41 It has been documented
that refugees’ successful escape from disas-
ter results in an initial stage of relief.
However, as these refugees become less
dependent on agencies for social services,
or when their cash assistance runs out,
their level and rate of behavioural accultur-
ation may diminish or vacillate.42 There is
an extensive literature which supports the
observation that levels of psychological dis-
tress may be higher among more accultur-
ated immigrants who find that their
attempts to achieve social and economic
status fall short of their expectations and
aspirations due to discriminatory barriers
and practices related to employment and
advancement.30,43-46
Study limitations include the classifica-
tion of the study population into two
groups, assuming that the process of accul-
turation is linear and unidirectional. For
example, one may falsely conclude that an
immigrant who speaks, reads and writes
English is highly acculturated. 47
Dichotomizing acculturation may have
also led one to suspect that the negative
consequences described could be attributed
to socioeconomic status (SES) rather than
acculturation. Although it has been sug-
gested that these two constructs need to be
examined as separate variables which inde-
pendently and interactively impact on
health, 48 this could not be done in our
qualitative study.
The results of this study have implica-
tions for the organization of health and
social services and future research. It has
frequently been assumed that new immi-
grants and refugees constitute a higher risk
group than other more established mem-
bers of their communities but our findings
imply that it is not the recentness of immi-
grant status that contributes to this risk.
Rather, other factors related to the conse-
quences of acculturation need to be con-
sidered (in addition to conventional fac-
tors) in perinatal and other health risk
CANADIAN JOURNAL OF PUBLIC HEALTH 359
identification systems. Findings also sug-
gest the need for culturally appropriate ser-
vices that address both physical and emo-
tional needs (e.g., health education, stress
management) and for peer support and
other community-based help networks for
immigrant women. Furthermore, even
though Southeast Asian women may not
exhibit many behavioural risk factors at
present, this may be changing. Klatsky &
Armstrong49 reported that U.S.-born Asian
American women were more likely to
smoke than their foreign-born counter-
parts. Mitchell and Mackerra50 found that
only 57% of pregnant Vietnamese
American women continue to follow tradi-
tional food habits.
Most immigrant studies focus on the
early years of resettlement. However, the
later periods during which time there may
be delayed reactions to earlier traumas, and
distress over changes in life and status,
have received less attention. Future
research is needed to explore the long-term
consequences of acculturation, particularly
with respect to changes in health behav-
iours and reactions to stress.
ACKNOWLEDGEMENTS
The authors wish to thank Louise
Séguin, MD who acted as a consultant on
this study and colleagues who reviewed
previous versions of this manuscript.
REFERENCES
1. Berry JW. Psychology of acculturation. In:
Goldberger NR, Veroff JB (Eds.), The Culture
and Psychology Reader. New York: New York
University Press, 1995; 457-88.
2. Beiser M, Devins G, Dion R, et al. Immigration,
Acculturation and Health: Final Report to the
National Health Research and Development
Program. Ottawa: 1997; 6606-6614-NPHS.
3. Hyman I, Dussault G. The effect of acculturation
on low birthweight in immigrant women. Can J
Public Health 1996;87(3):158-62.
4. Marin G, Peres-Stable EJ, Marin BV. Cigarette
smoking among San Francisco Hispanics: The
role of acculturation and gender. Am J Public
Health 1989;79:196-98.
5. Pumariega AJ. Acculturation and eating attitudes
in adolescent girls: A comparative and correla-
tional study. J Am Acad Child Adolesc Psych
1986;25(2):276-79.
6. Hrboticky N, Krondl M. Acculturation to
Canadian foods by Chinese immigrant boys:
Changes in the perceived flavour, health value
and prestige of foods. Appetite 1984;5(2):117-26.
7. Cardoso MA, Hamada GS, de Souza JM, et al.
Dietary patterns in Japanese migrants to south-
ACCULTURATION AND HEALTH BEHAVIOUR
eastern Brazil and their descendants. J Epidemiol
1997;7(4):198-204.
8. Schaefer O, Timmermans JF, Eaton RD, et al.
General and nutritional health in two Eskimo
populations at different stages of acculturation.
Can J Public Health 1980;71(6):397-405.
9. Kouris-Blazos A, Wahlqvist ML, Trichopoulou
A, et al. Health and nutritional status of elderly
Greek migrants to Melbourne, Australia. Age
Ageing 1996;25(3):177-89.
10. Cheung YW. Making sense of ethnicity and drug
use: A review and suggestions for future research.
Social Pharmacology 1989;3:55-82.
11. Markides KS, Krause N, Mendes de Leon CF.
Acculturation and alcohol consumption among
Mexican Americans: A three-generational study.
Am J Public Health 1988;78:1178-81.
12. Scribner R, Dwyer JH. Acculturation and low
birthweight among Latinos in the Hispanic
HANES. Am J Public Health 1989;79:1263-67.
13. Yudkin PL, Harlap S. High birthweight in an
ethnic group of low socio-economic status. Br J
Obstet Gynaecol 1983;90:291-96.
14. Rumbaut RG, Weeks JR. Unraveling a public
health enigma: Why do immigrants experience
superior perinatal health outcomes? Research in
the Sociology of Health Care 1996;13B:337-91.
15. Scribner R. Editorial: Paradox as paradigm-The
health outcomes of Mexican-Americans. Am J
Public Health 1996;86(3):303-5.
16. Shiono PH, Klebanoff MA, Graubard B, et al.
Weight among women of different ethnic groups.
JAMA 1986;255:48-52.
17. Kramer MS. Intrauterine growth and gestational
duration determinants. Pediatrics 1987;80:502-
11.
18. Dunn HG. Social aspects of low birth weight.
CMAJ 1984;130:1131-40.
19. Alberman E. Low birthweight. In: Bracken MB
(Ed.), Perinatal Epidemiology. New York: Oxford
University Press, 1984.
20. Showstack JA, Budetti PP, Minkler D. Factors
associated with birthweight: An exploration of
the roles of prenatal care and length of gestation.
Am J Public Health 1984;74:1003-8.
21. Kleinman JC, Madans JH. The effects of mater-
nal smoking, physical stature, and educational
attainment on the incidence of low birth weight.
Am J Epidemiol 1985;121:843-55.
22. Zimmer-Gembeck MJ, Helfand M. Low birth
weight in a public prenatal care program:
Behavioral and psychosocial risk factors and psy-
chosocial intervention. Soc Sci Med
1996;43(2):187-97.
23. Norbeck JS, DeJoseph DF, Smith RT. A ran-
domized trial of an empirically-derived social
support intervention to prevent low birthweight
360 REVUE CANADIENNE DE SANTÉ PUBLIQUE
among African American women. Soc Sci Med
1997;43(6):947-54.
24. Newton RW, Hunt LP. Psychosocial stress in
pregnancy and its relation to low birth weight.
BMJ 1994;288:1191-94.
25. Boyce WT, Schaeffer C, Harrison HR, et al.
Social and cultural factors in pregnancy compli-
cations among Navajo women. Am J Epidemiol
1986;124:242-53.
26. Shiono PH, Raugh VA, Park M, et al. Ethnic dif-
ferences in birthweight: The role of lifestyle and
other factors. Am J Public Health
1997;87(5):787-93.
27. Michielutte R, Ernest JM, Moore ML, et al. A
comparison of risk assessment models for term
and preterm low birthweight. Prev Med
1992;21:98-109.
28. Reeb KG, Graham AV, Zyzanski SJ, et al.
Predicting low birthweight and complicated
labour in urban black women: A biopsychosocial
perspective. Soc Sci Med 1987;25:1321-27.
29. Oakley A, MacFarlane A, Chalmers I. Social
class, stress and reproduction. In: Rees AR,
Purcell H (Eds.), Disease and the Environment.
Chichester; Toronto: John Wiley, 1982; 11-49.
30. Hurh WM, Kim KC. Adaptation stages and
mental health Korean male immigrants in the
United States. International Migration Review
1990;24:456-79.
31. Cortes DE. Acculturation and its relevance to
mental health. In: Malgady RG, Rodriguez O
(Eds.), Theoretical and Conceptual Issues in
Hispanic Mental Health. Malabar, FL: Robert E.
Krieger Publishing, 1994; 53-68.
32. Hazuda HP, Stern MP, Haffner SM.
Acculturation and assimilation among Mexican
Americans: Scales and population-based data. Soc
Sci Q 1988;69(3):687-706.
33. Deyo RA, Diehl AK, Hazuda H, et al. A simple
language-based acculturation scale for Mexican-
Americans and applications for health care
research. Am J Public Health 1985;75:51-55.
34. Beiser M, Hyman I. Southeast Asian refugees in
Canada. In: Tousignant M, Al-Issa I (Eds.),
Ethnicity, Immigration and Psychopathology. New
York: Plenum Publishing, 1997.
35. Cabral H, Fried LE, Levenson S, et al. Foreign-
born and US-born Black women: Differences in
health behaviours and birth outcomes. Am J
Public Health 1990;80:70-72.
36. Zambrana RE, Scrimshaw SCM, Collins N.
Prenatal health behaviors and psychosocial risk
factors in pregnant women of Mexican origin:
The role of acculturation. Am J Public Health
1997;87(6):1022-26.
37. Reynoso TC, Felice ME, Shragg GP. Does accul-
turation affect outcome of Mexican-American
teenage pregnancy? J Adolesc Health
1993;14:257-61.
38. Ventura S, Taffel S. Childbearing characteristics
of U.S.- and foreign-born Hispanic mothers.
Public Health Rep 1985;100:647-52.
39. Roberts EM. Neighborhood social environments
and the distribution of low birthweight in
Chicago. Am J Public Health
1997;87(4):597-603.
40. Kieffer EC, Alexander GR, Lewis ND, et al.
Geographic patterns of low birthweight in
Hawaii. Soc Sci Med 1993;36(4):557-64.
41. Stein BN. The experience of being a refugee:
Insights from the research literature. In: Williams
CL, Westermeyer JW (Eds.), Refugee Mental
Health in Resettlement Countries. Washington:
Hemisphere Publishing Company, 1986.
42. Celano MP, Tyler FB. Behavioral acculturation
among Vietnamese refugees in the United States.
J Soc Psychol 1991;131(3):373-85.
43. Kaplan MS, Marks G. Adverse effects of accul-
turation: Psychological distress among Mexican
American young adults. Soc Sci Med
1990;31:1313-19.
44. Markides KS, Ray LA, Stroup-Benham CA, et al.
Acculturation and alcohol consumption in the
Mexican American population of the
Southwestern United States: Findings from
HHANES 1982-84. Am J Public Health
1990;80:42-46.
45. Richman JA, Gavira M, Flaherty JA. The process
of acculturation: Theoretical perspectives and an
empirical investigation in Peru. Soc Sci Med
1987;25(7):839-47.
46. Janes CR, Pawson IG. Migration and biocultural
adaptation: Samoans in California. Soc Sci Med
1986;22:821-33.
47. Sodowsky GR, Lai EW, Plake BS. Moderating
effects of sociocultural variables on acculturation
attitudes of Hispanics and Asian Americans.
J Counseling and Development 1991;70:194-204.
48. Rogler LH, Cortes DE, Malgady RG.
Acculturation and mental health states among
Hispanics: Convergence and new directions for
research. Am Psychol 1991;46:585-97.
49. Klatsky AL, Armstrong MA. Cardiovascular risk
factors among Asian Americans living in
Northern California. Am J Public Health
1991;81(11):1423-28.
50. Mitchell J, Mackerras D. The traditional
humoral food habits of pregnant Vietnamese-
Australian women and their effect on birth
weight. Aust N Z J Public Health
1995;19(6):629-33.
Received: July 21, 1999
Accepted: March 6, 2000
VOLUME 91, NO. 5
A B S T R A C T
Using the Theory of Planned Behaviour
(Ajzen, 1988) as a conceptual framework, 705
secondary school students were surveyed to
identify their intentions to use birth control
pills, condoms, and birth control pills in com-
bination with condoms. Hierarchical multiple
regression revealed that the theory explained
between 23.5% and 45.8% of the variance in
intentions. Variables external to the model such
as past use, age, and ethnicity exhibited some
independent effects. Attitudes were consistently
predictive of intentions to use condoms, pills,
and condoms in combination with pills for
both male and female students. However, there
were differences by gender in the degree to
which subjective norms and perceived behav-
ioural control predicted intentions. The find-
ings suggest that programs should focus on:
creation of positive attitudes regarding birth
control pills and condoms; targeting important
social influences, particularly regarding males’
use of condoms; and developing strategies to
increase students’ control over the use of con-
doms.
A B R É G É
La théorie du comportement axé sur un
objectif (Ajzen, 1988) a servi de cadre de travail
conceptuel pour mener un sondage auprès de
705 étudiants à l’école secondaire visant à con-
naître leurs intentions concernant l’utilisation
de la pilule anticonceptionnelle, du condom ou
des deux méthodes combinées. L’analyse de
variance avec régression hiérarchique a révélé
que la théorie expliquait entre 23,5 % et
45,8 % de la variance dans les intentions. Les
variables extérieures au modèle comme l’utilisa-
tion qu’ils en ont fait dans le passé, l’âge et le
groupe ethnique ont révélé des effets indépen-
dants. Les attitudes étaient prévisibles tant chez
les filles que chez les garçons quant à l’intention
d’utiliser soit le condom, la pilule anticoncep-
tionnelle ou bien une combinaison des deux.
Cependant, on a noté des différences selon le
sexe concernant la subjectivité des normes et la
perception de contrôle des comportements des
intentions prévues. Les conclusions indiquent
que les programmes devraient s’attarder à :
développer des attitudes positives au sujet de la
pilule anticonceptionnelle et du condom; viser
les influences sociales importantes, en particu-
lier en ce qui a trait à l’utilisation du condom
chez les hommes; et mettre au point des straté-
gies pour accroître la maîtrise des étudiants
pour ce qui est de l’utilisation du condom.
SEPTEMBER – OCTOBER 2000
Factors Predictive of Adolescents’
Intentions to Use Birth Control Pills,
Condoms, and Birth Control Pills in
Combination with Condoms
Dorothy M. Craig, RN, MScN,1 Karen E. Wade, RN, MScN,1,2
Kenneth R. Allison, PhD,3 Hyacinth M. Irving, MA,1,4
J. Ivan Williams, PhD,1,3,5 Carole M. Hlibka, RN, BN, BEd 2
Surveys in Canada and the United States
have shown evidence of high levels of sexu-
al activity among adolescents and young
adults, with increasing rates by age. 1-4
According to the Canada Youth and AIDS
study, 31% of males and 29% of females
in grade nine have had sexual intercourse
at least once.1 Furthermore, several studies
suggest that many are not using condoms
or are using them inconsistently.1,2,5-10 Also,
studies have noted that young people tend
to use condoms for contraception rather
than protection against STDs11-13 and to
not use condoms when taking oral contra-
ceptives. 13,14 This is of concern because
unprotected sexual activity poses serious
health threats for youth, including sexually
transmitted diseases, HIV infection, and
unplanned pregnancy.2,15,16
The Theory of Planned Behaviour
(TPB)17 was used as the conceptual frame-
work for the study reported here. This the-
ory, which is an extension of the Theory of
1. Faculty of Nursing, University of Toronto
2. Toronto Public Health
3. Department of Public Health Sciences, Faculty
of Medicine, University of Toronto
4. Department of Sociology, University of Toronto
5. Institute for Clinical Evaluative Sciences,
Sunnybrook and Womens’ College Health
Sciences Centre
Correspondence and reprint requests: Dorothy M.
Craig, Professor Emeritus, Faculty of Nursing,
University of Toronto, 50 St. George Street,
Toronto, ON, M5S 3H4, Tel: 416-626-3529,
E-mail: dorothy.craig@utoronto.ca.
This study was funded by The National Health
Research and Development Program of Health
Canada, Project #6606-5631-201. The study was a
project of the North York Community Health
Promotion Research Unit. Kenneth Allison is sup-
ported by a Career Scientist award from the Ontario
Ministry of Health.
CANADIAN JOURNAL OF PUBLIC HEALTH 361
Reasoned Action (TRA),18 suggests that a
specific behaviour is a function of an indi-
vidual’s intention to perform that behav-
iour. Intention is a function of attitude
towards the behaviour, subjective norm,
and perceived behavioural control. In turn,
attitude is a function of the belief that per-
forming the behaviour leads to certain out-
comes and the evaluation of these out-
comes. Subjective norm is a function of
the beliefs that important referents value
the performance of a particular behaviour
and the motivation to comply with the ref-
erents. Perceived behavioural control,
which refers to the ease or difficulty of per-
forming a behaviour, is influenced by
internal factors such as information, skills,
and emotions, as well as external factors
such as dependence on others. Variables
such as age, gender, and ethnicity are seen
as affecting intentions and behaviour indi-
rectly, through the model variables.
Studies based conceptually on the TRA
and the TPB have provided evidence of pos-
itive associations between beliefs and atti-
tudes about the outcomes of using birth
control pills, condoms, and various contra-
ceptives, and intention to use, and/or use of
birth control pills;19-21 and intention to use
and/or use of condoms.9,19,22-27 Similarly,
there is some evidence, in studies using the
TRA and the TPB, for the relationship
between subjective norms and intention to
use and/or use of condoms;22-31 and inten-
tion to use and/or use of oral contracep-
tives.19-21 Although studied to a lesser extent,
perceived behavioural control has been
found to be associated with intention to use
condoms.26,27 No previous studies using the
ADOLESCENTS’ INTENTIONS TO USE PILLS AND/OR CONDOMS

TPB have explored young adolescents’ TABLE I

intentions to use birth control pills, or birth Descriptive Statistics for Variables Used in OLS Regressions of Adolescents’
control pills in combination with condoms. Intentions to Use Birth Control Pills, Condoms, and Condoms in Combination
Using the TPB as a conceptual framework, with Birth Control Pills
the analysis reported here examines the
extent to which students’ global attitudes, Females (N=326) Males (N=379)
Mean SD Mean SD
subjective norms, and perceived behavioural Intentions:
control predict their intention to use birth I intend to use birth control pills daily. 5.046 1.865 5.129 1.524
I intend to insist that my partner always uses a condom
control pills, condoms, and birth control pills if I am not taking the birth control pill every day. 6.472 1.204 6.419 0.986
in combination with condoms. A subsequent I intend to insist that my partner always uses a condom
even if I am taking the birth control pill every day. 6.160 1.457 5.676 1.426
paper will examine the relationship between
intentions and actual behaviour. Birth Control Pills
Attitudes (sum of two 7-point scales) 10.444 2.722 10.653 2.546
Using birth control pills daily is useful.
METHODS Using birth control pills daily is good.

Subjective Norm (multiplicative composite

Sample of two 7-point scales) 20.295 13.968 19.068 15.087
Generally speaking, most people who are important
A single-stage cluster sampling design to me think I should use birth control pills daily.
was used. From a sampling frame of 17 Generally speaking, I want to do what most people
who are important to me want me to do.
secondary schools in an Ontario city, six
schools stratified by socioeconomic status Perceived Behavioural Control 5.767 1.433 4.344 1.538
I feel I have control over whether or not
and ethnicity were originally selected and I use birth control pills daily.
four of these schools consented to partici- Condom Use
pate. Following negotiations with individ- Attitudes (sum of two 7-point scales) 12.559 2.086 11.832 2.13
ual schools, eligible students between Insisting that my partner always uses a condom is useful.
Insisting that my partner always uses a condom is good.
grades nine and OAC (formerly called
grade 13) were invited to participate. The Subjective Norm (multiplicative composite
of two 7-point scales) 27.028 13.927 24.249 15.422
study received approval from the relevant Most people who are important to me think that
ethics committees. All participants provid- I should insist that my partner use a latex condom
each time we have sexual intercourse.
ed written informed consent. In addition, Generally speaking, I want to do what most people
parental consent was obtained for those who are important to me want me to do.
below age 18. Perceived Behavioural Control (sum of two
Of a total of 1,597 eligible students, 711 7-point scales) 10.434 2.591 10.779 2.319
I feel I have control over whether or not my
participated, for a response rate of 44.5%. partner always uses a condom.
Six questionnaires were excluded due to Insisting that my partner always uses a condom will be easy.
incomplete data, leaving 705 available for Past Behaviour†
the analysis. The participants consisted of Used birth control pills during the past 3 months. 0.067 0.251 0.069 0.253
Used condoms during the past 3 months. 0.150 0.358 0.187 0.391
53.8% males (mean age 15.8 years) and Combined use of condoms and birth control
46.2% females (mean age 15.8 years). pills during the past 3 months. 0.031 0.173 0.053 0.224
Almost a third (31.4%) of males and Demographic Measures
24.1% of females had engaged in sexual Age (in years) 15.803 1.651 15.789 1.552
intercourse at least once in the past. Length of Residence in Canada†
0 - 2 years 0.144 0.352 0.166 0.373
3 - 5 years 0.129 0.336 0.103 0.304
Measures used More than 5 years 0.291 0.455 0.298 0.458
Data were collected using the Adolescent NOTE: The wording in the table is the language for female respondents. For males, the wording
Sexual Health Questionnaire (male and reflected the behaviour of using a condom or insisting that a partner use birth control pills.
female versions). This questionnaire used Questions concerning intentions, attitudes, subjective norms, and perceived behavioural control
were rated on 7-point Likert scales, ranging from strongly disagree (1) to strongly agree (7).
global measures of attitude, subjective † The means reflect the mean percentage. For example, 6.7% of females used birth control pills
norm, and perceived behavioural control, during the past 3 months.
suggested by Ajzen.17 A number of experts Data collection and analysis tion of each of the model variables. Past
reviewed the instruments for relevance, The questionnaires were completed by experience and demographic characteris-
clarity, and potential item bias. Based on students in classrooms or auditoriums dur- tics were also examined. Analyses were car-
their input, and the results of a pilot test ing the spring term, 1997. Hierarchical ried out separately by sex. In each hier-
with 37 students, the instruments were multiple regression was used to examine archical analysis, intention was regressed
modified. The measures used in the analysis the ability of the TPB to predict inten- first on the TPB variables. The second
reported here are summarized in Table I. tions and to assess the relative contribu- regression model added past experience

362 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5

 ADOLESCENTS’ INTENTIONS TO USE PILLS AND/OR CONDOMS

TABLE II towards using condoms in combination

Intentions to Use Birth Control Pills Progressively Regressed on with pills was a significant positive predic-
Measures of the Theory of Planned Behaviour, Past Behaviour, tor of intentions for females. In the case of
and Demographic Characteristics males, all three TPB predictors were signif-
icant positive predictors of intention.
Model 1 Model 2 Model 3 Additional significant predictors of inten-
Independent Variables b beta b beta b beta
tion to use condoms in combination with
Females pills included: past use of condoms in
Attitudes 0.375*** 0.547 0.368*** 0.537 0.357*** 0.520
Subjective Norm 0.009 0.070 0.009 0.069 0.009 0.068 combination with pills (a negative predic-
Perceived Behavioural Control 0.094 0.072 0.096 0.074 0.092 0.070 tor for females); age (a negative predictor
Used Birth Control Pills During for both males and females); and years of
the Past 3 Months† 0.931** 0.125 0.960** 0.129 residence compared to born in Canada
Age -0.102* -0.090 (0-2 years was a negative predictor of
0 - 2 years residence in Canada‡ 0.101 0.019 intention for females, while 0-2 years, and
3 - 5 years residence in Canada‡ -0.612* -0.110
More than 5 years residence more than 5 years residence, were positive
in Canada‡ -0.142 -0.035 predictors of intention among males).
R2 0.368 0.383 0.405

Males Predictive value of the theory of planned

Attitudes 0.395*** 0.660
Subjective Norm 0.004
Perceived Behavioural Control 0.038
Partners Used Birth Control Pills
During the Past 3 Months†
Age
0 - 2 years residence in Canada‡
3 - 5 years residence in Canada‡
More than 5 years residence
in Canada‡
R2 0.458
* p<0.05 ** p<0.01 ***p<0.001
† Comparison group is students without prior experience with birth control pills during the past
3 months.
‡ Comparison group is native-born students.
and the third model added demographic
variables. Mean substitution was used to
replace missing values for the three out-
come measures.32,33
RESULTS
Intention to use birth control pills
The findings indicate some support for
the TPB although, for both males and
females, the only significant predictor of
intention from the theory was attitude
toward using pills (Table II). Students with
positive (global) attitudes were more likely
to intend to use pills. Neither subjective
norm nor perceived behavioural control
was predictive of intention for males or
females. Other significant predictors of
intention included past use of birth control
pills (females); age (a negative predictor for
females); and years of residence compared
to born in Canada (3-5 years was a nega-
tive predictor of females’ intentions and a
positive predictor of males’ intentions to
insist that their partners use pills).
0.395*** 0.660 0.398*** 0.666
0.042 0.004 0.042 0.004 0.039 behaviour
0.038 0.038 0.038 0.037 0.037 The predictive value of the TPB as a
whole was moderately high, explaining
-0.058 -0.010 -0.016 -0.003 36.8% (females) and 45.8% (males) of the
-0.012 -0.012 variance in intention to use pills (Table II);
0.030 0.007 32.3% (females) and 23.5% (males) of the
0.435* 0.087
variance in intention to use condoms
0.058 0.017 (Table III); and 35.3% (females) and
0.458 0.465
26.9% (males) of the variance in intention
to use condoms in combination with pills
(Table IV). Additional variables, included
in subsequent models of the hierarchical
Intention to use condoms multiple regression analyses, explained less
As was the case with pills, the findings additional variance in intention regarding
indicate some support for the TPB regard- either pill use or condom use than they did
ing intention to use condoms (Table III). for intention to use condoms in combina-
For females, attitude toward using con- tion with pills. These additional variables
doms was a positive predictor of intention. explained 6.1% of the total variance for
For males, all three components of the TPB females in intention to use condoms in
(attitude, subjective norm, and perceived combination with pills and 5.4% for males
behavioural control) were significant posi- (Table IV).
tive predictors of intention. Other signifi-
cant predictors of intention included past DISCUSSION
use (a negative predictor for females); age (a
negative predictor for both males and The predictive value of the TPB was
females); and years of residence compared moderately high, explaining between
to born in Canada (more than 5 years was a 23.5% and 45.8% of the variance in inten-
negative predictor of females’ intentions to tions. Previous studies of adolescents’
insist that their partners use condoms). intentions to use birth control pills or con-
doms have shown that the TRA or the
Intention to use condoms in combination TPB accounts for between 10.6% to 50%
with birth control pills of the variance in intention. 9,19,22,26 The
The findings indicate some support for TPB is said to be a “sufficient” model –
the TPB as a framework for predicting suggesting that variables external to the
intention to use condoms in combination model, such as past use, age, and ethnicity,
with pills (Table IV). As was the case with affect intentions indirectly through their
the two discrete intentions, attitude influence on attitudes, subjective norm or

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 363

ADOLESCENTS’ INTENTIONS TO USE PILLS AND/OR CONDOMS

perceived behavioural control. However, in TABLE III

the current study, these variables exhibited Intentions to Use Condoms Progressively Regressed on Measures of the Theory
some independent effects, although not of Planned Behaviour, Past Behaviour, and Demographic Characteristics
always in the expected direction. Thus, it is
important to include demographic and Model 1 Model 2 Model 3
Independent Variables b beta b beta b beta
social (contextual) factors when using atti-
tudinal models either to predict intention Females
Attitudes 0.324*** 0.562 0.318*** 0.551 0.303*** 0.525
and behaviour, or as the basis for programs Subjective Norm 0.000 0.001 0.000 0.004 -0.001 -0.016
and policies in the field. Perceived Behavioural Control 0.010 0.022 0.025 0.053 0.024 0.052
Results of this study indicate that atti- Partner Used Condoms During
tudes were consistently, positively predic- the Past 3 Months† -0.532*** -0.158 -0.498*** -0.148
tive of intentions to use condoms, pills, Age -0.074* -0.102
and condoms in combination with pills for 0 - 2 years residence in Canada‡ -0.321 -0.094
3 - 5 years residence in Canada‡ -0.167 -0.047
both male and female students. Previous More than 5 years residence
studies also report attitudes to be a signifi- in Canada‡ -0.273* -0.103
R2 0.323 0.347 0.371
cant predictor of high school age students’
intentions to use condoms9,19,22,26 or pills Males
Attitudes 0.185*** 0.400 0.182*** 0.392 0.177*** 0.382
(females only).19 However, results of this Subjective Norm 0.008** 0.119 0.008** 0.120 0.008** 0.121
study found differences by gender in the Perceived Behavioural Control 0.041* 0.097 0.045* 0.104 0.041* 0.097
degree to which subjective norms and per- Used Condoms During the
ceived behavioural control predicted inten- Past 3 Months† -0.134 -0.053 -0.061 -0.024
tions. Previous studies using the TRA or Age -0.074** -0.117
the TPB have produced mixed results 0 - 2 years residence in Canada‡ -0.116 -0.044
3 - 5 years residence in Canada‡ 0.138 0.042
regarding the predictive value of subjective More than 5 years residence
norms as related to intentions to use con- in Canada‡ 0.149 0.069
R2 0.235 0.238 0.261
doms9,19,22,26 or pills.19 Similarly, two pre-
vious studies examining the relationship * p<0.05 ** p<0.01 ***p<0.001
† Comparison group is students without prior experience with condoms during the past 3 months.
between perceived behavioural control and ‡ Comparison group is native-born students.
high school students’ intentions to use
condoms found mixed results.9,26 The findings suggest that programs need focus on developing strategies to increase
Ajzen’s explanation for these differences to move beyond merely providing infor- students’ control over the use of condoms
would be that the relative importance of mation to students and focus on the cre- and to make condoms easily and cheaply
attitudes, subjective norms, and perceived ation of positive attitudes regarding birth available to students.
behavioural control in predicting intention control pills and condoms. At a population The fact that older students were less
varies across behaviours and populations. level, health communication strategies likely to intend to enact these behaviours
However, this does not explain why subjec- (e.g., social marketing campaigns) could be and that female students who had used
tive norm and perceived behavioural control developed to illustrate positive attitudes condoms in the past said they were less
are more predictive of intentions for males regarding these behaviours. At a small likely to do so in the future are of concern.
than females. One possible reason for the group or individual level, practitioners, Although the study did not explore these
greater influence of subjective norm among such as nurses, physicians, and teachers, issues, it is noteworthy that, in this com-
males is that they may be more inclined to should engage in discussion with adoles- munity, most students are not involved in
discuss condom use with friends in that cents to determine their specific attitudes, educational programs focussing on healthy
condom use may be a “bragging point” including the beliefs underlying the atti- sexuality beyond grade 9. These findings
relating to sexual activity. The intentions of tudes. Discussion could ensue regarding suggest that accurate information regard-
female students, perhaps, are not as influ- how their attitudes and beliefs are support- ing behaviours to prevent pregnancy and
enced by others because their sexual activity ed or not supported by current informa- sexually transmitted diseases need to be
may be seen as more of a private and confi- tion and research. reinforced as adolescents get older.
dential matter. In order to examine this Because the current study highlighted
explanation empirically, further research is the influential nature of subjective norms ACKNOWLEDGEMENTS
needed on male and female differences in on males’ intentions to use condoms, pro-
the influence of subjective norms and per- grams should focus on determining the The authors acknowledge the contribu-
ceived behavioural control on intentions. salient social influences regarding condom tion of the following to this project: Board
For example, focus groups could be used to use and targeting these individuals/groups of Education staff and students, Cindy-Lee
explore the social context surrounding the as well as the adolescents themselves. The Dennis, Leanne Burton, Maureen Cava,
influences on adolescent intentions. study also suggests that programs should Carol Farkas, Angela Golabek, Helen

364 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5


TABLE IV
Intentions to Use Condoms in Combination with Birth Control Pills
Progressively Regressed on Measures of the Theory of Planned Behaviour,
Past Behaviour, and Demographic Characteristics
Model 1
Independent Variables b beta
Females
Attitudes 0.406*** 0.582
Subjective Norm 0.000
Perceived Behavioural Control 0.020
Combined Use of Condoms and
Birth Control Pills During the
Past 3 Months†
Age
0 - 2 years residence in Canada‡
3 - 5 years residence in Canada‡
More than 5 years residence
in Canada‡
R2 0.353
Males
Attitudes 0.287*** 0.429
Subjective Norm 0.008*
Perceived Behavioural Control 0.082** 0.134
Combined Use of Condoms and
Birth Control Pills During the
Past 3 Months†
Age
0 - 2 years residence in Canada‡
3 - 5 years residence in Canada‡
More than 5 years residence
in Canada‡
R2 0.269
* p<0.05 ** p<0.01 ***p<0.001
† Comparison group is students without prior experience with the combined use of condoms and
birth control pills during the past 3 months.
‡ Comparison group is native-born students.
Hutton, Lisa Lai, Jennifer Leeyus, Joan
Osbourne, and Nancy Weir.
REFERENCES
1. King A, Beazley R, Warren W, et al. Canada,
Youth, and AIDS Study. Social Program
Evaluation Group, Queen’s University,
Kingston, ON, 1988.
2. Health Canada. Sexual Risk Behaviours of
Canadians. Bureau of HIV/AIDS and STD Epi
Update Series, Ottawa, ON: LCDC, 1997.
3. U.S. Department of Health and Human Services.
Trends in sexual risk behaviour among high
school students - United States, 1991-1997.
Morbidity and Mortality Weekly Report
1998;47(36):749-52.
4. Thomas BH, DiCenso A, Griffith L. Adolescent
sexual behaviour: Results from an Ontario sam-
ple. Part 1: Adolescent sexual activity. Can J
Public Health 1998;89(2):90-93.
5. Borges SS, Hollett RG, Down S, et al. AIDS,
sexually transmitted diseases, and birth control:
Knowledge, attitudes and behaviours among
grade 11 students. PHERO 1992;3(17):293-96.
6. Ramsum DL, Marion SA, Mathias RG. Changes
in university students’ AIDS-related knowledge,
attitudes, and behaviours, 1988 and 1992. Can J
Public Health 1993;84(4):275-78.
SEPTEMBER – OCTOBER 2000
ADOLESCENTS’ INTENTIONS TO USE PILLS AND/OR CONDOMS
14. Maticka-Tyndale E. Sexual scripts and AIDS pre-
vention: Variations in adherence to safer sex
guidelines by heterosexual adolescents. J Sex Res
1991;28:45-66.
15. Frank J. 15 Years of AIDS in Canada. Canadian
Social Trends 1996;41:4-10. Ottawa, ON:
Statistics Canada.
Model 2 Model 3 16. Wadhera S, Millar W. Trends in births to
b beta b beta teenagers in Ontario, 1974-1994. Health Rep
1997;9(3):9-17.
17. Ajzen I. Attitudes, Personality, and Behaviour.
0.407*** 0.583 0.378*** 0.542 Bristol, Great Britain: Open University Press,
0.009 0.001 0.011 -0.001 -0.012 1988.
0.035 0.024 0.043 0.024 0.043
18. Ajzen I, Fishbein M. Understanding Attitudes and
Predicting Social Behaviour. Englewood Cliffs,
NJ: Prentice-Hall, 1980.
-0.899* -0.107 -0.930** -0.110 19. Adler NE, Kegles SM, Irwin CE, et al.
Adolescent contraceptive behaviour and assess-
-0.157*** -0.177 ment of decision-making. J Pediatr
-0.455* -0.110 1990;116:463-70.
-0.359 -0.082 20. Jaccard JJ, Davidson AR. Toward an understand-
ing of family planning behaviours: An initial
-0.229 -0.072 investigation. J Appl Soc Psychol 1972;2:228-35.
0.364 0.414 21. Werner PD, Middlestadt SE. Factors in the use
of oral contraceptives by young women. J Appl
Soc Psychol 1979;9(6):537-47.
0.285*** 0.426 0.284*** 0.424 22. Basen-Engquist K, Parcel GS. Attitudes, norms,
0.088 0.008 0.086 0.009* 0.102 and self-efficacy: A model of adolescents’ HIV-
0.087*** 0.142 0.095*** 0.154 related sexual risk behaviour. Health Educ Q
1992;19(2):263-77.
23. Ewald BM, Roberts CS. Contraceptive behaviour
-0.451 -0.071 -0.291 -0.046 in college-age males related to Fishbein model.
Adv Nurs Sci 1985;7(3):63-69.
-0.187*** -0.204 24. Fisher WA. Predicting contraceptive behaviour
0.364* 0.095 among university men: The role of emotions and
0.387 0.083 behavioural intentions. J Appl Soc Psychol
1984;14(2):104-23.
0.367** 0.118 25. Jemmott LS, Jemmott JB. Applying the theory of
0.273 0.323 reasoned action to AIDS risk behaviour:
Condom use among Black women. Nurs Res
1991;40(4):228-34.
26. Jemmott JB, Jemmott LS, Hacker CI. Predicting
intentions to use condoms among African
American adolescents: The theory of planned
behaviour as a model of HIV risk-associated
7. Svenson LW, Varnhagen CK. Knowledge, atti- behaviour. Ethn Dis 1992;2:371-80.
tudes and behaviours related to AIDS among 27. Lavoie M, Godin G. Correlates of intention to
first-year university students. Can J Public Health use condoms among auto mechanic students.
1990;81(2):139-40. Health Educ Res 1991;6(3):313-16.
8. Svenson LW, Varnhagen CK, Godin AM, 28. Rannie KE, Craig DM. Adolescent females’ atti-
Salmon TL. Rural high school students’ knowl- tudes, subjective norms, perceived behavioural
edge, attitudes and behaviours related to sexually control and intentions to use latex condoms.
Public Health Nurs 1993;14(1):51-57.
transmitted diseases. Can J Public Health
29. Reinecke J, Schmidt P, Ajzen I. Application of
1992;83(4):260-63.
the theory of planned behaviour to adolescents’
9. Richardson HRL, Beazley RP, Delaney ME, et
condom use: A panel study. J Appl Soc Psychol
al. Factors influencing condom use among stu-
1996;26(9):749-72.
dents attending high school in Nova Scotia. Can
30. Wulfert E, Wan K. Safer sex intentions and con-
J Human Sexuality 1997;6(3):185-96.
dom use viewed from a health belief, reasoned
10. Thomas BH, DiCenso A, Griffith L. Adolescent
action, and social cognitive perspective. J Sex Res
sexual behaviour: Results from an Ontario sam-
1995;32(4):299-311.
ple Part II: Adolescent use of protection. Can J
31. Reinecke J, Schmidt P, Ajzen I. Birth control ver-
Public Health 1998;89(2):94-97.
sus AIDS prevention: A hierarchical model of
11. Baffi CR, Schroeder KK, Redican KJ, et al.
condom use among young people. J Appl Soc
Factors influencing selected heterosexual male
Psychol 1997;27(9):743-59.
college students’ condom use. J Am Coll Health
32. Aseltine RH. Pathways linking parental divorce
1989;38:137-41.
with adolescent depression. J Health Soc Behav
12. Grimley DM, Riley GE, Bellis JM, et al.
1996;37(2):133-48.
Assessing the stages of change and decision-
33. Tabachnick BG, Fidell LS. Using Multivariate
making for contraceptive use for the prevention
Statistics. New York: Harper Collins, 1996.
of pregnancy, sexually transmitted diseases, and
acquired immunodeficiency syndrome. Health
Received: July 26, 1999
Educ Q 1993;20(4):455-70.
Accepted: March 6, 2000
13. Weisman CS, Plichta S, Nathanson CA, et al.
Consistency of condom use for disease preven-
tion among adolescent users of oral contracep-
tives. Fam Plann Perspect 1991;23(2):71-74.
CANADIAN JOURNAL OF PUBLIC HEALTH 365
 A B S T R A C T
Objective: To estimate the prevalence of resis-
tance of Mycobacterium tuberculosis to first-line
antituberculosis drugs in Canada.
Methods: M. tuberculosis isolates from one
third of all culture-positive tuberculosis (TB)
cases diagnosed between February 1, 1993 to
January 31, 1994 in Canada were collected
prospectively. Proportion of drug-resistant iso-
lates and the factors related to drug resistance
were measured.
Results: Of 458 study cases, 40 (8.7%) had
resistance to at least one first-line antituberculo-
sis drug, of which 5.9% had mono-resistance,
0.7% had multidrug-resistance(MDR-TB) –
i.e., resistance to at least isoniazid and rifampin
– and 2.2% had other patterns. The overall
prevalence of resistance among the foreign-born
cases was 10.6% with the highest level among
those who resided in Canada for less than four
years (15.5%).
Conclusions: Canada has a relatively low
prevalence of antituberculosis drug resistance
and a very low prevalence of MDR-TB. Some
new immigrants to Canada may be at higher
risk for drug resistance and their initial treat-
ment needs to be tailored accordingly.
A B R É G É
Objectif : Évaluer la prévalence de la résis-
tance du mycobacterium tuberculosis contre le
médicament antituberculeux de première ligne
au Canada.
Méthodes : Le mycobacterium tuberculosis
repère un tiers de tous les cas de cultures tuber-
culosis qui ont fait l’objet d’un diagnostic positif
entre le 1er février 1993 et le 31 janvier 1994 au
Canada; ces données ont été collectées de
manière prospective. On a mesuré la proportion
de repère des médicaments résistants et les fac-
teurs reliés à la résistance du médicament.
Résultats : Sur 458 cas étudiés, 40 (8,7 %) se
sont avérés résistants à au moins un médicament
antituberculeux de première ligne dont 5,9 % se
sont avérés résistants à un seul médicament et
0,7 % ont une résistance à plusieurs médica-
ments (MDR-TB) – c.-à-d. résistants à au
moins deux médicaments : l’isoniazid et le
rifampin – et 2,2 % avaient d’autres caractéris-
tiques. La prévalence globale de la résistance
parmi les cas venus de l’étranger était de 10,6 %
dont le niveau le plus élevé a résidé au Canada
pendant moins de quatre ans (15,5 %).
Conclusions : Au Canada, la prévalence de la
résistance des médicaments antituberculeux est
relativement faible. Certains nouveaux arrivants
peuvent présenter un risque plus élevé de résis-
tance aux médicaments et leur traitement initial
doit être personnalisé en conséquence.
366 REVUE CANADIENNE DE SANTÉ PUBLIQUE
Drug Resistance Study of
Mycobacterium tuberculosis
in Canada, February 1, 1993
to January 31, 1994
Ezzat Farzad, MD, MSc, FRCPC,1 Donna Holton, MD, FRCPC,2
Richard Long, BSc, MD, FRCPC, FCCP,3 Mark FitzGerald, MD, FCCP, FRCPI,
FRCPC,4 Adalbert Laszlo, PhD,5 Howard Njoo, MD, MHSc, FRCPC,6
Anne Fanning, MD, FRCPC,7 Earl Hershfield, MD, FRCPC,8
Vernon Hoeppner, MD, FRCPC,9 Edward Allen, MB, FRCPC, FCCP10
Reports from many countries around the
world have documented an increasing
prevalence of drug-resistant tuberculosis.1-5
The last Canadian national drug resistance
survey6 conducted in 1975 reported a drug
resistance prevalence of 6.3% among new
cases. More recently, a drug resistance sur-
vey in Western Canada7 reported a preva-
lence of 6.9%. Prevalence of drug resis-
tance was reported to be higher among
foreign-born patients6-9 especially if they
developed tuberculosis (TB) within the
first few years of their arrival in Canada.7-9
Drug-resistant tuberculosis – i.e., resis-
tance to one or more of the first-line anti-
1. Medical specialist, Division of Blood Borne
Pathogens, LCDC. Formerly, TB epidemiolo-
gist, Division of Tuberculosis Prevention and
Control, LCDC, Ottawa
2. Regina Health District, Regina, Saskatchewan
3. Associate professor, Department of Medicine,
University of Alberta, Edmonton
4. Director, TB Services to Aboriginals, BC Centre
for Disease Control Society, Vancouver
5. A/Head National Reference Centre for
Tuberculosis, LCDC. Consultant TB
Bacteriology-International Union Against
Tuberculosis and Lung Disease (IUATLD),
Ottawa
6. Director, Division of Tuberculosis Prevention
and Control, LCDC
7. Professor, Department of Medicine, University
of Alberta
8. Professor, Department of Medicine, University
of Manitoba, Winnipeg
9. Professor, Department of Medicine, University
of Saskatchewan, Saskatoon
10. Clinical professor of Medicine Emeritus,
University of British Columbia. Formerly,
Director, Division of Tuberculosis Control,
Ministry of Health, British Columbia,
Vancouver
Correspondence: Dr. Ezzat Farzad, 1900 Montereau
Avenue, Orleans, ON, K1C 5X2, Tel: 613-837-
0199, E-mail: ezzat_farzad@hc-sc.gc.ca
tuberculosis drugs including isoniazid
(INH), rifampin (RMP), streptomycin
(SM), pyrazinamide (PZA) and ethambu-
tol (EMB), and specifically multidrug-
resistant tuberculosis (MDR-TB), i.e.,
resistance to at least isoniazid and rifampin
– poses new challenges for both clinical
management and TB control programs.10-13
In Canada from 1991 to 1996, about
2,000 active cases of TB (7 per 100,000
population) were diagnosed and reported
annually14 with more than 100 deaths per
year attributed to TB.15
The epidemiology of TB in Canada has
changed over the last 15 years. The propor-
tion of foreign-born cases has increased from
35% in 1980 to 63% in 1996 while the pro-
portion of cases occurring in the non-
Aboriginal Canadian-born has decreased
from 50% to 20% and the proportion of
cases among Aboriginal Canadians has
remained stable over this period.14
To provide a more recent estimate of the
prevalence of antituberculosis drug resis-
tance in Canada and to ascertain the
potential risk factors related to drug resis-
tance, a collaborative study was initiated by
the interested investigators, the Laboratory
Centre for Disease Control (LCDC), the
provincial and territorial TB Control units,
and the provincial and territorial TB labo-
ratories.
METHODS AND MATERIALS
Study design
A descriptive study was designed to
determine the prevalence of antituberculo-
VOLUME 91, NO. 5
 ANTITUBERCULOSIS DRUG RESISTANCE IN CANADA

sis drug resistance among TB cases in TABLE I

Canada. Data were collected prospectively Distribution of Sex, Age, Disease Activity Status, and Ethnic Origin Among
over a one-year period. Study Cases and Drug-resistant Cases

Sample size Number of TB Number of Prevalence of

Cases (%) Resistant Cases (%) Resistance (95% CI)
The required sample size was estimated
based on the prevalence of drug resistance Total 458 (100) 40 (100) 8.7 (6.4, 11.8)
Sex
reported from previous drug resistance sur- Male 268 (58.5) 24 (60) 8.9 (5.9, 12.8)
veys and the number of culture-positive Female 190 (41.5) 16 (40) 8.4 (5.1, 13.0)
Age (years)
TB cases reported in Canada in 1990 16 <15 14 (3.1) 0 (0.0) 0.0
(1,274), the most recent national statistic 15-45 219 (47.8) 24 (60) 10.9 (7.3, 15.6)
>45 225 (49.1) 16 (40) 7.1 (4.3, 11.1)
available when the study was initiated. The Activity Status
Epi-Info 6.03 statistical package was used New 401 (87.6) 36 (90) 9.0 (6.5, 12.1)
Relapsed 50 (10.9) 4 (10) 8.0 (2.6, 18.2)
to calculate the sample size. It was deter- Unknown 7 (1.5) 0
mined that a sample size of 455 culture- Ethnic Origin
Foreign-born 255 (55.7) 27 (67.5) 10.6 (7.2, 14.8)
positive cases was needed to detect a drug Canadian-born 194 (42.3) 12 (30) 6.2 (3.4, 10.3)
resistance level of 8% ± 2%, with a 95% Aboriginal 86 (18.8) 2 (5) 2.3 (0.3, 7.5)
Non-Aboriginal 108 (23.5) 10 (25) 9.3 (4.8, 15.9)
confidence interval (CI). Unknown 9 (2) 1 (2.5) 11.1 (0.5, 43.9)

Sample selection TABLE II

Since the number of reported TB cases Pattern of Drug Resistance in Canada
varied extensively among different
Number of Prevalence
provinces and territories, using a systematic TB Cases of Resistance
sampling, one third of all culture-positive (95% CI)
Total Number of Isolates 458
cases from Quebec, Ontario, Alberta and Sensitive to All 5 Drugs 418
British Columbia; one half of cases from Any Resistance 40 8.7 (6.4, 11.8)
SM 25 5.5 (3.6, 8.1)
Saskatchewan and Manitoba; and all cases INH 20 4.4 (2.8, 6.5)
from Newfoundland, Prince Edward PZA 7 1.5 (0.7, 3.3)
RMP 3 0.7 (0.2, 2.1)
Island, New Brunswick, the Yukon, and EMB 1 0.2 (0.0, 1.4)
the Northwest Territories were selected. Mono-resistance 27 5.9 (4.0, 8.3)
SM 14 3.1 (1.7, 4.9)
This meant that only one half or one third INH 7 1.5 (0.7, 3.0)
of all cases from provinces with higher case PZA 6 1.3 (0.5, 2.7)
Multidrug Resistance* 3 0.7 (0.16, 1.77)
loads were selected, whereas all cases from Other Patterns
provinces with lower case loads were select- INH + SM 10 2.2 (1.11, 3.86)
ed. Because this sampling method may not * Resistance to at least INH & RMP
have provided a representative sample
within the Canadian population, a random Comparative studies17 have shown the reli- origin (foreign-born, non-Aboriginal
sample of cases from the original sample ability of the radiometric modified propor- Canadian-born, and Aboriginal Canadian),
was selected so that one third of all culture- tion method when compared to the con- TB sites (based on ICD-9 codes), and dis-
positive TB cases from each of the provin- ventional proportion method. ease activity status (new disease, relapsed
cial/territorial laboratories, diagnosed An isolate was defined as: 1) drug-resistant disease). The completed questionnaires
between February 1, 1993 and January 31, if it was resistant to one or more of the five were then forwarded to LCDC.
1994, were included in the analyses. drugs tested, or 2) drug-susceptible if it was The definitions of new and relapsed dis-
The enrolled cases were assigned a sensitive to all drugs. Drug resistance was ease were consistent with the definitions
unique study number and their isolates categorized as: 1) mono-resistance if the used in the Canadian Tuberculosis
were forwarded to the National Reference isolate was resistant to only one of the first- Reporting System. New disease was
Centre for Tuberculosis (NRCT) at line antituberculosis drugs, 2) multidrug- defined as a case of TB with no history or
LCDC, where they were tested for resis- resistance (MDR) if the isolate was resis- documentation of previously active tuber-
tance to the five first-line antituberculosis tant to at least INH and RMP, and culosis and relapsed disease was defined as
drugs: isoniazid (INH), rifampin (RMP), 3) other patterns. a case of TB with a history or documenta-
streptomycin (SM), pyrazinamide (PZA), For each case in the study, a question- tion of previously active tuberculosis that
and ethambutol (EMB). The radiometric naire was completed by the provincial/ter- became inactive. Foreign-born individuals
modified proportion method of drug sus- ritorial TB Control units. The same study were categorized, according to the WHO,
ceptibility testing of M. tuberculosis by the number was assigned to the questionnaire. as having been born in one of six geo-
BACTEC 460 system was employed. Data were collected on gender, age, ethnic graphical regions (the Americas, Europe,

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 367

ANTITUBERCULOSIS DRUG RESISTANCE IN CANADA
TABLE III
Prevalence of Drug Resistance by Province/Territory
Province/Territory Number of TB
Cases (%)
Atlantic Provinces 24 (5.2)
Quebec 107 (23.4)
Ontario 152 (33.2)
Manitoba 31 (6.8)
Saskatchewan 20 (4.4)
Alberta 37 (8.1)
British Columbia 79 (17.2)
Yukon & Northwest Territories 8 (1.7)
Total 458 (100)
South East Asia, Western Pacific, Eastern
Mediterranean, and Africa).18
Case inclusion criteria
TB cases were entered into the study if:
1) the NRCT received a viable M. tubercu-
losis specimen for the case, 2) the sampling
method had been followed, 3) the culture
date was consistent with study time frame,
February 1, 1993 to January 31, 1994, and
4) the epidemiological questionnaire was
available.
Statistical analysis
The Epi-info version 6.04 statistical
package was used to perform descriptive
and univariate statistical analyses to deter-
mine the prevalence of drug resistance in
the study population (i.e., culture-positive
cases) and to compare frequency of occur-
rence of potential risk factors among resis-
tant TB cases. For simple proportions,
confidence intervals were obtained by
Fisher exact method. For comparison of
categorical data, χ2 test or Fisher exact test
(if the expected cell values were less than
five) were calculated to determine the sta-
tistical significance. For comparison of
continuous data, the student t-test was
used. For all analyses, a p-value of < 0.05
was considered to be significant.
RESULTS
During the study period, a total of 536
culture-positive TB cases were enrolled in
the study. This was slightly in excess of all
culture-positive cases that were reported to
Statistics Canada in 1993.19 Upon arrival
to NRCT, 78 (14.6%) isolates were
excluded from further analyses for the fol-
lowing reasons: 70 isolates did not contain
viable M. tuberculosis, 6 cases were not
368 REVUE CANADIENNE DE SANTÉ PUBLIQUE
Number of Prevalence of
Resistant Cases (%) Resistance (95% CI)
2 (5) 8.3 (1.4, 24.9)
13 (32.5) 12.1 (6.9, 19.4)
14 (35) 9.2 (5.3, 14.6)
0 (0) 0.0
1 (2.5) 5.0 (0.2, 22.2)
2 (5) 5.4 (0.9, 16.7)
8 (20) 10.1 (4.8, 18.3)
0 (0) 0.0
40 (100) 8.7 (6.4, 11.6)
selected according to the study sampling
frame, and 2 cases were not selected within
the study time period. These patients were
similar to the 458 patients included in the
study except they were more likely to be
foreign-born (p = 0.006). However, the
proportion of foreign-born cases as well as
the distribution of sex, age, ethnic origin,
and disease activity status among the final
study population were similar to those of
the total TB cases that were reported to
Statistics Canada in 1993.19
Overall, resistance to one or more anti-
tuberculosis drugs was reported in 40 (8.7%)
of the total 458 cases (Table I). Prevalence
of resistance was similar among males and
females. No drug resistance occurred in
children younger than 15 years of age. The
rate of resistance was 10.9% among
patients aged 15-45 years and 7.1% for
those over 45 years, however, the difference
was not statistically significant (p = 0.16).
The overall prevalence of drug resistance
was 9% among those with new disease and
8% among relapsed cases. Further analyses
of data showed that among foreign-born
cases, prevalence of drug resistance was
11% for the new cases and 8.3% for the
relapsed cases. On the other hand among
non-Aboriginal Canadian-born, these pro-
portions were 8.6 and 14.3% respectively.
However, these differences were not statis-
tically significant. Both Aboriginal resistant
cases were new TB cases.
Of the drug-resistant cases, 27 (67.5%)
were foreign-born, 10 (25%) were non-
Aboriginal Canadian-born, and 2 (5%)
were Aboriginal Canadians. The preva-
lence of resistance was 10.6% among
foreign-born cases versus 6.2% for the
Canadian-born cases (p = 0.1). Further
analysis showed that the prevalence of
resistance among both foreign-born and
non-Aboriginal Canadian-born cases were
significantly higher than that of Aboriginal
cases (p = 0.02 and p = 0.05).
There were 297 pulmonary TB cases of
which 30 exhibited resistance (10.1%),
123 with extra pulmonary TB of which 9
were resistant (7.3%), and 25 with both
pulmonary and extra pulmonary TB. The
difference between proportion of drug
resistance among pulmonary and extra pul-
monary cases did not achieve statistical sig-
nificance.
The patterns of drug resistance among
study population are shown in Table II.
The highest level of resistance for individ-
ual drugs was to streptomycin (5.5%), fol-
lowed by isoniazid (4.4%) and pyrazi-
namide (1.5%). Of the drug-resistant
cases, 27 (5.9%) had mono-resistance,
3 (0.7%) had MDR-TB. The resistance to
SM was significantly higher among
foreign-born cases (7.1%) than among
Aboriginal and non-Aboriginal Canadian
cases, 0.0% and 5.6% respectively
(p = 0.04)(not shown in Table II). The
resistance to INH in foreign-born cases
was 5.1% versus 2.3% for Aboriginal cases
and 3.7% for non-Aboriginal Canadian-
born cases. These differences were not sta-
tistically significant.
The prevalence of drug resistance varied
widely between provinces/territories (Table
III). Quebec, Ontario and British
Columbia had the highest prevalence of
drug resistance while Manitoba, the Yukon
and Northwest Territories had the lowest.
The overall drug resistance in the provinces
of Quebec, Ontario and British Columbia
(i.e., provinces with the highest reporting
number of TB) was 10.3% which was sig-
nificantly higher than the drug resistance
level of 4.2% for the rest of Canada
(p = 0.04). A high proportion of resistant
cases in Quebec, Ontario and British
Columbia occurred among foreign-born
groups – 54%, 86%, and 63% respective-
ly; and these proportions were not signifi-
cantly different.
The prevalence of drug resistance among
foreign-born cases varied between geo-
graphical regions (Table IV). More than
85% of the foreign-born drug-resistant
cases were born in countries located in the
Western Pacific, African, and South East
Asian regions.
VOLUME 91, NO. 5
 ANTITUBERCULOSIS DRUG RESISTANCE IN CANADA

The five countries with the highest pro- TABLE IV

portion of resistance are listed in Table V. Prevalence of Drug Resistance Among Foreign-born TB Cases
The total resistant cases from these coun- by WHO Geographical Region
tries accounts for two thirds of foreign-born
cases. Their overall drug resistance level of Number of Number of Prevalence of
TB Cases Resistant Cases Resistance (95% CI)
15.1% is significantly higher than the
Canadian national level of 8.7% (p = 0.04). African Region 27 4 14.8 (4.9, 31.9)
Western Pacific Region 125 15 12.0 (7.1, 18.6)
The mean duration of residence in South East Asian Region 37 4 10.8 (3.5, 24.0)
Canada for foreign-born individuals with Eastern Mediterranean Region 10 1 10.0 (0.5, 40.3)
American Region* 27 2 7.4 (1.3, 22.4)
drug resistance was 6.3 years versus 10.6 European Region 29 1 3.4 (0.2, 15.8)
years for those without drug resistance. Total 255 27 10.6 (7.2, 14.8)
This difference did not reach the signifi- * Does not include Canada
cance level (p = 0.1). Table VI shows the
trend of drug resistance among foreign- TABLE V
born cases in relation to the duration of Antituberculosis Drug Resistance Among Foreign-born Cases from the
their residence in Canada. The prevalence Five Countries with Highest Number of Drug Resistance
of drug resistance among those who lived
Number of Number of Prevalence of
in Canada less than four years (15.5%) was TB Cases Resistant Cases Resistance (95% CI)
significantly higher than the average
Vietnam 30 7 23.3 (10.8, 40.8)
national level of 8.7% (p = 0.04). Somalia 14 3 21.4 (5.8, 48)
China 31 4 12.9 (4.2, 28.3)
India 21 2 9.5 (1.6, 28.1)
DISCUSSION Hong Kong 23 2 8.7 (1.5, 25.9)
Total 119 18 15.1 (9.5, 22.4)
In this study, resistance to antituberculo-
sis drugs occurred in 8.7% of the 458 TABLE VI
culture-positive TB cases. In the last The Relationship Between Antituberculosis Drug Resistance and Duration of
national survey, Eidus6 reported a resistance Residence in Canada Among Foreign-born Cases*
rate of 6.3% for new active cases – a rate
Duration of Number of Number of Prevalence of
lower than the 9% found in the present Residence (years) TB Cases Resistant Cases Resistance (95% CI)
study (p = 0.07). Other Canadian studies
<4 97 15 15.5 (9.3, 23.7)
have reported different rates of resistance 4-10 63 6 9.5 (3.9, 18.7)
from different parts of the country, ranging >10 83 6 7.2 (3.0, 14.4)
from 6.9% in Western Canada7 to 16.2% * The cases for whom the duration of residence in Canada were not reported were excluded
in Montreal,20 which indicate regional dif-
ferences. The prevalence of drug resistance lence of resistance to other first-line anti- respectively, whereas the corresponding
in this study is lower than those reported to tuberculosis drugs including pyrazinamide, figures for non-Aboriginal Canadian-born
WHO 1 by many countries including rifampin, and ethambutol were all low, cases were 8.6 and 14.3%. It is probable
Australia (9.5%), USA (12.9%), Spain justifying their use in the initial treatment that, due to lack of documentation of prior
(12.9%), Netherlands (14%) and Portugal of TB. The sample size of the present antituberculosis drug use and because of
(16.5%). study was not sufficient for a precise calcu- recall bias, some relapsed foreign-born
MDR-TB was demonstrated in only lation of resistance for individual antitu- cases might have been misclassified as new
0.7% of TB cases. A recent survey of berculosis drugs. cases.
MDR-TB from Alberta and British There was no case of drug resistance 55.6% of TB cases and 67.5% of resis-
Columbia reported the same rate (0.7%) among children 14 years of age and tant cases were among the foreign-born,
with all cases having MDR-TB on the ini- younger in our data (Table I). Children are who constitute only 17% of the total pop-
tial isolate being foreign-born.21 Based on less likely to produce sputum or to be ulation of Canada. Prevalence of drug
the WHO report,1 this is similar to the culture-positive and therefore are under- resistance was 10.6% among foreign-born
reported figure from Australia and New represented in the study population.19 and 6.2% among Canadian-born cases, a
Zealand (0.7%) but lower than those Surprisingly, our data did not show that difference that did not achieve statistical
reported from USA (2%) and among many resistance was more frequent among significance. Whereas a survey from
European countries including England relapsed cases than among new cases. Montreal 20 showed no difference in the
(1.9%), Spain (2%) and Portugal (3.7%). However, stratified analyses of the data prevalence of drug resistance between the
Resistance was highest to streptomycin showed that among foreign-born cases, the foreign-born and Canadian-born (16.3%
and isoniazid, probably because these drugs proportion of resistance was 11% and and 15.9%), a Western Canada survey 7
have been in use the longest. The preva- 8.3% among new and relapsed cases demonstrated a significantly higher preva-

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 369

ANTITUBERCULOSIS DRUG RESISTANCE IN CANADA
lence of resistance in foreign-born cases as
compared to Canadian-born cases (11%
and 3.1% respectively).
Further stratification of data showed a
significantly lower prevalence of resistance
among Aboriginal cases (2.3%) as com-
pared to other groups. This low prevalence
of resistance among Aboriginal cases was
also shown in the Western Canada study.7
The likelihood of drug resistance among
foreign-born cases decreased as the dura-
tion of residence in Canada increased
(Table VI), with more than 50% of drug-
resistant cases having resided in Canada for
less than four years. These findings are
consistent with those of previous drug
resistance surveys in Canada.6-9
The recent changes in immigration pat-
terns might have contributed to the high
proportions of TB and resistant TB cases
among new immigrants. The number of
immigrants from areas of the world where
the prevalence of TB and drug resistance
are high, has increased over the past 20
years. As a result, a large pool of infected
individuals or previously inadequately
treated TB cases with resistant strains, may
have immigrated to Canada with later pro-
gression to active resistant TB.
More than 85% of the foreign-born
resistant cases were born in countries locat-
ed in the Western Pacific, African, and
South East Asian regions. Five countries,
including China, Vietnam, Somalia, Hong
Kong, and India, were stated as the birth-
place for 66.6% of the foreign-born drug-
resistant cases. The findings in this study
are consistent with other drug-resistant
surveys7,9 and reflect the rate of resistance
prevailing in those geographical areas.
CONCLUSION
This study demonstrated that Canada
has a relatively low prevalence of antituber-
culosis drug resistance, including MDR-
TB. Foreign-born TB cases from some
regions of the world may have a higher risk
of developing drug-resistant TB, especially
if they develop TB within the first three
years of arrival in Canada. It is recom-
mended that initial therapy of foreign-born
TB cases, especially those from countries
with high prevalence of TB, include at
least four first-line drugs until susceptibili-
370 REVUE CANADIENNE DE SANTÉ PUBLIQUE
ty testing allows a more individualized reg-
imen. Furthermore, adequate treatment of
TB cases must be assured to prevent devel-
opment of drug resistance. Since treatment
of TB cases with antituberculosis drugs is
the mainstay of TB control, drug resistance
studies should be repeated at regular inter-
vals to assess the trends and pattern of drug
resistance. This information may be used
as an indicator of the effectiveness of TB
control programs in Canada and as a guide
to the clinical management of TB cases.
Although the study sample size was suf-
ficient to produce a meaningful estimate of
the overall prevalence of antituberculosis
drug resistance in Canada, further studies
with larger sample size are required to
allow for precise subgroup analyses.
ACKNOWLEDGEMENTS
We are indebted to the following persons
from the Provincial/Territorial Tuberculosis
Control Units, the Provincial/Territorial
Tuberculosis Laboratories, and the
Provincial Epidemiologists for their generous
contribution to provision of epidemiological
and laboratory data: Dr. Robert Cowie, Dr.
James Talbot, Sylvia Chomyc, Dr. Kevin
Elwood, Dr. William Black, Dr. Amin
Kabani, Joyce Wolfe, Dr. Christofer Balram,
Dr. John MacKay, Dr. Glenna Hardy,
Mona Crowley, Joan MacDonald, Dr. Paul
Fardy, Dr. Faith Stratton, Sandra March,
Dr. Ian Gilchrist, Wanda White, Dr.
Maureen Baikie, Dr. Jeff Scott, Dr. David
Haldane, Dr. Monika Naus, Dr. Adeola
Jaiyeola, Dr. Chander Krishnan, Albert
Haddad, Dr. Lamont Sweet, Dr. Lewis
Abbott, Dr. André Beauchense, Dr. Michel
Brazeau, Dr. Thérèse Leboeuf-Trudeau, Dr.
Terry Nan Tannenbaum, Louise Alain,
Odette Laplante, Louise Thibert, Dr.
Edward Chan, Dr. Gregory Horsman, Tom
Martin, Iris Natyshak, Anne Grauwiler,
Cheryl Hobson, Joanne Crosbie. Thanks to
Ms. Dorothy Helbecque, National
Reference Centre for Tuberculosis, LCDC,
for her valuable contribution to provision of
laboratory data; Dr. Jure Manfreda and Dr.
Rama Nair for their constructive advice; and
Dr. Louise Pelletier and Penny Nault,
Division of Tuberculosis Prevention and
Control, LCDC, for technical support and
data management.
REFERENCES
1. Antituberculosis Drug resistance in the World,
The WHO/IUATLD Global Project on
Antituberculosis Drug Resistance Surveillance
1994-1997, WHO 1997.
2. Edlin BR, Tokars JL, Grieco MH, et al. An out-
break of multidrugresistant tuberculosis among
hospitalized patients with the acquired immun-
odeficiency syndrome. N Engl J Med
1992;326:1514-21.
3. Frieden T, Sterling T, Pablos-Méndez A, et al.
The emergence of drug-resistant tuberculosis in
New York City. N Engl J Med 1993;328:521-26.
4. Monno L, Angarano G, Carbonara S, et al.
Emergence of drug resistant Mycobacterium
tuberculosis in HIV-infected patients. Lancet
1991;337:852.
5. Herrera D, Cano R, Godoy P, et al. Multidrug-
resistant tuberculosis outbreak on an HIV ward-
Madrid, Spain, 1991-1995. MMWR
1996;45:330-33.
6. Eidus L, Jessamine AG, Hershfield ES,
Helbecque DM. A national study to determine
the prevalence of drug resistance in newly discov-
ered previously untreated tuberculosis patients as
well as in retreatment cases. Can J Public Health
1978;69:146-53.
7. Long R, Fanning A, Cowie R, et al., and the
Western Canadian Tuberculosis Group.
Antituberculous drug resistance in western
Canada. Can Respir J 1997;4(2):71-75.
8. Long R, Manfreda J, Mendella L, et al.
Antituberculous drug resistance in Manitoba
from 1980 to 1989. CMAJ 1993;148:1489-95.
9. Manns BJ, Fanning EA, Cowie RL.
Antituberculosis drug-resistance in immigrants to
Alberta, Canada, with tuberculosis, 1982-1994.
Int J Tuberc Lung Dis 1997;1(3):225-30.
10. Kochi A, Vareldzis B, Styblo K. Multidrug resis-
tant tuberculosis and its control. Res Microbiol
1993;144:104-10.
11. Goble M, Iseman MD, Madsen LA, et al.
Treatment of 171 patients with pulmonary
tuberculosis resistant to isoniazid and rifampin.
N Engl J Med 1993;328:527-32.
12. Mahmoudi A, Iseman MD. Pitfalls in the care of
patients with tuberculosis. JAMA 1993;270:65-
68.
13. Canadian Tuberculosis Standards, Fourth
Edition, Canadian Lung Association, 1996.
14. Tuberculosis in Canada, 1996, Health Canada.
15. Statistics Canada. Causes of Death, 1995, cata-
logue number 84-208-XPB, Ministry of
Industry, Ottawa, Canada, 1997.
16 Tuberculosis Statistics, 1990, Statistics Canada.
17. Laszlo A, Gill P, Handzel V, et al. Conventional
and radiometric drug susceptibility testing of
Mycobacterium tuberculosis. J Clin Microbiol
1983;18:1335-39.
18. Case Notification Update, 1996, Global
Tuberculosis Program, World Health
Organization, Geneva.
19. Tuberculosis Statistics, 1993, Statistics Canada.
20. Rivest P, Tannenbaum T, Bédard L.
Epidemiology of tuberculosis in Montreal.
CMAJ 1998;158:605-9.
21. Hersi A, Elwood K, Cowie R, et al. Multidrug-
resistant tuberculosis in Alberta and British
Columbia 1989 to 1998. Can Respir J
1999;6:155-60.
Received: August 16, 1999
Accepted: February 15, 2000
VOLUME 91, NO. 5
A B S T R A C T
This paper examines structured opportunities
for student physical activity in Ontario elemen-
tary and secondary schools. Random samples of
elementary and secondary schools were selected,
and telephone surveys of 353 elementary and
360 secondary school personnel were conducted
in 1998. The findings indicate that elementary
schools offered physical education, on average,
just under three days per week. The duration of
physical education class and, in some cases, the
duration of vigorous physical activity in class
and the weekly amount of vigorous activity in
class, were significantly higher at successive
grade levels. The physical education enrollment
rate in secondary schools was significantly lower
at successive grade levels. Student participation
in secondary school intramurals and inter-
school sports was 22.8% and 28.7% respective-
ly. These findings suggest increasing the weekly
frequency of physical education in elementary
schools and increasing participation in physical
education classes in secondary schools. Also,
intramural opportunities and participation
should be promoted.
A B R É G É
Ce document étudie la pratique de l’activité
physique encadrée chez les écoliers du primaire
et du secondaire en Ontario. On a sélectionné
des échantillons aléatoires puis mené des
sondages par téléphone auprès de 353 membres
du personnel d’écoles primaires et de 360 mem-
bres du personnel d’écoles secondaires, en 1998.
Les conclusions montrent que les écoles élémen-
taires offrent des cours d’éducation physique un
peu moins de trois jours par semaine en
moyenne. On observe que la durée du cours
(dans certains cas, l’activité physique y est pra-
tiquée de façon intense), et le nombre de cours
d’éducation physique par semaine, augmentent
de manière importante en proportion des
niveaux successifs. On constate, par contre, que
l’inscription aux cours d’éducation physique
dans les écoles secondaires baisse de manière sig-
nificative avec les niveaux successifs. La partici-
pation des étudiants du niveau secondaire dans
les sports internes et interscolaires était de
22,8 % et de 28,7 % respectivement. Ces con-
clusions indiquent une augmentation de la
fréquence hebdomadaire aux cours d’éducation
physique dans les écoles primaires et une aug-
mentation de la participation aux cours d’éduca-
tion physique dans les écoles secondaires. Par
conséquent, on devrait promouvoir les possibi-
lités internes d’activité physique et la participa-
tion à l’activité physique.
SEPTEMBER – OCTOBER 2000
Structured Opportunities for Student
Physical Activity in Ontario
Elementary and Secondary Schools
Kenneth R. Allison, PhD,1 Edward M. Adlaf, PhD 1,2
Increasing evidence supports the positive
role of regular physical activity both in pre-
venting disease and promoting health and
well-being.1 Physical activity and fitness are
associated with lower adult mortality rates
for several conditions, 1-5 as well as
enhanced physical and mental health. 6
Physical activity during childhood and
adolescence also has health benefits7 and
may have important implications for activ-
ity levels in adult life.8,9 The problem is
that participation in regular physical activi-
ty declines with age, and among females,
within the teenage years.10-12 Participation
is at least partly influenced by opportuni-
ties to engage in regular physical activi-
ty.13,14 Thus, policies and curricula con-
cerning the type and amount of activity
offered in physical education class and
other school-related settings (intramurals,
inter-school sports) represent structured
opportunities for students.
A number of federal and provincial orga-
nizations in Canada have promoted the
notion of “quality daily physical educa-
tion” (QDPE), or daily physical activity,
for elementary and secondary level stu-
dents,15-17 and this program’s effectiveness
was assessed recently.18
The study reported here did not assess
QDPE specifically, but examined the avail-
1. Department of Public Health Sciences, Faculty
of Medicine, University of Toronto
2. Centre for Addiction and Mental Health,
Toronto, ON
Correspondence and reprint requests: Dr. Kenneth
R. Allison, Associate Professor, Department of Public
Health Sciences, Faculty of Medicine, University of
Toronto, 12 Queen’s Park Crescent West, Toronto,
ON, M5S 1A8, Tel: 416-978-5869, Fax: 416-978-
2087, E-mail: k.allison@utoronto.ca.
This research was supported by a grant from the
Ontario Ministry of Health. Dr. Kenneth R. Allison
is supported by an Ontario Ministry of Health
Career Scientist award.
CANADIAN JOURNAL OF PUBLIC HEALTH 371
ability and utilization of physical education
and other types of physical activity in
Ontario elementary and secondary schools.
An examination of opportunities con-
tributes to a better understanding of the
role of schools in helping children and
youth attain sufficient levels of physical
activity. Findings from this study can be
used to provide baseline data and to help
set objectives for school-based physical
activity.
METHODS
Sample
The data are based on two systematic
random samples generated by the Institute
for Social Research (ISR), York University,
using information from the Ontario
Ministry of Education’s Directory and
MIDENT file. The target population rep-
resents 3,900 elementary schools and 800
secondary schools of the regular public and
Catholic school systems. It was determined
that a sample of 350 schools per school
type would provide sufficient precision for
both cross-sectional and baseline estimates.
The sample resulted in completions of 353
elementary and 360 secondary schools,
representing response rates of 73.2% and
74.5%, respectively.
Description of respondents and schools
Respondents at the elementary school
level consisted primarily of principals or
vice-principals (58.3%), physical education
teachers (17.8%), or others (23.9%, i.e.,
classroom teachers, resource teachers, and
others). Approximately half (49.9%) of the
schools participating in the survey repre-
sented a range of grades from JK to grade
8, while 12.7% of schools included JK to
grade 6, and the remaining schools includ-
OPPORTUNITIES FOR PHYSICAL ACTIVITY
ed other grade ranges. The mean number
of students enrolled in the elementary
schools was 365.3. Grades 1, 3, 6, and 8
were selected for the study since provincial
guidelines provided in the Common
Curriculum describe learning outcomes for
physical education for some of these grades
(3 and 6).
Unlike the elementary school survey,
secondary school respondents consisted
primarily of physical education teachers
(72.5%). Additional respondents consisted
of principals or vice-principals (11.1%),
and others (16.4%). Most of the schools
represented in the survey (86.1%) con-
tained a range of grades beginning with
grade 9 and ending with OAC (previously
known as grade 13). The mean number of
students enrolled in secondary schools
included in the study was 880.9.
With one exception, there were no dif-
ferences in outcomes based on the type of
respondent at either the elementary or sec-
ondary school level.
Data collection
The study was approved by the universi-
ty ethics committee. The questionnaires,
which included items dealing with oppor-
tunities and participation in physical edu-
cation, intramurals, and inter-school sports
(Table I), were developed with input from
an advisory committee. A computer-assisted
telephone interviewing (CATI) version of
the questionnaire was produced and pre-
tested by ISR. The data collection period
was June 1998 for the elementary schools
and June-October, 1998 for the secondary
schools. Telephone interviews were con-
ducted with a representative from each
school using trained ISR interviewers. The
interviewers asked to speak with an indi-
vidual knowledgeable about the school’s
physical education program. Completed
interviews averaged nine minutes in both
elementary and secondary schools.
Data analysis
The findings reported here include
both descriptive and bivariate analysis.
Pearson correlations were used to examine
the linear relationship between continu-
ous variables. Cross-tabulations of cate-
gorical variables utilized the chi-square
test. Analysis of variance was used to
372 REVUE CANADIENNE DE SANTÉ PUBLIQUE
TABLE I
Measures Used
Elementary Schools Only
Frequency - the average number of days per week that students have physical education class in
grades 1, 3, 6, and 8. (range 1-5)
Duration of physical education class - the number of minutes in the typical physical education class
in grades 1, 3, 6, and 8. (range 10-61)
Duration of vigorous physical activity - in a physical education class, the number of minutes of vig-
orous physical activity the typical student receives in grades 1, 3, 6, and 8. (A physical activity is
vigorous if it makes a student’s heart beat faster and makes them breathe a lot faster than normal).
(range 0-80)
Weekly minutes of vigorous physical activity - a derived variable created by multiplying frequency
times duration of vigorous physical activity for grades 1, 3, 6, and 8. (range 0-300)
Participation rate in intramural program - the percentage of students participating in an intramural
program of organized physical activity between January, 1998 and the end of the school term.
Participation rate in inter-school sports - the percentage of students participating in an inter-school
sports program between January, 1998 and the end of the school term.
Responsibility for teaching physical education - who teaches most of the physical education classes
in the school: a classroom teacher, a physical education specialist, a combination of classroom
teachers and physical education specialist, a parent or volunteer, or some other arrangement.
Secondary Schools Only
Offering physical education - whether the school offers curriculum-based physical education in
classes in grades 9, 10, 11, 12, and OAC. (yes/no)
Enrollment in physical education courses (PEER) - a derived variable based on enrollment in physi-
cal education courses as a proportion of the total number of students enrolled in a particular grade
in schools offering physical education classes.
Participation rate in intramural program - a derived variable indicating the proportion of students
participating in an intramural program during the spring (1998) term, based on the number of stu-
dents participating divided by the number of students enrolled in schools offering an intramural pro-
gram.
Participation rate in inter-school sports - a derived variable indicating the proportion of students par-
ticipating in inter-school sports programs during the spring (1998) term, based on the number of stu-
dents participating divided by the number of students enrolled in schools offering an inter-school
sports program.
Both Elementary and Secondary Schools
Offering intramural program - having an intramural program of organized physical activity in the
school. (yes/no)
Offering inter-school sports program - having an inter-school sports program. (yes/no)
School board policy - whether school boards have written guidelines to ensure safe participation in
curriculum-based physical education classes. Similar questions dealt with intramural physical activi-
ty programs and inter-school sports. (yes/no)
Use of community recreation resources - school use of community recreation resources or programs
that are available to students. (yes/no)
Region - based on telephone area exchange numbers of schools included in the survey. (area codes
416, 519, 613, 705, 807, and 905)
Respondent’s position - principal, vice-principal, physical education specialist/physical education
teacher, other.
examine differences in continuous out- RESULTS
comes by categorical level variables. In
order to examine the significance of grade Elementary schools
level differences in outcomes, sample
t-tests with unequal variance were used. Physical Education Classes
Nine key indicators (outcome variables) All respondents in the survey (100%)
in the analysis included: frequency of reported that students participate in physi-
physical education class, duration, min- cal education classes as part of the curricu-
utes of vigorous activity in class, and lum at their school. There was little vari-
weekly amount of vigorous activity (ele- ability by grade level in the average num-
mentary level); physical education enroll- ber of days per week (just under 3) that
ment rate (secondary level); and intra- students engaged in physical education
mural offering and participation, inter- classes (not shown in tables). However,
school sports offering and participation there was more variability by grade in the
(both levels). average number of minutes included in a
VOLUME 91, NO. 5
 OPPORTUNITIES FOR PHYSICAL ACTIVITY

TABLE II
Grade Differences in Duration of Physical Education Class, Duration of Vigorous Physical Activity, and Weekly Minutes
of Vigorous Physical Activity
Duration of Physical Duration of Vigorous Weekly Minutes of
Education Class (mins.) Physical Activity (mins.) Vigorous Physical Activity (mins.)
Grade N M P* SD N M P* SD N M P* SD

One 311 35.9 - 6.7 238 14.2 - 7.4 205 42.6 - 30.9
Three 309 37.5 0.002 6.2 238 16.4 0.002 7.9 204 46.7 NS 28.5
Six 292 39.7 0.001 7.1 230 19.7 0.001 10.1 193 55.9 0.005 35.3
Eight 235 42.3 0.001 7.7 178 21.5 NS 10.2 145 57.9 NS 33.2

* Based on two sample t-tests, comparing subsequent grade levels (e.g., grade three compared to grade one).
amount of weekly vigorous physical activi-
TABLE III ty was unrelated to type of respondent,
Intramural and Inter-school Sports Participation region, or primary responsibility for teach-
Elementary Schools ing physical education classes (not shown
N Percent Mean Percent in tables).
Schools with intramurals 353 88.4
Schools with inter-school sports 352 92.0 Intramural and Inter-school Sports Programs
Students participating - intramurals 301 57.8 The majority of elementary schools
Students participating - inter-school sports 314 31.4
(88.4%) reported offering an intramural
N = the number of schools responding. program (Table III). Having an intramural
program was related to school size (F=6.57,
TABLE IV p<0.01), with schools offering programs
Grade Differences in Physical Education Enrollment Rate (PEER)* more likely to have higher enrollment.
Secondary Schools However, offering an intramural program
Grade N PEER (M) P† SD
was unrelated to region, respondent’s posi-
tion, responsibility for teaching physical
Nine 337 0.95 - 0.16 education, and whether or not schools
Ten 288 0.63 0.001 0.21
Eleven 275 0.54 0.001 0.35 make use of community recreation
Twelve 258 0.45 0.001 0.23 resources. During the period January-June,
OAC 52 0.33 0.002 0.25
1998, 57.8% of elementary school stu-
* PEER represents enrollment in physical education courses as a proportion of the total number of dents participated in these intramural pro-
students enrolled in a particular grade in schools offering physical education classes.
† based on t-tests comparing subsequent grade levels (e.g., grade ten compared to grade nine). grams. The intramural participation rate
was inversely related to school size
TABLE V (r=-0.293, p<0.001), and was related to
Intramural and Inter-school Sports Participation type of respondent (F=3.17, p<0.05),
Secondary Schools although multiple comparison tests showed
no significant differences between specific
N Percent Participation Rate (M %)
pairs of respondent (principals vs. teachers).
Schools with intramurals 355 67.0 The intramural participation rate was unre-
Schools with inter-school sports 355 86.2
Students participating - intramurals 217 22.8 lated to region, responsibility for teaching
Students participating - physical education, and use of community
inter-school sports 292 28.7
recreation resources (not shown in tables).
N = the number of respondents at the school level reporting the information. Almost all (92.0%) elementary schools
reported offering an inter-school sports
typical physical education class. The length II). Furthermore, the average weekly program (Table III), and offering a sports
of class time devoted to physical education amount of vigorous physical activity was program did not differ significantly by
was significantly greater (by approximately significantly higher at grade 6 compared to school size, region, responsibility for teach-
two minutes) for each of the years included grade 3 (Table II). Within grade levels, the ing physical education, type of respondent,
(Table II). weekly amount of vigorous physical activi- or use of community recreation resources.
The findings also indicate a significantly ty was negatively correlated with school About one third (31.4%) of students par-
greater average number of minutes of vig- size (ranging from r=-0.188 to r=-0.233). ticipated in these inter-school sports pro-
orous physical activity in physical educa- That is, the weekly amount of vigorous grams between January-June, 1998. School
tion class at grade 3 compared to grade 1, physical activity was lower in larger schools size was inversely related to participation
and at grade 6 compared to grade 3 (Table at each grade level surveyed. However, the (r= -0.161, p<0.01), but region, responsi-

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 373

OPPORTUNITIES FOR PHYSICAL ACTIVITY
bility for teaching physical education, type
of respondent, and use of community
recreation resources were unrelated to
sports participation (not shown in tables).
Secondary schools
Curriculum-based Physical Education
With the exception of OAC, there was
little variability between grades regarding
the offering of curriculum-based physical
education classes. Most schools (about
98%) included curriculum-based physical
education classes offered for grades 9-12.
However, only 21.1% of the schools
reported offering curriculum-based physi-
cal education classes at the OAC level (not
shown in tables).
Physical Education Enrollment Rate
Table IV indicates that the physical edu-
cation (course) enrollment rate (PEER)
was significantly lower by grade level, with
the highest physical education class enroll-
ment rate in grade 9 (0.95), and the lowest
rate at the OAC level (0.33). PEER was
inversely correlated with school size in the
case of grades 9 (r=-0.109, p<0.05) and 10
(r=-0.20, p<0.001), but was not correlated
with school size for grades 11, 12, and
OAC. PEER did not differ significantly by
region or type of respondent (not shown in
tables).
Intramural and Inter-school Sports Programs
Sixty-seven percent of schools reported
offering an intramural program at the sec-
ondary level (Table V). Whether or not
schools offered an intramural program was
unrelated to school size, geographic region,
respondent’s position, and whether schools
make use of community recreation
resources. Within schools offering an intra-
mural program, an intramural participa-
tion rate of 22.8% was calculated, based
on the number of students participating in
the spring term (January-June,1998) divid-
ed by the number of students enrolled in
schools offering an intramural program.
The intramural participation rate was neg-
atively correlated (r= -0.34, p<0.001) with
school size. Thus, larger schools were more
likely to have lower rates of intramural par-
ticipation than smaller schools. The intra-
mural participation rate was unrelated to
374 REVUE CANADIENNE DE SANTÉ PUBLIQUE
geographic region, type of respondent, and
whether or not schools make use of com-
munity recreation resources (not shown in
tables).
Most schools (86.2%) reported having
an inter-school sports program (Table V).
Whether or not schools offered an inter-
school sports program was related to geo-
graphic region (Chi-square=23.52,
p<0.001). Whether or not schools offered
inter-school sports was unrelated to school
size, type of respondent, and whether or
not schools used community recreation
resources (not shown in tables).
For schools offering inter-school sports,
a participation rate of 28.7% was calculat-
ed (Table V). The inter-school sports par-
ticipation rate was related to geographic
region, with multiple comparison tests
showing a significantly higher rate in one
of the northern Ontario regions (area code
705) compared to Metropolitan Toronto
(area code 416). The inter-school sports
participation rate was negatively related to
school size (r=-0.23, p<0.001). However,
the inter-school sports participation rate
was unrelated to type of respondent, and
whether or not schools used community
recreation resources (not shown in tables).
DISCUSSION
The findings raise interesting questions
concerning whether opportunities provid-
ed for structured school-based physical
activity are sufficient in terms of such char-
acteristics as activity type, frequency, dura-
tion, and intensity. Several organizations
recommend offering physical education on
a daily basis.15-17,19,20 Moreover, an interna-
tional consensus statement on physical
activity guidelines recommends that, in
addition to daily (total) physical activity,
“adolescents should engage in three or
more sessions per week of activities that
last 20 minutes or more at a time and that
require moderate to vigorous levels of exer-
tion.”21 If the standard of comparison is
the notion of quality daily physical educa-
tion, it would appear that Ontario elemen-
tary schools are at a lower than optimal
level. Although school-based curricula for
physical education are available on a uni-
form basis for all grades surveyed, physical
education in elementary and middle
schools is offered just under three days per
week.
The findings indicate that the reported
duration of typical physical education
classes and, in some cases, the duration of
vigorous physical activity in class and the
weekly amount of vigorous physical activi-
ty were significantly higher by grade level
up to grade 6. However, even the upper
level elementary and middle school stu-
dents are offered amounts of vigorous
physical activity in physical education
classes only approaching, or marginally
within, the suggested guidelines for chil-
dren and adolescents.21 From another per-
spective, one of the Healthy People 2000
(U.S.) objectives is “to increase to at least
50% the proportion of school physical
education time that students spend being
physically active… .”22 In the current study
(using information from Table II), the
proportion ranged from 39.5% (grade 1)
to 50.8% (grade 8).
At the secondary school level, curriculum-
based physical education classes appear to
be available to students at most grade levels
(other than OAC). However, there were
significantly lower physical education
course enrollment rates at subsequent
grades examined. This phenomenon is
partly explained by provincial require-
ments for a single physical education cred-
it, normally taken in grade nine. However,
the findings suggest declining enrollment
at each grade level surveyed, supported by
additional studies on age differences in
physical activity participation in Canada
and elsewhere.7-9
The problem of declining physical edu-
cation course enrollment by grade is com-
pounded by low levels of participation
(22.8%) in school-based intramural pro-
grams at the secondary school level. The
reasons for low levels of participation in
intramural programs were not determined
in this study. Additional information con-
cerning the range of intramural activities is
needed, along with data dealing with barri-
ers to participating in intramurals. Such
issues as the time and facilities available,
and students’ interest in participating, need
to be explored more fully.
Limitations of the study included the
use of different types of respondents, the
use of measures of unconfirmed validity
VOLUME 91, NO. 5

and reliability, and the sole reliance on
data provided by respondents. We found
only one instance of a significant difference
in outcome by type of respondent.
Regarding validity and reliability of the
measures, while we did not undertake to
examine these in detail, a pre-test conduct-
ed by ISR indicated that the measures had
face validity and could be used reliably by
trained interviewers. Although the accura-
cy of data obtained was not confirmed
through observation or other means, we
attempted to obtain the best reported
information possible in a telephone-
administered short survey of school per-
sonnel.
While the current study examined some
important quantitative indicators of struc-
tured opportunities for physical activity in
elementary and secondary schools, further
questions remain. Additional information
on patterns and predictors of student par-
ticipation in vigorous physical activity is
needed. In addition, information is needed
on more qualitative aspects of structured
physical activity opportunities.
Based on the current study, implications
for the field include the need to establish
objectives for school-based physical activity
in the various settings – physical education
class, intramurals, and inter-school sports.
In particular, the adequacy of an average of
three physical education classes per week in
elementary schools needs to be addressed.
Also, the problem of declining enrollment
in secondary school physical education
courses requires attention, both in terms of
the factors influencing this pattern, and
ways of increasing enrollment. Finally,
school-based intramural activity, particu-
larly at the secondary school level, needs to
SEPTEMBER – OCTOBER 2000
be further developed through policy sup-
port, increased opportunities for a variety
of activities, and the promotion of these
programs to students.
ACKNOWLEDGEMENTS
The following individuals provided use-
ful advice to the study: Charles Clayton,
Jennifer Cowie Bonne, Michelle
Brownrigg, David Carmichael, Erica de
Ruggiero, Terry McKinty, David
Northrup, and two anonymous reviewers.
REFERENCES
1. Pate R, Pratt M, Blair S, et al. Physical activity
and public health: A recommendation from the
Centers for Disease Control and Prevention and
the American College of Sports Medicine. JAMA
1995;273:402-7.
2. Blair S, Kohl H, Paffenbarger R, et al. Physical
fitness and all-cause mortality: A prospective
study of healthy men and women. JAMA
1989;262:2395-401.
3. Powell K, Thompson P, Caspersen C, Kendrick
J. Physical activity and the incidence of coronary
heart disease. Annu Rev Public Health
1991;8:253-87.
4. Powell K, Blair S. The public health burdens of
sedentary living habits: Theoretical but realistic
estimates. Med Sci Sports Exerc 1996;26:851-56.
5. U.S. Department of Health and Human Services.
Physical Activity and Health: A Report of the
Surgeon General. Atlanta: U.S. Department of
Health and Human Services, Centers for Disease
Control and Prevention, National Center for
Chronic Disease Prevention and Health
Promotion, 1996.
6. Stephens T. Physical activity and mental health
in the United States and Canada: Evidence from
four population surveys. Prev Med 1988;17:35-
47.
7. Guidelines for school and community programs
to promote lifelong physical activity among
young people. National Center for Chronic
Disease Prevention and Health Promotion,
Centers for Disease Control and Prevention.
J Sch Health 1997;67:202-19.
8. Dishman R, Dunn A. Exercise adherence in chil-
dren and youth: Implications for adulthood. In:
Dishman R, Dunn A (Eds.), Exercise Adherence:
CANADIAN JOURNAL OF PUBLIC HEALTH 375
OPPORTUNITIES FOR PHYSICAL ACTIVITY
Its Impact on Public Health. Champaign: Human
Kinetics, 1988.
9. Malina R. Tracking of physical activity and phys-
ical fitness across the lifespan. Res Q Exerc Sport
1996;57:48-57.
10. Sallis J. Epidemiology of physical activity and fit-
ness in children and adolescents. Crit Rev Food
Sci Nutr 1993;33:403-8.
11. Heath G, Pratt M, Warren C, Kann L. Physical
activity patterns in American high school stu-
dents. Arch Pediatr Adolesc Med 1994;148:1131-
36.
12. Allison K, Adlaf E. Age and sex differences in
physical inactivity among Ontario teenagers. Can
J Public Health 1997;88:177-80.
13. King A, Jeffery R, Fridinger F, et al.
Environmental and policy approaches to cardio-
vascular disease prevention through physical
activity: Issues and opportunities. Health Educ Q
1995;22:499-511.
14. McKenzie T. School health-related physical activ-
ity programs: What do the data say? JOPERD
1999;70:16-19.
15. The Canadian Association for Health, Physical
Education and Recreation. The QDPE Leader’s
Lobbying Kit. Ottawa: Government of Canada
Fitness and Amateur Sport, 1993.
16. The Ontario Physical and Health Education
Association. Towards QDPE: Final Summary
Report on 1995 Needs Assessment Survey.
Toronto: Ontario Physical and Health Education
Association, 1995.
17. Canadian Fitness and Lifestyle Research Institute.
Physical Inactivity in Ontario: A Framework for
Action. Ottawa: Canadian Fitness and Lifestyle
Research Institute, 1996.
18. Chad K, Humbert M, Jackson P. The effective-
ness of the Canadian Quality Daily Physical
Education program on school physical education.
Res Q Exerc Sport 1999;70:55-64.
19. Ontario Ministry of Health. Promoting Heart
Health. Toronto: Queen’s Printer for Ontario,
1993.
20. Report of the Royal Commission on Learning.
Toronto: Queen’s Printer for Ontario, 1995.
21. Sallis J, Patrick K. Physical activity guidelines for
adolescents: Consensus statement. Ped Exerc Sc
1994;6:302-14.
22. Healthy people 2000: National health promotion
and disease prevention objectives and healthy
schools (adapted from Healthy People 2000:
National Health Promotion and Disease
Prevention Objectives (DHHS Pub. No. (PHS)
91-50212). J Sch Health 1991;61(7):298-328.
Received: August 3, 1999
Accepted: March 6, 2000
 A B S T R A C T
Using data from a 1996 random-digit-
dialling computer-assisted telephone survey of
Ontario adults, 424 smokers and 1,340 non-
smokers were compared regarding knowledge
about the health effects of tobacco use, attitudes
toward restrictions on smoking and other tobac-
co control measures, and predictions of compli-
ance with more restrictions. The response rate
was 65%. Smokers were less knowledgeable
than nonsmokers. Smokers were also less likely
to support bans on smoking in specific loca-
tions, but majorities of both groups supported
some restriction in most settings. Smokers were
more likely than nonsmokers to predict that
most smokers would comply with more restric-
tions, and more than three quarters indicated
that they, themselves, would comply. Sizable
proportions of both groups, especially smokers,
failed to appreciate the effectiveness of taxation
in reducing smoking. Support for other control
measures also differed by smoking status. Both
knowledge and smoking status were indepen-
dently associated with support for more restric-
tions and other tobacco control policy measures.
A B R É G É
À partir de données recueillies en 1996 lors
d’un sondage téléphonique à numéros aléatoires
assisté par ordinateur et mené auprès d’adultes
de l’Ontario, nous avons comparé les réponses
de 424 fumeurs et de 1 340 non-fumeurs con-
cernant les effets du tabagisme sur la santé, les
attitudes à l’égard de restrictions et d’autres
mesures de contrôle de l’usage du tabac et les
prédictions quant au respect de restrictions plus
nombreuses. Le taux de réponse au sondage a
été de 65 %. Les fumeurs avaient moins de con-
naissances que les non-fumeurs. Les fumeurs
étaient également moins portés à appuyer
l’interdiction de fumer à certains endroits précis,
mais la majorité des répondants des deux
groupes appuyaient l’idée d’imposer certaines
restrictions dans la plupart des endroits. Les
fumeurs étaient plus portés que les non-fumeurs
à prédire que la plupart des fumeurs se con-
formeraient à un plus grand nombre de restric-
tions; plus des trois-quarts d’entre eux ont
indiqué qu’eux-mêmes s’y conformeraient. Une
grande proportion des deux groupes, surtout
parmi les fumeurs, ne croyait pas que les taxes
étaient un moyen efficace de réduire l’usage du
tabac. Les deux groupes ne donnaient pas le
même appui à d’autres mesures de contrôle du
tabagisme. La connaissance et le statut de
fumeur ou de non-fumeur étaient associés de
façon indépendante à l’appui donné à des
restrictions plus nombreuses et à d’autres
mesures de contrôle de l’usage du tabac.
376 REVUE CANADIENNE DE SANTÉ PUBLIQUE
Knowledge About Tobacco and
Attitudes Toward Tobacco Control:
How Different are Smokers and
Nonsmokers?
Mary Jane Ashley, MD,1 Joanna Cohen, PhD,2 Shelley Bull, PhD,3
Roberta Ferrence, PhD,4 Blake Poland, PhD,2 Linda Pederson, PhD,5 Joseph Gao, PhD6
The public health impact of tobacco use
in Canadian society is well documented.
Recent estimates indicate that each year
tobacco accounts for 33,500 to 41,500
deaths, almost half a million years of lost
life, more than 200,000 hospital separa-
tions, 3 million hospital days and $15 bil-
lion in economic costs.1-4 Despite this enor-
mous toll, smoking remains prevalent; in
1996-97 almost 7 million Canadians were
smokers. 5 Public policies and programs
aimed at reducing smoking and exposure to
environmental tobacco smoke (ETS) are
key components in a comprehensive
approach to eliminating the use of tobacco
products in Canada.6 Although progress has
been made, including the enactment of
restrictions on smoking in some loca-
tions,7,8 much remains to be done.
Current information about knowledge of
the health impacts of tobacco use and atti-
tudes toward control measures in smokers
and nonsmokers can be valuable in inform-
ing the public education and policy devel-
opment processes. In previous work,9,10,11-13
1. Professor, Department of Public Health Sciences,
University of Toronto
2. Assistant Professor, Department of Public Health
Sciences, University of Toronto
3. Senior Scientist, Division of Clinical
Epidemiology, Samuel Lunenfeld Research
Institute
4. Senior Scientist, Centre for Addiction and
Mental Health, Addiction Research Foundation
Division, Toronto
5. Professor, Director of Research, Department of
Community Health and Preventive Medicine,
Morehouse School of Medicine, Atlanta, Georgia
6. Biostatistician, Division of Clinical
Epidemiology, Samuel Lunenfeld Research
Institute
Correspondence: Dr. Mary Jane Ashley, Department
of Public Health Sciences, University of Toronto,
Toronto, ON, M5S 1A8, Tel: 416-978-2751, Fax:
416-978-8299, E-mail: maryjane.ashley@utoronto.ca
differences between smokers and nonsmok-
ers were found. Informational strategies
specifically targeted to smokers may be
needed. Further, explicit data on the atti-
tudes of smokers to restrictions on smoking
and their predictions of compliance with
more restrictions may help counteract
activities promoted by the tobacco industry
regarding “smokers’ rights” and “smokers’
revolts”,14,15 and allay anxieties that policy
makers may have about noncompliance.
We present comparative information on
smokers and nonsmokers in Ontario and
discuss the implications of the findings for
policy and program development.
METHODS
A telephone survey of a representative
sample of adult Ontarians was conducted
in 1996.16,17 A two-stage probability process
was used to select respondents 18 years of
age and older. First, households were select-
ed using random digit dialling. A respon-
dent within each household was then
selected based on most recent birthday.18
To maximize the chances of getting a com-
pleted interview from each sample number,
at least 12 call attempts were made during
the day and evening, both during the week
and on the weekend. Because Metropolitan
Toronto was over-sampled to allow for
comparison with previous surveys, both
household weights and regional weights
were computed for each respondent.
A computer-assisted telephone inter-
view, lasting about 20 minutes on average,
addressed the respondent’s sociodemo-
graphic characteristics; smoking history;
knowledge of the health effects of smoking
and exposure to ETS and of the public
VOLUME 91, NO. 5
 DIFFERENCES BETWEEN SMOKERS AND NONSMOKERS

health impact of tobacco in Canada; atti- TABLE I

tudes towards restrictions on smoking in Knowledge of Health Effects of Smoking and
specific settings and predicted compliance Environmental Tobacco Smoke (ETS)
with more restrictions; and attitudes
toward other tobacco-control measures, Nonsmokers (n=1340) Smokers (n=424)
Percent 95% CI Percent 95% CI
including packaging and sales, prohibition
of advertising, restrictions on sales to Smoking is a cause of:
Lung cancer* 87.0 84.5, 89.4 67.4 61.3, 73.6
minors, and informational package inserts. Chronic bronchitis* 75.4 72.2, 78.7 65.7 59.4, 72.0
Perceptions of the effectiveness of taxation Pregnancy complications* 67.3 63.8, 70.8 51.1 44.5, 57.8
Heart attacks* 60.2 56.5, 63.9 45.1 38.4, 51.8
measures in reducing smoking were also ETS is a cause of:
assessed (a copy of the questionnaire can Chest problems in children* 58.0 54.3, 61.8 35.8 29.5, 42.2
Lung cancer* 54.1 50.3, 57.8 32.6 26.2, 39.1
be obtained from the first author). Heart attacks* 30.4 26.9, 33.8 17.7 12.5, 22.9
Interviews were completed with 1,764 Ear problems in children 14.3 11.7, 16.8 11.4 7.2, 15.6
respondents, yielding a response rate of Daily smokers addicted: All/most 87.6 85.4, 89.9 82.6 77.5, 87.8
65%, based on the estimated number of Quitting smoking can improve
eligible households. By comparing the health even after having smoked
sociodemographic characteristics of the a lot for a long time: Strongly agree 73.1 69.5, 76.6 66.7 60.2, 73.2
sample with the adult population (1991 Number of cigarettes that can be
census), it was determined that the sample smoked daily without increasing
the risk of a serious health
was representative with respect to age, sex, problem: None* 49.7 46.2, 53.3 33.8 28.0, 39.6
and marital status, but under-represented Tobacco causes a lot more deaths than:
those with low levels of education, a com- Alcohol 23.3 20.0, 26.6 17.3 11.6, 22.9
mon finding in telephone surveys.19 The AIDS* 34.6 30.8, 38.4 25.2 18.9, 31.4
proportions of never smokers (51.2%), for- * indicates that the difference between nonsmokers and smokers is statistically significant at the
mer smokers (24.7%) and current smokers p < 0.05 level.
(24.7%) were in keeping with those found Summaries of knowledge regarding the problems. Overall, smokers were much less
in other recent provincial surveys.20 health effects of both smoking and ETS likely than nonsmokers to indicate that
Based on findings from an earlier survey exposure, each ranging between 0 and 4, smoking is a cause of lung cancer, chronic
showing similar knowledge and attitudes were obtained by summing each correct bronchitis, complications in pregnancy,
among never and former smokers,9 these answer to the conditions reported in Table and heart attacks (Table I). Both groups
groups were combined, yielding 424 smok- I. We defined support as “majority” if the were most knowledgeable about the causal
ers for comparison with 1,340 nonsmok- lower CI of the percentage in question role of smoking in lung cancer and least
ers. These sample sizes were sufficient to exceeded 50%; “clear majority” support aware about smoking and heart attacks.
detect differences of eight percentage required that the lower CI be at least 60%, While majorities of nonsmokers recog-
points. Observations were weighted while “substantial minority” support nized the causal role of smoking in all four
according to the probability of being required the lower CI to be at least 40%. conditions, among smokers this was the
selected into the sample. Responses were case only for lung cancer and chronic
tabulated as weighted percentages, and cor- RESULTS bronchitis. Although both nonsmokers and
responding 95% confidence intervals (CIs) smokers knew less about the health effects
were calculated using standard errors esti- Sociodemographic characteristics of ETS, majorities of nonsmokers
mated according to the survey design.21-23 The male/female ratio was less than one acknowledged the causal role of ETS in
A statistically significant difference at the in nonsmokers and greater than one in chest problems in children and in lung
p = 0.05 level was declared when the 95% smokers. In both sexes smokers tended to cancer. Both groups were less aware of the
CI for the difference in responses between be younger than nonsmokers. Although causal role of ETS in heart attacks and ear
smokers and nonsmokers excluded zero the majority of both groups were married problems in children. This is not surpris-
(denoted by * in the tables). Non-overlapping or living with partners, smokers were more ing because strong evidence of the causal
95% CIs for smokers versus nonsmokers is likely to be divorced or separated. They role of ETS in these conditions has
approximately equivalent to statistically also tended to report lower levels of formal emerged only recently.25 More than three
significant differences at the p = 0.005 education. quarters of each group thought that all or
level. Multiple logistic regression was used most daily smokers are addicted, and clear
to examine the independent relationships Knowledge of health effects and impact of majorities also strongly agreed that quit-
of knowledge about the health effects of smoking in Canada ting smoking can improve health even after
smoking and ETS exposure, as well as Respondents were asked whether smok- smoking a lot for a long time. However,
smoking status, to attitudes, controlling for ing or exposure to ETS is a cause, may be a knowledge that any amount of smoking
sociodemographic characteristics. 24 cause, or is not a cause of specific health represents a serious risk was lower, espe-

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 377

DIFFERENCES BETWEEN SMOKERS AND NONSMOKERS

TABLE II ers would ignore the rules. When smokers

Attitudes Toward Restrictions on Smoking in Specific Settings† were asked to predict their own compliance
with more restrictions on smoking, a large
Nonsmokers (n=1340) Smokers (n=424) majority (78%) predicted that they, them-
Percent 95% CI Percent 95% CI
selves, would go along with the rules.
Smoking should not be permitted at all in
Family fast food restaurants* 71.0 67.5, 74.5 47.0 40.3, 53.7
Indoor public gatherings* 62.4 58.7, 66.1 39.9 33.4, 46.4 Perceptions of the effectiveness of taxes
Food courts in malls* 58.6 54.8, 62.3 31.6 25.4, 37.8 on tobacco and attitudes toward tobacco
Hockey arenas* 58.4 54.6, 62.1 41.5 34.9, 48.1
Workplaces* 45.5 41.8, 49.3 22.9 17.4, 28.5 policy measures
Restaurants* 44.5 40.8, 48.3 13.6 8.7, 18.4 Sizable proportions of both groups, but
Bingo halls* 36.1 32.6, 39.7 16.7 11.7, 21.7
Bars and taverns* 17.9 15.1, 20.7 3.0 1.0, 5.0 especially smokers, failed to recognize the
† listed in descending order of support among nonsmokers for not permitting smoking at all
effectiveness of tax measures in reducing
* indicates that the difference between nonsmokers and smokers is statistically significant at the smoking among both children and adults
p < 0.05 level. (Table IV). Smokers also differed from
nonsmokers in their support for restric-
TABLE III tions on the sale of tobacco products. In
Predicted Compliance with More Restrictions both groups, support was strongest for
Nonsmokers (n=1340) Smokers (n=424)
banning sales in drug stores. A majority of
Percent 95% CI Percent 95% CI nonsmokers and about one in four smok-
“If there were more restrictions on smoking MOST SMOKERS WOULD”
ers favoured restricting the sale of cigarettes
Go along with the rules* 36.8 33.2, 40.4 50.0 43.2, 56.7 to special stores, as is done with alcohol.
Go along only if there is a big fine 31.7 28.1, 35.2 26.7 20.6, 32.9 Although nonsmokers were also more
Ignore restrictions* 26.9 23.6, 30.3 19.4 13.9, 25.0
Don’t know/refused to answer 4.6 3.1, 6.1 3.9 1.5, 6.3 likely than smokers to support other tobac-
“If there were more restrictions on smoking, I WOULD” (smokers only)
co policy measures, clear majorities of both
Go along with the rules 77.8 71.9, 83.7 groups agreed that stores convicted of sell-
Go along only if there is a big fine 10.6 6.3, 15.0 ing tobacco to minors should be prohibit-
Ignore restrictions 9.4 5.3, 13.6
Don’t know/refused to answer 2.1 0.0, 4.4 ed from selling tobacco, and majorities
* indicates that the difference between nonsmokers and smokers is statistically significant at the
agreed that cigarette packages should
p < 0.05 level. include an insert fully describing the health
hazards of smoking and tips on how to
cially among smokers. Most respondents and indoor public gatherings. Both groups quit. Clear majorities of nonsmokers also
in both groups were not aware that tobac- were least supportive of bans in bars and supported the plain packaging of cigarettes
co causes a lot more deaths than alcohol or taverns. and bans on advertising. Even among
AIDS. When nonsmokers and smokers were smokers, there was substantial minority
compared regarding support for at least support for bans on advertising. However,
Attitudes toward restrictions on smoking some restriction on smoking (either a com- consistent with these findings, smokers
in specific settings plete ban or restriction to certain areas) in were more likely than nonsmokers to agree
For each of eight settings, respondents the eight settings, most of the intergroup that tobacco companies should be allowed
were asked whether smoking should not be differences disappeared. Clear majorities of to sponsor sporting and cultural events.
permitted at all, smoking should be per- both groups favoured at least some degree
mitted in restricted areas, or smoking of restriction on smoking in seven of the Relationships of smoking status and
should not be restricted at all. For all set- eight locations. Only for bars and taverns knowledge to attitudes
tings, nonsmokers were significantly more did nonsmokers and smokers differ sub- Multiple logistic regression analysis that
likely than smokers to support bans on stantially, with 73% and 45%, respective- controlled for age, sex, marital status and
smoking (Table II); a clear majority sup- ly, supporting some degree of restriction. educational attainment showed that both
ported a ban in family fast food restau- smoking status and knowledge were inde-
rants, while majorities supported bans at Predicted compliance with more restric- pendently associated with support for
indoor public gatherings, in food courts in tions restrictions and other tobacco control policy
malls, and in hockey arenas, and substan- Respondents were asked to predict what measures. Nonsmokers were more support-
tial minorities supported bans in work- most smokers would do if there were more ive of these measures, as were respondents
places and restaurants. In contrast, bans on restrictions on smoking (Table III). who were more informed about either the
smoking were not supported by a majority Smokers were more likely than nonsmok- health risks of smoking or exposure to ETS.
of smokers for any setting. The largest ers to predict that most smokers would go However, the strength of the relationship
minorities of smokers favoured bans in along with rules (50% versus 37%, respec- between knowledge and support was similar
family fast food restaurants, hockey arenas, tively), and less likely to predict that smok- in smokers and in nonsmokers.

378 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5

 DIFFERENCES BETWEEN SMOKERS AND NONSMOKERS

DISCUSSION TABLE IV
Perceptions of Effectiveness of Taxes on Tobacco and
These findings have implications for Attitudes Toward Tobacco Policy Measures
tobacco control programs and policy. First,
deficits in knowledge, particularly with Nonsmokers (n=1340) Smokers (n=424)
Percent 95% CI Percent 95% CI
regard to ETS, risks of smoking even small Strongly agree/agree
amounts, and public health impact, were Higher taxes on tobacco will
help prevent children from
found in both groups, but especially in becoming smokers* 54.9 51.3, 58.4 40.5 34.3, 46.8
smokers. Explanations for previous find- Higher taxes on tobacco would
help people quit smoking* 51.6 48.0, 55.1 27.2 21.6, 32.9
ings of accentuated knowledge deficits in
smokers have been offered.26 The findings Tobacco products should not be sold in:
Drug stores* 73.8 70.8, 76.8 56.0 49.7, 62.2
imply that while effective educational pro- Grocery stores* 50.0 46.4, 53.5 22.9 17.5, 28.2
grams aimed at the entire population are Variety stores* 23.7 20.7, 26.6 4.6 2.2, 7.0
needed, specific efforts must be directed at Cigarettes should be sold only
smokers. Appropriately designed health in special stores, like alcohol:* 54.3 50.8, 57.8 24.1 18.9, 29.4
warnings on cigarette packages 27,28 and Strongly agree/agree
package inserts are potential vehicles. The Stores convicted of selling
tobacco to young people
latter measure was supported by a majority < 19 should lose licence
of smokers in this survey. In designing to sell tobacco* 90.3 88.2, 92.4 74.6 69.2, 80.1
Cigarette packages should
educational strategies, the lower education- include insert describing
al attainment of smokers compared to health hazards and tips
on quitting* 83.7 81.2, 86.3 59.5 53.3, 65.7
nonsmokers must be taken into account. Cigarettes should be sold in
Using data from the 1994-95 National plain white packages to
discourage smoking
Population Health Survey, Miller29 showed by children* 71.5 68.4, 74.6 44.5 38.2, 50.7
that while all smokers cited the mass media All advertising about tobacco
products should be
as their major source of information about forbidden by law* 68.4 65.0, 71.7 46.9 40.6, 53.2
smoking, those with lower education Tobacco companies should
be allowed to sponsor
reported the mass media less often than sporting and cultural events* 51.1 47.6, 54.7 72.8 67.2, 78.5
did smokers with higher education.
* indicates that the difference between nonsmokers and smokers is statistically significant at the
Furthermore, they were less likely to p < 0.05 level.
obtain information from books, pamphlets
or magazines and less likely to recall print- and 1996, there was some increase in sup- to minors is likewise encouraging.
ed warnings about heart disease on ciga- port for bans on smoking in workplaces However, the relative failure of both
rette packages. These findings must be and restaurants, however in 1996, such groups to recognize the effectiveness of tax
considered in selecting channels for educa- support still fell short of majorities among measures in reducing smoking among both
tion and designing materials. The fact that both groups for both settings. The recent children and adults is consistent with the
knowledge was found to be independently findings do suggest that family fast food findings of earlier studies in Ontario.9,10
associated with supportive attitudes toward restaurants and hockey arenas should be This should be a matter of concern. Many
tobacco control suggests that educational priority settings for the implementation of studies have demonstrated the effectiveness
interventions will not only help to inform complete bans. They further suggest that of tax policies as part of a comprehensive
the public, they may increase support for for settings where there is only weak sup- tobacco control strategy (e.g. refs. 32-35).
other interventions. port for complete bans, a requirement for Specific interventions to increase under-
Second, clear majorities of both groups appropriate restrictions, in the form of standing about the effectiveness of tobacco
supported some degree of restriction on enclosed, separately ventilated areas, 30 taxes and support for tax measures are
smoking in specific settings; they differed, should be considered as an interim step. It needed.
however, in their support for complete is most encouraging that almost 80% of Certain limitations are inherent in these
bans. In a 1991 survey, it had been found smokers indicated that they would go findings. The growing social unacceptabili-
that 50% or more of both smokers and along with more restrictions on smoking, ty of smoking may have biased respon-
nonsmokers supported complete bans in even without the threat of a fine. dents, particularly smokers, to indicate
city buses, doctors’ offices, day-care cen- It is also encouraging that majorities of stronger support for various control mea-
tres, stores, schools, banks, movie theatres, both groups supported banning cigarette sures than was really the case, thus dimin-
airplanes, and hospitals.10 Therefore, these sales in drug stores, a prohibition that ishing the actual extent of differences
locations were not reassessed. Only restau- came into effect in Ontario some months between smokers and nonsmokers.
rants, workplaces and indoor public gath- before the survey.31 The widespread sup- Regarding knowledge, however, it is reas-
erings were re-examined. Between 1991 port for punishing stores that sell tobacco suring that only 6% and 5% of nonsmok-

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 379

DIFFERENCES BETWEEN SMOKERS AND NONSMOKERS
ers and smokers indicated a causal relation-
ship between smoking and arthritis. Such a
relationship has not been established. As
well, the comprehensiveness of the topics
addressed prohibited in-depth probing in
the telephone interview of perceptions and
experiences that may underpin attitudes
toward tobacco control. Nonetheless, the
findings provide guidance for the tobacco
control agenda.
ACKNOWLEDGEMENTS
This research was carried out under the
auspices of the Ontario Tobacco Research
Unit, Centre for Health Promotion,
University of Toronto, with support from
Health Canada (NHRDP). Dr. Bull is a
National Health Research Scholar. The
Institute for Social Research, York
University conducted the interviews and
prepared the data for analysis. David
Northrup provided advice on the question-
naire and other aspects of the survey.
REFERENCES
1. Peto R, Lopez AD, Boreham J, et al. Mortality
from Smoking in Developed Countries 1950-2000:
Indirect Estimates from National Vital Statistics.
Oxford: Oxford University Press, 1994.
2. Makomaski Illing EM, Kaiserman MJ. Mortality
attributable to tobacco use in Canada and its
regions, 1991. Can J Public Health 1995;86:257-65.
3. Single E, Robson L, Xie XD, Rehm J. The eco-
nomic costs of alcohol, tobacco, and illicit drugs
in Canada, 1992. Addiction 1998;93:991-1006.
4. Kaiserman MJ. The cost of smoking in Canada,
1991. Chron Dis Can 1997;18:13-19.
5. Health Canada, Statistics Canada, and the
Canadian Institute for Health Information.
Statistical Report on the Health of Canadians.
Ottawa: 1999.
6. Health Canada. Directions. The Directional
Paper of the National Strategy to Reduce
Tobacco Use. An Update: 1993. Ottawa: Health
Canada, 1993.
380 REVUE CANADIENNE DE SANTÉ PUBLIQUE
7. Office of Tobacco Control. Smoking By-Laws in
Canada 1995. Health Canada, Health Protection
Branch, Ottawa: Minister of Health Canada, 1995.
8. National Clearinghouse on Tobacco and Health.
Federal and Provincial Legislation in Canada: An
Overview. Ottawa, March 1995.
9. Pederson LL, Bull SB, Ashley MJ, Lefcoe NM. A
population survey in Ontario regarding restrictive
measures on smoking: Relationship of smoking
status to knowledge, attitudes, and predicted
behaviour. Int J Epidemiol 1987;16:383-91.
10. Ashley MJ, Bull SB, Pederson LL. Support
among smokers and nonsmokers for restrictions
on smoking. Am J Prev Med 1995;11:283-87.
11. Martin G, Steyn K, Yach D. Beliefs about smok-
ing and health attitudes toward tobacco control
measures. S Afr Med J 1992;82:241-45.
12. Velicer WF, Laforge RG, Levesque DA, Fava JL.
The development and initial validation of the
smoking policy inventory. Tobacco Control
1994;3:347-55.
13. Laforge RG, Velicer WF, Levesque DA, et al.
Measuring support for tobacco control policy in
selected areas of six countries. Tobacco Control
1998;7:241-46.
14. Samuels B, Glantz SA. The politics of local
tobacco control. JAMA 1991;266:2110-17.
15. Traynor MP, Begay ME, Glantz SA. New tobac-
co industry strategy to prevent local tobacco con-
trol. JAMA 1993;270:479-86.
16. Ashley MJ, Pederson L, Poland B, et al.
Smoking, Smoking Cessation, Tobacco Control
and Programming: A Qualitative and
Quantitative Study. Final Report NHRDP
Project No. 6606-6006-801. Submitted to
Health Canada. March 31, 1997.
17. Northrup DA, Rhyne D. Smoking, Smoking
Cessation, Tobacco Control and Programming:
A Qualitative and Quantitative Study. Technical
Documentation. North York: Institute for Social
Research, York University, 1996.
18. O’Rourke D, Blair J. Improving random respon-
dent selection in telephone surveys. J Marketing
Research 1983;20:428-32.
19. Groves RM, Lyberg LE. An overview of nonre-
sponse issues in telephone surveys. In: Groves
RM, Biemer PP, Lyberg LE, et al. (Eds.),
Telephone Survey Methodology. New York: John
Wiley and Sons, 1988;191-211.
20. Bondy S, Ialomiteanu A. Smoking in Ontario, 1991
to 1996. Can J Public Health 1997;88:225-29.
21. Cox BG, Cohen SB. Methodological Issues in
Health Care Surveys. New York: Marcel Dekker,
Inc., 1985.
22. Goodman LA. On simultaneous confidence
intervals for multinomial proportions.
Technometrics 1965;7:247-54.
23. Fleiss JL. Statistical Methods for Rates and
Proportions Second Edition. New York: John
Wiley and Sons, 1981.
24. Hosmer DWJr, Lemeshow S. Applied Logistic
Regression. New York: John Wiley and Sons,
1989.
25. U.S. Department of Health and Human Services.
Reducing the Health Consequences of Smoking.
25 Years of Progress. A report of the Surgeon
General. 1989. Chapter 4. Trends in public
beliefs, attitudes, and opinions about smoking.
Public Health Service, Centers for Disease
Control. Office of Smoking and Health.
Rockville, Maryland. pages 171-258.
26. Law MR, Morris JK, Wald NJ. Environmental
tobacco smoke exposure and ischaemic heart dis-
ease: An evaluation of the evidence. Br Med J
1997;315:973-80.
27. Borland R, Hill D. Initial impact of the new
Australian tobacco health warnings on knowledge
and beliefs. Tobacco Control 1997;6:317-25.
28. Borland R. Tobacco health warnings and smoking-
related cognitions and behaviours. Addiction
1997;92:1427-35.
29. Millar WJ. Reaching smokers with lower educa-
tional attainment. Health Reports 1996;8:11-19.
30. Repace JL. Risk management of passive smoking
at work and at home. Saint Louis Univ Public
Health Law Rev 1994;13:763-85.
31. Province of Ontario. Bill 119, 35th Legislature.
Chapter 10, Statutes of Ontario, 1994. An Act to
Prevent the Provision of Tobacco to Young
Persons and to Regulate its Sale and Use by
Others. Toronto: Legislative Assembly of
Ontario, 1994.
32. U.S. Department of Health and Human Services.
Preventing Tobacco Use Among Young People.
A Report of the Surgeon General. Atlanta, GA:
Public Health Service. Centers for Disease
Control and Prevention. Office on Smoking and
Health, 1994.
33. Lewit EM, Hyland A, Keerebrock N, Cummings
KM. Price, public policy, and smoking in young
people. Tobacco Control 1997;6 (Suppl. 2):S17-
S24.
34. Koh HK. An analysis of the successful 1992
Massachusetts tobacco tax initiative. Tobacco
Control 1996;5:220-25.
35. Chaloupka FJ, Wechsler H. Price, tobacco con-
trol policies and smoking among young adults.
J Health Economics 1997;16:359-73.
Received: November 15, 1999
Accepted: January 12, 2000
VOLUME 91, NO. 5
A B S T R A C T
Purpose: A national survey of adult Canadians
(n=1,544) was recently undertaken (1997-
1998) to monitor whether changes in dietary
intake have occurred since the last Canadian
dietary survey, conducted a generation ago
(1970).
Methods: Individuals from randomly selected
households from a stratified sample of 80 enu-
meration areas from five regions of Canada were
interviewed by a dietitian at home for a 24-hour
dietary recall and food frequency questionnaire.
Results: The mean dietary % energy from
protein (16-18%), carbohydrate (50-56%) and
fat (29-31%) was close to recommended levels
in the different age-sex groups. Fat intake was
reduced from previous surveys. Intakes of dairy
products and fruits and vegetables are lower
than recommended. Nutrient values at the 25th
percentile of the nutrient distribution, were
below recommended levels for calcium, folate,
iron and zinc among women.
Conclusion: Despite the growing problem of
obesity, Canadians are eating less fat than a gen-
eration ago but intake of certain nutrients are
still suboptimal.
A B R É G É
Objectif : Une étude nationale sur des
Canadiens adultes (n=1 544) fut récemment
entreprise (1997-1998) pour vérifier si des
changements dans la prise alimentaire étaient
survenus depuis la dernière étude il y a une
génération (1970).
Méthodes : Des individus furent interviewés à
la maison pour répondre à un questionnaire sur
le rappel nutritionnel de 24 heures et la
fréquence de prise alimentaire parmi des foyers
sélectionnés au hasard à partir d’un échantillon
stratifié de 80 territoires énumérés venant de
cinq régions du Canada.
Résultats : Le pourcentage moyen d’énergie
provenant de protéines (16-18 %), des hydrates
de carbone (50-56 %) et du gras (29-31 %) était
près des niveaux recommandés. La prise de
matières grasses avait diminué depuis les études
précédentes. La prise de produits laitiers et de
fruits et légumes est au-dessous de la recomman-
dation. Les valeurs nutritives au 25e percentile
de la distribution nutritionnelle, étaient au
dessous des niveaux recommandés pour le calci-
um, l’acide folique, le fer et le zinc chez les
femmes.
Conclusion : Malgré le problème croissant
d’obésité, les Canadiens consomment moins de
gras qu’il y a une génération mais la prise de cer-
tains nutriments est toujours sous-optimal.
SEPTEMBER – OCTOBER 2000
Food Habits of Canadians:
Reduction in Fat Intake Over a
Generation
Katherine Gray-Donald, PhD, Linda Jacobs-Starkey, PhD,
The role of diet in the prevention of
chronic disease is well established: fruit and
vegetable consumption has a strong protec-
tive effect on the development of cancer at
numerous sites; 1 saturated fat intake is
clearly associated with coronary heart dis-
ease;2 and the total direct cost of obesity in
Canada was estimated to be $1.8 billion
for 1997.3
In the U.S., the National Health and
Nutrition Examination Surveys and the
Continuing Survey of Food Intake of
Individuals4 provide useful data on nutri-
tional intake and food trends.4,5 These data
show a downward trend in energy intake
levelling off in the 1990s, but total fat con-
sumption continuing to decline from 42%
of energy intake in 1965 to 33% in 1995.2
The most recent national data in Canada
derive from the Nutrition Canada survey
(1970-72) conducted a generation ago.6,7
Four provincial surveys were completed
more recently. The Ontario 8,9 and
Manitoba10 surveys used semi-quantitative
food frequency questionnaires while the
Nova Scotia11 and Santé Québec surveys12
used 24-hour recalls which provide quanti-
tative measures of dietary intake. Total fat
intakes have declined but remain above the
recommended 30% of energy and low
intakes of calcium, iron and folate were
identified for specific age-sex groups.11,12
National data estimated from Canada’s
family food expenditure survey also indi-
School of Dietetics and Human Nutrition, McGill
University, Montreal, QC
Correspondence and reprint requests: Dr. Katherine
Gray-Donald, Director, School of Dietetics and Human
Nutrition, McGill University, Macdonald Campus, 21,
#111 Lakeshore Road, Ste. Anne de Bellevue, QC,
H9X 3V9, Tel: 514-398-7842, Fax: 514-398-7739,
E-mail: gray-donald@macdonald.mcgill.ca
Financial support by the Beef Information Centre
with funds obtained from the Beef Industry
Development Fund is acknowledged.
CANADIAN JOURNAL OF PUBLIC HEALTH 381
Louise Johnson-Down, MSc
cate declines in total fat content of foods
purchased.13
Our survey “Food Habits of Canadians”
provided data between August 1997 and
July 1998 in five regions of Canada
(Atlantic, Québec, Ontario, Prairies and
British Columbia) on current food and
nutrient intake. This report focusses on
current nutrient intake and compares these
data to earlier studies.
METHODS
The sampling of respondents (18-65
years and adolescents 13-17 years) was
done using a multi-stage random sample of
adult Canadians living in five regions of
Canada: Atlantic (New Brunswick, Nova
Scotia, Prince Edward Island,
Newfoundland), Québec, Ontario, Prairies
(Manitoba, Saskatchewan, Alberta), and
British Columbia. Fifteen percent of the
Canadian population who lived in regions
far from major population centres were not
sampled. Aboriginal communities were not
included. In each region, four Canada cen-
sus divisions were randomly chosen with a
probability proportional to the population.
For each of the 20 divisions, a random
sample of two subdivisions was similarly
selected, and for each subdivision, two
enumeration areas were selected, resulting
in 80 enumeration areas across the coun-
try. Boundaries of enumeration areas were
identified using Statistics Canada maps
and address ranges within an enumeration
area. The sampling for individual random
households from within each enumeration
area was done using the 1996 computer-
ized telephone directory (Pro CD Inc.,
Mass.). Each household received a person-
alized letter to explain the study prior to
telephone contact. Interviewers attempted
CANADIAN DIETARY SURVEY

TABLE I
Weighted Mean Nutrient Intake of Canadian Men and Women Aged 18-65
18-34 Years 35-49 Years 50-65 Years Recommendations*
M F M F M F
(n=125) (n=207) (n=266) (n=459) (n=181) (n=306)

Energy (MJ) 13.75 8.35 11.27 7.48 10.20 7.34 -

Energy (kcal) 3290 1998 2696 1789 2440 1756 -
Protein (% energy) 16.5 15.8 17.1 17.1 17.1 17.7 -
Carbohydrate (% energy) 53.1 55.9 51.6 53.0 50.1 54.0 55
Total fat (% energy) 29.1 28.8 30.2 29.7 30.6 28.9 30
Saturated fat (% energy) 10.2 9.5 9.6 9.7 9.8 9.5 10
Polyunsaturated fat (% energy) 4.5 5.0 5.1 5.2 5.3 5.0 -
Monounsaturated fat (% energy) 11.1 10.6 11.6 11.5 11.7 10.7 -
Cholesterol (mg) 395 238 340 224 343 230 <300
Total fibre (g) 21 14 18 14 16 16 -
Vitamin A (RE) 1988 1362 2014 1782 2050 1846 800-1000
Folate (µg) 375 275 311 238 308 241 180-230
B12 (µg) 5.8 5.0 8.3 5.3 7.0 5.0 1.0
Vitamin C (mg) 242 144 154 132 137 143 30-40
Calcium (mg) 1376 813 1020 764 901 777 700-900
Iron (mg) 24 14 19 13 18 14 8-13†
Zinc (mg) 18 11 15 10 14 10 9-12
* Nutrition Recommendations 1990
† For women aged 19-49, the recommendation is 13 mg.

to enrol one adult per household (the adult TABLE II

with the next birthday) for a total of 20 Mean Nutrient Intake of Canadian Adolescents Aged 13-17
adult respondents from each enumeration Compared to Recommended Levels (RNI*)
area.
Nutrients Males Recommendations Females Recommendations
Appointments for interviews were made n = 84 n = 94
on different days of the week including
Energy (MJ) 13.4 11.7-13.4 9.2 8.78-9.20
weekends. Interviews were held in the Energy (kcal) 3206 2800-3200 2201 2100-2200
respondents’ homes or at other convenient Protein (% energy) 14.6 - 14.4 -
Carbohydrate (% energy) 55.9 55 60.3 55
locations. A repeat 24-hour recall was con- Fat (% energy) 30.8 30 27.2 30
ducted on 30% of the adult sample within Saturated fat (% energy) 10.2 10 9.1 10
Monounsaturated fat (% energy) 11.9 - 9.9 -
approximately one week of the initial inter- Polyunsaturated fat (% energy) 5.2 - 4.6 -
view in order to estimate intra-individual Cholesterol (mg) 303 - 196 -
Total fiber (g) 17.9 - 14.5 -
variability. Vitamin A (RE) 1888 900-1000 1434 800
Each adult participant was asked Folate (µg) 299 175-220 274 170-190
B12 (µg) 6.1 1.0 5.0 1.0
whether there was a potential adolescent Vitamin C (mg) 173 30-40 214 30
(13-17 years) participant living in the Calcium (mg) 1407 900-1100 1004 700-1000
Iron (mg) 22.2 10 15.1 12-13
household. The adolescent sample was not Zinc (mg) 15.8 12 9.8 9
proportional to the population and is not
* Nutrition Recommendations, 1990.
independent of the adult sample.
language of interview, civil status, birth when possible or from the American data
Measurements date, educational level, smoking status, base.17 Food portion sizes from the four
Dietary intake was measured using the number of adults and children in the food groups of Canada’s Food Guide to
24-hour recall, commonly used for nation- household and perceived health status as Healthy Eating were calculated based on
al surveys.4,5,11,12 A single 24-hour recall is well as reported height and weight. categories of foods and standard weights
an appropriate method to assess the aver- Average income per enumeration area was assigned to those categories.18
age intake of a large group of individuals obtained from the 1991 census.16
provided all days of the week and seasons Double verification of all 24-hour recalls ANALYSIS
are covered.14,15 Interviews were conducted was done and data were entered into the
in French or English by professional dieti- nutrient analysis program Candat (Godin In order to eliminate potential bias
tians who received a two-day training ses- London Inc., London, Ontario, 1991). resulting from the sampling strategy, indi-
sion in Montreal. Food portion models The most recent Canadian Nutrient File vidual results were multiplied by weights
(graduated cup, two bowls, a plate, spoons (Health Canada, Ontario, 1997) was used calculated using 1991 Canadian census
and a ruler) were used to quantify intake. and a total of 267 food items were added data of the total population and the num-
A sociodemographic questionnaire was to the data set using nutrient information ber of households in each region. The
used, including data on country of origin, obtained from food manufacturers’ data weighting process involved three compo-

382 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5

 CANADIAN DIETARY SURVEY

nents: a ratio of populations esti- and 30.5% in the National Population

mating how many people each
respondent represented within
their region; the number of house-
holds in a given region because we
sampled by household not individ-
uals; and the number of adults in
each household responding to the
Distribution of Intakes for Selected Nutrients for Adults 18-65 Years in the Food Habits of Canadians Survey
Health Survey.21
Nutrient intake data for the country,
weighted for the sampling strategy, are pre-
sented separately for males and females in
three age groups (Table I).22 A comparison
of reported energy intake compared to
basal metabolic requirements yielded ratios

survey. of 1.3 to 1.7 for men and 1.2 to 1.4 for

7.87
1883
79.9
260
60.5
32.8
20.3
9.80
208
907
280
14.7
10.9
25th 50th 75th

In order to provide the 25th and women indicating some under-reporting

75th percentiles of the distributions for women. 22,23 The mean values for all
(n=306)
50-65

of nutrients, repeat measures were nutrients listed for men and women are
6.90
1651
64.7
217
49.9
28.6
15.9
9.53
160
679
213
12.2
8.6 used to calculate the ratio of above the recommended intakes for
within-person to between-person Canadians22 with the exception of calcium
5.91
1414
52.7
179
39.1
24.6
12.0
9.21
115
492
163
10.1
7.0

variability for each nutrient for in some age/sex groups. The mean intake
men and women separately. Log of calcium for most groups, however, is
8.53
2041
82.1
272
66.9
33.7
21.6
9.83
226
934
268
14.4
11.2
25th 50th 75th

and square root transformations below the new Dietary Reference Intake (a
were used to normalize the data. new common standard for Canadians and
Females

(n=459)
35-49

7.20
1722
66.2
224
54.0
29.7
17.2
9.54
167
679
201
12.2
8.8

The nutrient distribution for usual Americans).24 The percent of energy from
intake of these nutrients was mod- fat (29-31%) was close to the recommend-
elled using the method described ed value of 30% and saturated fat repre-
5.98
1431
54.4
178
42.1
25.2
12.7
9.26
115
477
156
10.0
7.0

by the National Research sented approximately 10% of energy

Council.19 The basal metabolic rate intake. Mean intakes of adolescents (Table
8.69
2079
81.6
300
66.6
33.0
23.0
9.90
216
980
271
15.9
11.2
25th 50th 75th

(BMR) for each subject was calcu- II) indicated higher energy intakes than for
lated.20 adults as expected, however, fat intakes as a
(n=206)
18-34

7.34
1756
65.2
244
54.0
29.0
17.5
9.56
154
664
200
12.2
8.7

percent of energy were similar to those of

TABLE III

RESULTS adult participants. Mean micronutrient

6.08
1455
55.2
189
42.2
24.7
12.5
9.22
113
473
154
9.3
6.8

levels met intake recommendations except

The study sample consisted of for vitamin A in adolescent girls.
Data are adjusted for within-subject variability by the National Research Council Method.20

1,544 adults and 178 adolescents Table III provides the 25th, 50th and
9.27 11.03
2218 2639
113
357
74.1 95.1
30.1 34.6
23.2 31.3
9.75 10.1
347
803 1070
327
16.0 20.1
11.4 15.3
25th 50th 75th

from 80 enumeration areas across 75th percentiles of intake for each adult
(n=181)
50-65

Canada. These ranged from a fish- age-sex group adjusted for within-person
90.7
246

224

246

ing village in Newfoundland to a variability. (The sample size for the adoles-
suburb of Victoria, British cent population was not sufficiently large
7.44
1780
71.6
173
54.6
26.3
15.1
9.41
151
570
195
12.8
9.0

Columbia. The response rate to do this adjustment.) The median intake

(interviewed/interviewed and for the percentage of energy from total fat
8.34 10.37 12.83
25th 50th 75th

1995 2481 3069

126
458
106
25.9 30.1 34.3
17.6 24.7 34.8
9.46 9.75 10.1
347
821 1208
365
13.2 17.1 21.3
10.0 12.9 16.8

refusals) was calculated for each was 29-30% in all age-sex groups while the
enumeration area. The average rate 75 th percentile was 33-35%. Calcium
(n=266)
35-49
Males

was 30%. In three enumeration intake among women at the 25th percentile
78.4 97.9
326
61.4 79.5

250

276

areas in inner city Montreal and of intake was under 500 mg for all age
Vancouver, we were unable to groups indicating very low intakes in many
211

163
621
197

obtain more than 10 subjects women. Similarly iron intakes in women

because of the very low response of reproductive age were low in a substan-
25th 50th 75th

9.72 11.50 14.44

2325 2751 3455
137
496
117
25.0 29.3 34.5
20.7 30.4 40.2
9.42 9.80 10.2
417
739 1018 1490
382
14.7 19.0 23.5
11.2 14.0 18.0

rate. More women than men tial portion of the population.

accepted to be interviewed (972 vs. A comparison with the Nutrition
(n=125)
18-34

105
365
63.3 86.7

292

274

572 respectively). Our sample was
similar to the Canadian population
(1991 Census) in number of peo-
84.6
285
Canada Survey indicates major changes in
the intake of a number of nutrients over a
generation (Table IV). Mean nutrient

189

213

ple born in Canada (86% vs. intakes, using the two age groups originally
84%), number of subjects with less reported on in the Nutrition Canada
% Energy Sat. Fat
Carbohydrate (g)

Cholesterol (mg)
Saturated Fat (g)

than high school education (22% Survey of 1970, 6,7 are compared to our
% Energy Fat
Energy (kcal)

vs. 26%), and single marital status data (1997-1998). Mean energy intakes
Calcium (g)
Energy (MJ)

Folate (µg)
Protein (g)

Zinc (mg)
Iron (mg)

was 26% vs. 32% respectively. The were lower in our survey in most age-sex
Fat (g)

percentage of adults reporting a groups. Total fat intake, however, declined

BMI of >27 was 32% in our survey far more over the last 27 years, as did the

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 383

CANADIAN DIETARY SURVEY

intake of saturated fats. Micronutrient TABLE IV

density, however, has improved. Absolute Comparison of Food Habits Data 1997-98 to Nutrition Canada 1970
intakes of most vitamins and minerals have
increased. 20-39 Years 40-64 Years
Males Females Males Females
The mean number of servings of the Year 1970 1997-98 1970 1997-98 1970 1997-98 1970 1997-98
four food groups – grain products, vegeta-
Sample Size 999 203 1347 343 1222 348 1500 608
bles and fruit, milk products, and meat Energy (MJ) 14.12 12.21 8.37 7.82 11.18 10.31 7.22 7.32
and alternatives – for adults are shown in Energy (kcal) 3378 2921 2002 1871 2675 2467 1727 1751
Protein (g) 119 123 72 75 94 104 63 75
Table V. The consumption of vegetables Fat (g) 154 98 89 63 118 84 75 58
and fruits were low, particularly among % energy from fat 41.0 29.4 40.0 29.2 39.7 30.3 39.1 29.3
Calcium (mg) 1081 1177 709 781 883 896 613 745
women and among men aged 50+ years. Iron (mg) 18 21 12 14 16 17 11 13
Similarly the consumption of dairy prod- Thiamin (mg) 1.57 2.57 1.02 1.58 1.32 2.23 0.90 1.70
Riboflavin (mg) 2.59 2.70 1.70 1.68 2.09 2.27 1.49 1.67
ucts is below the suggested intake of two Niacin (NE) 48 52 28 33 37 46 25 32
servings a day among women and men 35 Vitamin C (mg) 118 204 89 143 101 134 106 132
Folate (µg) 221 322 146 239 183 301 148 241
years and older. Foods considered as extras Vitamin A (RE) 1551 2018 1292 1462 1332 1854 1031 1832
and not part of the four food groups of
the Food Guide 18 provided 26-29% of TABLE V
energy and 24-34% of fat intake in the Average Number of Servings for Each Food Group According to “Canada’s
different age-sex groups, but only very Food Guide to Healthy Eating” by Gender and Age on the Day of their Recall
small amounts of protein and micronutri-
ents. Food Group Recommended Males Females
Portions
18-34 35-49 50-65 Mean 18-34 35-49 50-65 Mean
DISCUSSION Grain Products 5-12 8.0 6.8 6.4 6.9 5.2 5.0 4.5 4.9
Vegetables & Fruit 5-10 5.8 5.3 4.6 5.2 4.1 4.5 5.0 4.6
This first national dietary survey since Milk Products 2-4 2.5 1.7 1.5 1.8 1.6 1.5 1.3 1.4
Meat & Alternatives 2-3 3.9 3.3 3.0 3.4 2.0 2.1 2.0 2.0
1970 indicates important reductions in
dietary fat intake and substantially higher In terms of dietary changes for disease be accurate according to physiological
intakes of a number of important micro- prevention, the decreases in saturated fat needs for age, height and weight, but
nutrients. Significant numbers of intake may explain a part of the observed intakes of women indicate some under-
Canadians, particularly women, still con- decreases in coronary heart disease in reporting.23 The response rate was low, in
sume inadequate intakes of calcium, folate Canada. The decline in fat intake, how- part due to out-of-date telephone listings
and iron during the reproductive years. ever, is not related to the decreases in obe- in areas of high mobility and the need to
This same trend was seen in a comparison sity. In fact, Willett suggests that dietary compete with telephone soliciting. The
of the diet of Quebecers between 1971 and fat is not a major determinant of body Nutrition Canada Survey had a response
1990 where the authors conclude, “Les fat.27 The vegetable and fruit consumption rate of 46%, the U.S. Continuing Food
Québécoises et les Québécois mangent is difficult to compare to Nutrition Survey 1987, 35%. The particular diffi-
donc mieux, mais pas encore bien...” Canada because of changing definitions of culty of recruiting participants in very low
(Quebecers eat better but still not well).12 portion sizes over time, but the intake of income housing settings, means that the
Although there are a number of indicators nutrients that are usually found in fruits poorest Canadians in large cities may not
of an improvement in diet quality, it is and vegetables – such as vitamin C and be well represented in this survey. In
very evident from statistics on body weight folate – has increased. addition, high-risk groups with well-
and height for the population that there is As with any dietary survey, there are known nutritional problems – such as
a serious problem of obesity in Canada,3 important limitations in terms of the children living in poverty 30,31 and frail
and increasing rates of obesity are reported accuracy of the reporting of dietary data elderly32 – were not included in the sam-
in Europe25 and the United States.26 The and the response rate. The accuracy of the ple.
total energy that we observed was less than reported intake cannot be ascertained in a In summary, the diets of Canadians
that reported in 1971, most particularly survey, but 24-hour recall was observed to appear to be improving in general, with
among young men. This trend is similar to compare closely the energy expenditure in important decreases in fat intake and high-
that observed in Quebec, 12 where young women.28 There is the concern for er intakes of some micronutrients. Given
Nutrition Canada Data for Quebec alone underreporting – particularly in dietary the very important role of diet in the pre-
were compared to Santé Québec data of survey – and the trend for underreporting vention of cardiovascular disease, cancer
1990. Energy expenditure is widely may be increasing.29 Our data, however, and diabetes, it is important to encourage
believed to have decreased substantially in indicate the expected higher intake in further improvements and to monitor the
the last three decades and is blamed for the males and the expected age gradient. The changes in the diet of Canadians on a regu-
increase in obesity.25 data for energy intakes of men appear to lar basis.

384 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5


REFERENCES
1. World Cancer Research Fund Report. Food,
Nutrition and the Prevention of Cancer.
American Institute for Cancer Research.
Washington DC, 1997.
2. Lichtenstein AH, Kennedy E, Barrier P, et al.
Dietary fat consumption and health. Nutr Rev
1998;56:S3-S19.
3. Birmingham LC, Muller JL, Spielli JJ. The cost
of obesity in Canada. CMAJ 1999;160:483-88.
4. Ganji V, Betts N. Fat, cholesterol, fibre and sodi-
um intakes of US population: Evaluation of diets
reported in 1987-88 Nationwide Food
Consumption Survey. Eur J Clin Nutr
1995;49:915-20.
5. Anderson SA, Waters JH. Executive summary from
the Third Report on Nutrition Monitoring in the
United States. J Nutr 1996;126:1907S-1936S.
6. Nutrition Canada. Nutrition: A national priority.
Ottawa: Information Canada, 1973.
7. Nutrition Canada. Food Consumption Patterns
Report. Bureau of Nutritional Sciences, Health
Protection Branch, Department of National
Health and Welfare (undated).
8. Pomerleau J, Østbye T, Bright-See E. Place of
birth and dietary intake in Ontario: Energy, fat,
cholesterol, carbohydrate fiber and alcohol. Prev
Med 1998;27:32-40.
9. Pomerleau J, Østbye T, Bright-See E. Place of
birth and dietary intake in Ontario II: Protein and
selected micronutrients. Prev Med 1998;27:41-49.
10. Sevenhuysen GP, Gelenskey D, Macdonald S.
The Manitoba Heart Health Project: Nutrition
Survey Technical Report, Department of
Community Health Sciences, Faculty of
Medicine, University of Manitoba, 1991.
11. Nova Scotia Department of Health. Report of
the Nova Scotia Nutrition Survey 1993.
12. Les Québécoises et les Québécois mangent-ils
mieux? Éd : Bertrand L. Gouvernement du
Québec, 1995.
SEPTEMBER – OCTOBER 2000
13. Robichon-Hunt L, Robbins L. Food sources of
nutrients available in the Canadian diet. 1990.
An estimation based on food purchase data.
J Can Diet Assoc 1993;54:185-89.
14. Beaton GH, Milner J, Corey P, et al. Sources of
variance in 24-hour dietary recall data:
Implications for nutrition study design and inter-
pretation. Am J Clin Nutr 1979;32:2456-559.
15. Guenther PM, Kott PS, Carriquiry AL.
Development of an approach for estimating usual
intake distributions at the population level.
J Nutr 1997;127:1106-112.
16. Statistics Canada. Selected Income Statistics -
The Nation. Ottawa: Industry, Science and
Technology Canada, 1993. Catalogue 93-331.
17. American Data Base. Adams CF. Nutritive value
of American foods in common units, Agricultural
Handbook No. 456. Agr. Res. Serv., U.S. Dept
Agric, Washington, DC, 1975.
18. Health and Welfare Canada. Using the Food
Guide. Ottawa: Supply and Services Canada, 1992.
(Cat No H39-253/1992E; ISBN 0-662-19649-X).
19. National Research Council Subcommittee on
Criteria for Dietary Evaluation. Nutrient
Adequacy: Assessment Using Food Consumption
Surveys. Washington, DC: National Academy
Press, 1986.
20. FAO/WHO/UNU. Energy and Protein
Requirements. Report of a Joint
FAO/WHO/UNU Expert Consultation. WHO
Tech. Rep. Ser. 724, 1985.
21. National Population Health Survey, 1994-1995.
Ottawa: Statistics Canada, 1995,Cat No 82-
F0001XCB.
22. Nutrition Recommendations. The report of the
Scientific Committee. Ottawa: Minister of
Supply and Services, 1990.
23. Goldberg GR, Black AE, Jebb SA, et al. Critical
evaluation of energy intake using fundamental
principles of energy physiology. Derivation of cut
off limits to identify under-recording. Eur J Clin
Nutr 1991;45:569-81.
CANADIAN JOURNAL OF PUBLIC HEALTH 385
CANADIAN DIETARY SURVEY
24. Standing Committee on the Scientific Evaluation
of Dietary Reference Intakes. Food and Nutrition
Board Institute of Medicine. Dietary Reference
Intakes for Calcium, Phosphorus, Magnesium,
Vitamin D and Fluoride. Washington, DC:
National Academy Press, 1999.
25. Seidell JC. Time trends in obesity: An epidemio-
logical perspective. Horm Metab Res
1997;29:155-58.
26. Flegal KM, Carroll MD, Kucmarski RJ, Johnson
CL. Overweight and obesity in the United States:
Prevalence and trends, 1964-1994. Int J Obes
Met Dis 1998;22:39-47.
27. Willett WC. Is dietary fat a major determinant of
body fat? Am J Clin Nutr 1998;67:5565-625.
28. Sawaya AL, Tucker K, Tsay R, et al. Evaluation
of four methods for determining energy intake in
young and older women: Comparison with dou-
bly labeled water measurements of total energy
expenditure. Am J Clin Nutr 1996;63:491-99.
29. Hirvonen T, Männistö S, Roos E, Pietinen P.
Increasing prevalence of underreporting does not
necessarily distort dietary surveys. Eur Clin Nutr
1997;51:297-301.
30. Lehmann F, Gray-Donald K, Nongeon M, Di
Tommasso S. Iron deficiency anemia in 1-year-
old children of disadvantaged families in
Montreal. CMAJ 1992;146:1571-77.
31. Johnson Down L, O’Loughlin J, Koski KG,
Gray-Donald K. High prevalence of obesity in
low income and multiethnic school children: A
diet and physical activity assessment. J Nutr
1997;127:2310-15.
32. Gray-Donald K, Payette H, Bouthier V, Page S.
Evaluation of the dietary intake of homebound
elderly and the feasibility of dietary supplementa-
tion. J Amer Coll Nutr 1994;13:277-84.
Received: December 14, 1999
Accepted: May 11, 2000
 A B S T R A C T
The custom of displaying Chinese-style
barbecued meats at room temperature has
been a controversial food safety issue in North
America. This article is intended to facilitate
development of a risk-based food safety policy
for this unique food by providing a brief
overview of the recent study findings and
Canadian disease surveillance data. Despite
the lack of temperature control after cooking,
Chinese barbecued meats were rarely implicat-
ed in foodborne incidents in Canada between
1975 and 1993. This might be due to the
food’s ability to delay pathogen growth during
the first 5 hours immediately after cooking,
and the conventional trade practices of sepa-
rating the retail area from the main kitchen
(i.e., reducing risk of cross-contamination).
However, recent studies also pointed out the
high potential for cross-contamination during
the retail stage (i.e., chopping and packaging
the food) as a result of lack of proper hand-
washing and equipment sanitation. A risk-
based food safety policy is proposed.
A B R É G É
L’habitude de présenter à la température de
la pièce des viandes cuites au barbecue à la chi-
noise est une question qui suscite de la contro-
verse en Amérique du Nord. Cette coutume
représente-t-elle un risque pour la santé? Cet
article a pour but de favoriser l’élaboration
d’une politique basée sur la sécurité des ali-
ments, pour ce qui est de cet aliment particu-
lier, en donnant un bref aperçu des conclusions
de récentes études et des données de surveil-
lance des maladies au Canada. En dépit du
manque de contrôle des températures après la
cuisson, les viandes cuites à la broche à la chi-
noise sont rarement en cause dans des intoxi-
cations d’origine alimentaire au Canada entre
1975 et 1993. Cette situation peut résulter de
la capacité de l’aliment de retarder la crois-
sance de pathogènes durant les cinq premières
heures qui suivent la cuisson et de la pratique
commerciale qui consiste à séparer l’aire de
vente au détail de la cuisine principale (ce qui
a pour effet de réduire les risques de contami-
nation croisée). De récentes études ont toute-
fois démontré le risque élevé de contamination
à l’étape du détail (lors du découpage et de
l’emballage des viandes) en raison du manque
de désinfection de l’équipement et du fait que
les préposés négligent de se laver les mains. Il
est donc proposé d’adopter une politique de
sécurité alimentaire basée sur les risques.
386 REVUE CANADIENNE DE SANTÉ PUBLIQUE
Chinese-style Barbecued Meats:
A Public Health Challenge
A crucial preventive measure against
foodborne illness is the control of tempera-
ture to eliminate opportunities for incuba-
tion of pathogens. Potentially hazardous
foods should be retained either below 5°C
(refrigerated) or above 60°C (hot holding).
Chinese-style barbecued meats, however,
are conventionally displayed at room tem-
perature in enclosed showcases. In North
America some of these showcases may be
equipped with a warming device, but it is
rarely switched on during normal opera-
tion. The Chinese food industry in Canada
has for years argued that hot or cold hold-
ing can render the products undesirable for
their customers, and that the tradition of
displaying at ambient temperature has
been practised in Asia for many years with-
out being known to cause any major food-
borne outbreak.
This article will provide a closer exami-
nation of the food safety risk associated
with Chinese-style barbecued meats and
will discuss a risk-based public health
approach to this particular food.
Brief overview of study findings
Although a number of studies have been
undertaken to assess the food safety risk of
Chinese-style barbecued meats, most of
them are unpublished and thus their find-
ings remain largely unknown. The follow-
ing is a brief overview of both the pub-
lished and unpublished studies and their
findings.
The first study in Canada focussing on
the Chinese-style barbecued meats dates
Correspondence and reprint requests: Ms. Jane
Ying, Health Promotion and Environmental
Protection Office, Toronto Public Health, 277
Victoria Street, 7th Floor, Toronto, ON, M5B 1W1,
Tel: 416-338-8101, Fax: 416-392-7418, E-mail:
jying@city.toronto.on.ca
Jane Ying, MHSc, CPHIC
back to 1976 by Tiwari et al.1 Among the
total 138 meat samples collected from four
retail outlets in Edmonton, the study
found low levels of contamination in those
taken within two hours after cooking, but
substantial increases in the number of col-
iforms and other pathogens in those that
had been stored at 22°C for 20 hours.
Stiles and Ng in 1977 conducted a labo-
ratory study in which a number of entero-
toxigenic bacteria (Bacillus cereus, 2 strains
of Clostridium perfringens, Escherichia coli,
Salmonella typhimurium, and
Staphylococcus aureus) were inoculated onto
Chinese barbecued chickens, ducks, and
pork tenderloins purchased from two
Chinese stores in Edmonton.2 A number
of unexpected results were found: after 2-3
hours of incubation at 30°C, the bacterial
counts decreased to levels lower than the
initial inoculation; at 5 hours, most bacte-
ria showed little or no growth with the
exception of one strain of Clostridium per-
fringens; the lag phase of most bacteria per-
sisted up to 8 hours after incubation.
Substantial growth was noted among all
organisms after 20-22 hours of incubation
at 30°C. Figure 1 shows the growth rates
for three E.coli replicates. Stiles and Ng
concluded that Chinese barbecued prod-
ucts should be stored at temperatures out-
side the danger zone even though they
appeared to exhibit a long lag phase for
most of the tested organisms.
Robinson and Mathews in 1990 con-
ducted a similar study inoculating E.coli,
Salmonella muenchen and Staphylococcus
aureus onto Chinese barbecued ducks
obtained from a Chinese outlet in
Toronto. 3,4 The study found that most
bacterial counts fell steadily after the initial
inoculation and by the end of 22 hours
incubation at 30°C, the counts either
VOLUME 91, NO. 5
 CHINESE-STYLE BARBECUED MEATS

dropped to zero or to levels lower than the

initial inoculum (Figure 2).
As shown in Figures 1 and 2, a major
discrepancy between the findings of the
two studies is the bacterial counts at the
end of the 22 hours incubation. A number
of factors could have accounted for this
difference:
1. The recipes used to prepare the prod-
ucts might vary substantially since the
two studies were conducted 13 years
apart in two different regions in the
country;
2. The inoculates were poured onto the
meat products in Stiles and Ng’s study,
whereas in Robinson and Mathews’ study
they were rubbed onto the duck surfaces Figure 1. E. coli Growth in BBQ Ducks, Stiles and Ng 1977
using a sponge. This, according to
Robinson, was necessary because the
poured cultures rolled off the smooth and
greasy duck surfaces in a manner similar
to “water rolling off a teflon surface”;3
3. All the tested products in Stiles and
Ng’s study were wrapped with alumini-
um foil for incubation whereas repli-
cates 1 and 2 in Robinson and
Mathews’ study were covered only
loosely with a sheet of aluminium foil,
and replicate 3 was covered tightly with
aluminium foil similar to that described
in Stiles and Ng’s study. Wrapping the
tested products with foil could help
maintain the moisture (i.e., higher water
activity) and hence might render the
products more favourable for pathogen Figure 2. E. coli Growth in BBQ Ducks, Robinson and Mathews 1990
growth. Since Chinese barbecued meats
are conventionally displayed uncovered preparation procedures for barbecued Prompted by the findings of Chan et al.,
in the stores, the study findings of repli- ducks (dipping the ducks in a malt and Sahota in 1996 carried out a series of envi-
cates 1 and 2 from Robinson and vinegar mixture and air drying the skin) ronmental samples to determine the extent
Mathews’ study may be more represen- can render the outer surfaces less favourable to which the conventional methods for
tative of the actual growth rates. for bacterial growth. The study also found cleaning equipment in Chinese barbecued
Both studies noted that the tested prod- that the ducks that were aseptically meat outlets is effective in reducing micro-
ucts had either no or very low levels of retrieved from the oven and placed whole bial load.6 As it is impossible to immerse
contamination immediately after cooking. into the display cabinet showed no or min- the large wooden cutting block into a sink,
While the initial inoculum levels varied imal signs of bacterial growth even after the conventional cleaning procedures
between the studies, both found that the having been displayed at room temperature involve scraping the grease from the wood-
counts either declined or grew very slowly for five hours. Ducks that had undergone en surface between orders by means of a
during the first five hours of incubation. the normal retailing process (display, chop- cleaver, and then wiping off the remaining
In 1994, Chan et al. undertook an ping and packaging) showed increases in E. grease using a dish cloth. The study found
observational study including a Hazard coli and coliform levels. The probability of high bacterial counts on both the cutting
Analysis Critical Control Point (HACCP) cross-contamination during the vending blocks and the cleavers throughout the
analysis of the complex preparation proce- process was strengthened when the study day, and even prior to use in the morning.
dures, and were the first to identify a authors observed a lack of handwashing During operation, the bacterial counts
potential problem at the retail stage.5 The among the food handlers and an absence of increased after scraping and wiping with
authors suggest that the conventional equipment sanitization. the dish cloth, which was also found to be

SEPTEMBER – OCTOBER 2000 CANADIAN JOURNAL OF PUBLIC HEALTH 387

CHINESE-STYLE BARBECUED MEATS
heavily loaded with bacteria. Sahota also
noticed the lack of personal hygiene and
equipment sanitation during the retailing
process.
According to the Ontario Food Premises
Regulation 562 (1993), large equipment
should be sanitized with a 200 ppm chlo-
rine solution. Since it will be impractical to
require transporting a large butcher block
to a sink for cleaning and sanitizing, the
(former) Toronto Public Health
Department tested the method of spread-
ing a 200 ppm chlorine solution onto the
block for 5-minute contact time in two
retail premises. The method was capable of
reducing microbial load only on the
cleavers, but not on the wooden cutting
blocks. It is hypothesized that the porous
nature of the wood, and the grease from
the meat products, might have reduced the
disinfecting effect of chlorine.
A laboratory study has subsequently
been conducted to determine what (if any)
means of disinfecting is effective for the
wooden cutting block. 7 Two chlorine-
based commercial sanitisers designed to
penetrate and remove soil from food
preparation equipment were tested for
their ability to reduce microbial load
(E.coli was used for this study) from a
wooden cutting surface. A standard
amount of fat extracted from Chinese bar-
becued pork was evenly spread onto the
cutting surface to simulate typical condi-
tions in these restaurants. Unfortunately,
neither sanitiser was able to meet the crite-
rion of a 3-log unit reduction in surface-
adherent cells, a benchmark for determin-
ing sanitation effectiveness.8
Surveillance data
Despite the concerns about lack of tem-
perature control and poor sanitation,
Chinese-style barbecued meats have rarely
been implicated in foodborne incidents in
Canada. Between 1975 and 1993 (the only
period for which records are available),
about 7% of the 16,634 reported food-
borne incidents in Canada were suspected
to be associated with consumption of
Chinese foods, but most of them were fried
rice and egg rolls.9,10 Chinese barbecued
meat was specifically implicated in only 14
incidents (less than 0.1%) over the 18
years. Although the number of foodborne
388 REVUE CANADIENNE DE SANTÉ PUBLIQUE
incidents associated with Chinese barbe-
cued meat might have risen slightly in the
latter half of the 1990s as consumption of
this food continued to rise, the risk of con-
tracting foodborne illnesses from Chinese
barbecued meat is very small compared
with other potentially hazardous foods.
DISCUSSION
While most of the studies thus far are
small-scale investigations incapable of
offering conclusive evidence, several consis-
tent findings did emerge:
1) The conventional cooking procedures
for Chinese barbecued ducks, and prob-
ably for barbecued pork, produce inter-
nal temperatures high enough to destroy
all vegetative pathogens resulting in
absence or very low levels of microbial
load on the products immediately after
cooking (note: not precluding possible
undercooking due to human or equip-
ment error in individual outlets);
2) When challenged with a number of
common foodborne organisms in labo-
ratory settings, the outer surface of the
freshly cooked Chinese barbecued ducks
appeared to be able to delay pathogen
growth during the initial 5 hours of
storage at 30°C;
3) The retail process (cutting, handling,
and packaging) could represent a high
potential for cross-contamination due to
lack of proper handwashing and equip-
ment sanitation.
The conventional trade practices of
Chinese-style barbecued meats also provide
some additional safeguards against food
poisoning. The products are usually
cooked in the main kitchen and then
promptly transported to a separate retail
area, often located in the front of the
premises, where no or only limited cooking
takes place (e.g., boiling noodles and pre-
formed dumplings). This separation from
the main kitchen area is vital in avoiding
cross-contamination between cooked and
raw meat. In addition, most outlets tend to
make only a small batch at a time to ensure
their products are sold quickly. This is
especially important for barbecued ducks
because prolonged storage will cause the
skin to lose its crispness and shine, becom-
ing less desirable to customers.
An unexpected characteristic of these
products is their ability to inhibit bacterial
growth during the first five hours after
cooking. This protective factor against the
risk of foodborne illness is, however, sub-
ject to several conditions. For example, the
outer layer of barbecued ducks is protective
only when the cooked surface is intact.
Once it is cut and the skin is broken, the
meat underneath may be subject to
pathogen growth as is any other potentially
hazardous food, and cross-contamination
can easily happen during cutting of the
meat. Unfortunately, for convenience rea-
sons, most customers, when purchasing
barbecued meats for take-out, prefer to
have them chopped into bite-size pieces at
the store. Since the food is kept at room
temperature at point-of-purchase, con-
sumers may have the wrong impression
that refrigeration is not necessary. The haz-
ard can thus be further increased when the
meats are stored unrefrigerated in a car or
in home for several hours prior to con-
sumption (i.e., allowing opportunity for
bacterial incubation), and this is of particu-
lar concern during the summer months. In
addition, although most outlets follow the
traditional methods of preparing Chinese-
style barbecued meats, the exact recipes
pertaining to steps such as skin coating and
drying time before cooking may vary
slightly from outlet to outlet, and from
region to region. It is not known to what
extent these variations may affect the pro-
tective outer layer.
While it would be useful to gain better
understanding of the nature and risk fac-
tors of this food, developing a sound food
safety policy remains difficult for public
health officials. Many factors need to be
considered, including risk management
and assessment, burden of illness, legal lia-
bility, enforceability of any food safety
requirements, and last but not least, cultur-
al sensitivity. The problem in equipment
sanitizing presents an additional challenge
to public health officials as studies have
suggested that even adhering to the regula-
tory requirement (i.e., 200 ppm chlorine
solution) may not necessarily result in a
safe food contact surface.
From the risk management perspective,
the evidence thus far suggests that it would
be more cost-effective to focus the limited
VOLUME 91, NO. 5

inspection resources on reducing risk of
cross-contamination during the retail stage
than on enforcing the debatable tempera-
ture control requirements for the cooked
meats on display. While further studies are
needed in this area, an interim policy may
be to allow room temperature display of
the whole Chinese-style barbecued ducks
and barbecued pork for no more than five
hours providing the following conditions
are met:
1) The meats are prepared according to the
conventional method including steps
such as marinating the meats with vine-
gar and malt, and air drying of the duck
skin prior to cooking;
2) All the critical control points identified
by HACCP audits are properly executed
(except allowing for up to five hours of
room temperature display);
3) Basic standards in equipment sanitizing
and personal hygiene are incorporated
into daily operation and practised con-
sistently; and
4) Food samples are regularly taken, and
bacterial counts do not exceed the safe
limits.
More important is to develop a compre-
hensive food safety program to address the
following aspects:
1) Food handler education – food handlers
need to understand the potential risks of
the products and precautionary mea-
sures such as thorough cooking, ade-
quate handwashing and equipment sani-
tation;
2) Public health staff training –
Environmental Health Officers (Public
Health Inspectors) must be knowledge-
able about the potential risk and nature
of this food so that they can assist the
food handlers to incorporate the proper
critical control measures into daily oper-
ation;
3) Public awareness – consumers need to
be informed of the risk potential of this
food and the safeguard measures (e.g.,
refrigeration prior to consumption);
strategies such as placing a multi-
language warning label on take-out
boxes should be considered by public
health departments;5
SEPTEMBER – OCTOBER 2000
4) Partnership with the Chinese food
industry – public health officials need to
work with the industry to develop a
practical retailing routine that also
meets current food safety standards
(e.g., equipment cleaning and sanitizing
procedures);
5) Research – further studies are required
to systematically assess the risk of
Chinese barbecued meat products and
other non-barbecued products (e.g., soy
sauce chicken, steamed chicken) that are
conventionally displayed at room tem-
perature, as well as in the area of equip-
ment sanitation.
The issues with Chinese-style barbecued
meats represent only the tip of the iceberg
of new food safety challenges being faced
by public health officials today. As immi-
gration continues, and the food industry
becomes increasingly globalized, public
health officials frequently encounter food
ingredients and preparation methods that
are unfamiliar in North America. Many of
these foods have not gone through any risk
assessment process in their countries of ori-
gin and thus very little food safety infor-
mation is available. It is the officials’
responsibility to respect the different food
cultures and at the same time to ensure
that the basic safety standards are in place
even though this often requires extra effort
in research and inspection resources.
CONCLUSION
A consistent and risk-based food safety
policy for Chinese-style barbecued meats is
long overdue in North America. Studies
have repeatedly shown that certain barbe-
cued meats possess a long lag phase
(5 hours) before pathogen growth, but this
protective characteristic will vanish once
the surface is cut open. Poor retail practices
(lack of proper handwashing and equip-
ment sanitation) may represent a much
higher risk for foodborne illness than previ-
ously thought. Development of a food safe-
ty policy should seek input from the
Chinese food industry to ensure the
requirements are practical and can be incor-
porated into daily operation. Further
CANADIAN JOURNAL OF PUBLIC HEALTH 389
CHINESE-STYLE BARBECUED MEATS
research is needed to fully assess the risk of
different types of Chinese meat products
that are displayed at room temperature, and
it is hoped that this review article will serve
as a catalyst for further studies in this area.
ACKNOWLEDGEMENTS
This study is a result of Toronto Public
Health’s commitment to more effective pub-
lic services. The author wishes to thank
Ryerson Polytechnic University for their stu-
dent and faculty researches; Tom Wong,
Devinder Sahota, and Mark Shaw of Toronto
Public Health for their participation in the
pilot study; and Dr. Tim Sly and Professor
Pat Robinson of Ryerson Polytechnic
University for reviewing the manuscript.
REFERENCES
1. Tiwari NP, Kadis VW, Kemp GC. Comparison
of the microbiological quality of Chinese and
non-Chinese barbecued meats obtained from
Edmonton retail outlets. Can J Public Health
1976;67:485-89.
2. Stiles ME, Ng LK. The fate of enterotoxigenic
bacteria inoculated onto Chinese barbecued
foods. Can J Public Health 1977;68(5):389-94.
3. Robinson P. Fate of Bacteria (E. coli, Salmonella
muenchen and Staphylococcus aureus) inoculat-
ed onto Chinese Barbecued Duck. Report to the
Research and Scholarship Committee, Faculty of
Community Services, Ryerson Polytechnic
University, 1991.
4. Mathews E. Chinese Barbecued Duck and
Escherichia coli: A Challenge Test. Student
Thesis, Ryerson Polytechnic University, 1990.
5. Chan JI, Chan E, Leung P. HACCP Survey on
Chinese BBQ Duck - A Practical Approach.
Environmental Health Review 1994 (Winter).
6. Sahota D. Chinese Style BBQ Meats - Utensil
Disinfection. Number 42 in Series of Student
Reports on Environmental Health Issues.
Metropolitan Toronto Teaching Health Units
and Ryerson Polytechnic University. Toronto:
1996.
7. Quiros MN. A Study on Efficacy of Proposed
Sanitizing Agents in Disinfecting Greasy Cutting
Board. Student Thesis. Ryerson Polytechnic
University, 1998.
8. Mosteller TM, Bishop JR. Sanitizer efficacy
against attached bacteria in a milk biofilm. J
Food Protection 1993;56:34-41.
9. Todd ECD. Foodborne Disease in Canada: A
10-year Summary, 1975-1984. Health Protection
Branch, Health and Welfare Canada, Ottawa,
1991.
10. Todd ECD. Foodborne Disease in Canada:
Annual Summaries, 1985-86, 1988-89, 1990-91,
1992-93. Health Protection Branch, Health
Canada, Ottawa.
Received: September 20, 1999
Accepted: February 28, 2000
 A B S T R A C T
Accurate and complete registries are an
important source of knowledge about cancer.
The concordance of the recording of neo-
plasms in the Saskatchewan cancer registry
with that in hospital charts and death regis-
trations was evaluated for 368 patients. The
agreement between registry and hospital
charts or death registrations was excellent
(kappa: 0.93; 95% confidence interval: 0.89,
0.97), with 91.3% of those with cancer hav-
ing the same neoplasm recorded in their
chart or death registration as in the registry.
There was only one patient whose hospital
chart indicated cancer who was not in the
registry and one apparent major discrepancy
relating to the cancer site, which was due to
the recording of the primary site in the reg-
istry and a secondary in the hospital chart.
Although based on a relatively small number
of patients, these results suggest a high degree
of consistency between cancer registry, hospi-
tal charts and death registrations in
Saskatchewan.
A B R É G É
Des registres précis et complets sont une
source importante d’information sur le can-
cer. La concordance de l’enregistrement des
néoplasmes dans le registre du cancer dans la
Saskatchewan, les dossiers des hôpitaux et les
certificats de décès a été évaluée pour 368
malades. Le degré d’accord entre le registre,
les dossiers des hôpitaux et les certificats de
décès était excellent (kappa: 0,93; intervalle
de confiance à 95 %: 0,89, 0,97), avec
91,3 % de ceux avec le cancer ayant le même
néoplasme enregistré dans leurs dossiers ou
leurs certificats de décès comme dans le reg-
istre. Il y avait un seul malade où le dossier
d’hôpital indiquait un cancer qui n’était pas
dans le registre et apparemment, un désac-
cord important concernant le site d’un can-
cer où le site primaire avait été inclus dans le
registre mais le site secondaire pour le dossier
de l’hôpital. Bien qu’ils soient basés sur rela-
tivement un petit nombre de malades, ces
résultats suggèrent un haut degré de
cohérence entre le registre du cancer, les
dossiers des hôpitaux et les certificats de
décès dans la Saskatchewan.
390 REVUE CANADIENNE DE SANTÉ PUBLIQUE
Concordance on the Recording of
Cancer in the Saskatchewan Cancer
Agency Registry, Hospital Charts and
Death Registrations
Nigel S.B. Rawson, PhD,1 Diane L. Robson, BA2
Accurate and complete registries are an
important source of information for
research studies investigating a variety of
issues related to disease occurrence.1 Many
countries have established cancer registries
and endeavoured to make them as accurate
as possible,2 although studies to evaluate
completeness of coverage and data accura-
cy have shown a considerable disparity
between registries.3-12 Even if the overall
accuracy is good, wide variation may exist
in the quality of information from differ-
ent reporting sources or about individual
cancers.
In Saskatchewan, cancer is a reportable
disease, with information on cases being
collected and maintained by the
Saskatchewan Cancer Agency (SCA) in its
population-based cancer registry, which is
one of the oldest in the world having been
started in 1932.13 The principal source of
registrations (91% of all cases excluding
non-melanoma skin cancers) is pathology
reports. A further 3% of the registrations
come from death certificates, and the rest
are from physician service claims, hospital
reports, physician referrals and an inter-
provincial data exchange. All physician ser-
vice claims with a cancer disease or proce-
1. Professor of Pharmacoepidemiology, Division of
Community Health, Faculty of Medicine, and
School of Pharmacy, Memorial University of
Newfoundland, St. John’s, Newfoundland
2. Director of the Cancer Registry, Saskatchewan
Cancer Agency, Regina, Saskatchewan
Correspondence and reprint requests: Dr. Nigel S.B.
Rawson, Division of Community Health, Memorial
University of Newfoundland, Health Sciences
Centre, St. John’s, NF, A1B 3V6, Fax: 709-737-
7382, E-mail: nrawson@mun.ca
Although this study is based on data provided by
Saskatchewan Health and the Saskatchewan Cancer
Agency, the interpretation and conclusions are those
of the authors and do not necessarily represent those
of Saskatchewan Health, the Saskatchewan Cancer
Agency or the Government of Saskatchewan.
dure code are referred to the SCA for
approval before payment and any case not
registered is followed up with the appropri-
ate physician to verify the diagnosis. The
SCA registry is reported to be one of the
most complete in Canada,14 although the
evidence is limited to the high rate of
microscopically confirmed tumours2 and
an analysis of the recording of Wilms’
tumour cases.15
Knowledge about the consistency
between different data sources within a
health care system is important for patient
care and research purposes.16,17 In a study
of aplastic anemia and agranulocytosis18 –
conditions that can result from the disease
and treatment processes of cancer – infor-
mation on patients was obtained from the
SCA registry, hospital chart abstracts and
death registrations and an evaluation of the
concordance between these data sources
was possible.
METHODS
The Saskatchewan government supplies
a wide range of publicly funded health care
programs to provincial residents, while the
SCA provides cancer therapy.14,19 Each eli-
gible resident is issued with a unique
health service number (HSN) that is
required to obtain benefit from health care
programs (>95% of the population of just
over one million are eligible residents; the
rest are served by federal programs). The
HSN is recorded in the relevant datafile
when service is provided and is the key to
data linkage. The accuracy of the HSNs is
maintained on a daily basis and the relia-
bility of the linkage using the HSN is
good. The accuracy and reliability of the
Saskatchewan health care datafiles have
been found to be acceptable for distinct
VOLUME 91, NO. 5

conditions, but poorly defined diagnoses
and those determined from laboratory
results should be examined with caution.20-
24
Considerable use has been made of the
datafiles for research purposes.14,19,25,26
In the aplastic anemia and agranulocyto-
sis study,18 all patients with primary dis-
charge diagnosis codes for these dyscrasias
but without a secondary code for cancer
between 1982 and 1991 were identified
from the provincial hospitalization datafile.
Information on each patient included gen-
der, age, admission and discharge dates,
and primary and secondary discharge diag-
noses using four-digit International
Classification of Diseases (ICD-9) codes.27
Unique dummy identifiers were substitut-
ed for the HSNs before data were supplied
for analysis.
Because the computer records do not
contain clinical or laboratory data either to
confirm the disorder or to rule out alterna-
tive causes, health record abstractors
obtained this information from hospital
charts for the relevant admissions. Hospital
charts can be identified accurately, using
the HSN and other codes recorded in both
the computer record and the chart, and the
retrieval rate is normally close to 100%.24
The abstractors were trained to ensure a
standardized approach to data collection
and used a purpose-designed form to
record the information. Death registrations
were obtained from the provincial vital sta-
tistics records for most of the patients who
had died.
The hospital chart abstraction process
suggested that many dyscrasia patients
with cancer reported in their chart did not
have the cancer recorded as a secondary
diagnosis in the hospitalization datafile,
even though the dyscrasia appeared to be
directly related to the neoplasm or its treat-
ment. To maximize the identification of
cancer cases, the HSNs, names and
addresses of all patients in the study were
supplied by Saskatchewan Health directly
to the SCA for a check against its registry.
The SCA provided International
Classification of Diseases for Oncology
(ICD-O) diagnostic codes28 for the identi-
fied patients.
The concordance regarding cancer
occurrence between the SCA registry, hos-
pital charts and death registrations was
SEPTEMBER – OCTOBER 2000
CONCORDANCE ON RECORDING OF CANCER IN SK
TABLE I
Availability of Information for the 368 Patients in the Evaluation
Cancer Patients 127
Died, hospital chart abstract and death registration obtained 57
Died, death registration obtained, but no hospital chart abstract 26
Alive, hospital chart abstract obtained 44
Non-cancer Patients 241
Died, hospital chart abstract and death registration obtained 88
Died, hospital chart abstract obtained, but death registration unobtainable 6
Alive, hospital chart abstract obtained 147
quantified and evaluated, using kappa 29 TABLE II
with a 95% confidence interval (CI), for Sites of Cancer Recorded for the
those patients for whom information was 127 patients
available. Because there is no direct equiva-
Site of Cancer No.*
lence between many of the ICD-9 and
ICD-O codes and diagnoses from the hos- Blood 42
Lymph 19
pital charts and death registrations were Breast 13
recorded in a verbatim manner, “concor- Colo-rectal 11
Lung 11
dance” was defined as the same site for Prostate 10
solid tumours and the same broad type, Bladder/kidney 8
Eye/nasopharynx/sinus 5
e.g., leukemia, for hematological cancers. Lip 3
Ovary 3
Adrenal 2
RESULTS Brain 2
Multiple myeloma 2
Leg (osteosarcoma) 2
A total of 397 patients with a primary Ewing’s 1
diagnosis code for aplastic anemia or Pancreas 1
Thymus 1
agranulocytosis were identified from the Other 3
hospitalization datafile. Twenty-seven
* 10 patients had two sites and 1 had three
patients (9 in the cancer registry and 18 sites
not in it) had neither a hospital chart
abstract nor a death registration because the study from which the patients were
the chart was unavailable or the hospital- obtained focussed on blood dyscrasias, it is
ization occurred prior to the patient being not surprising that a high proportion of
entered in the SCA registry, and, for those the cancers (48.0%) were hematological or
who had died, no registration was found. lymphomas. Nevertheless, over half were
In addition, 1 patient had thrombo- solid tumours with a range of sites.
cythemia and another had neurofibro- Table III shows the concordance of the
matosis recorded on their respective death recording of cancer in the SCA registry,
registrations; these conditions are consid- hospital charts and death registrations for
ered to be benign disease by the SCA not all 368 patients. For 124 of the 127 cancer
requiring registration unless chemotherapy patients (97.6%), the SCA and one or
is needed and neither was in the cancer both of the other data sources agreed that
registry. These 29 patients were excluded the person had cancer and, for 116
from further analysis. (91.3%), the other source(s) recorded the
The availability of information about the same cancer. The kappa value for Table III
remaining 368 patients is described in is 0.93 (95% CI: 0.89, 0.97), which indi-
Table I. One hundred and twenty-seven cates excellent agreement.30 When hospital
(34.5%) were either recorded in the cancer charts and death registrations were exam-
registry or had a hospital chart or death ined separately, the concordance with the
registration that mentioned cancer. The SCA registry was 90.9% and 82.0%,
median age of both these patients and the respectively, and the kappa values were
other 241 was 68 years, with a range of less 0.93 (95% CI: 0.88, 0.97) and 0.82 (95%
than 1 year to 95 years. CI: 0.74, 0.91).
The sites of the cancers recorded for the One patient, a 78-year-old male, had a
127 patients are listed in Table II. Since hospital chart that mentioned adenocarci-
CANADIAN JOURNAL OF PUBLIC HEALTH 391
CONCORDANCE ON RECORDING OF CANCER IN SK

TABLE III The only major discrepancy was the 67-

Concordance of Cancer Registry Data, Hospital Charts and Death Registrations year-old female (case 9) recorded by the
SCA as having melanoma of the choroid (a
Cancer Registry Hospital Chart Death Registration No. % vascular membrane covering most of the
(n=368)
posterior of the eye between the retina and
Registered Same neoplasm Same neoplasm 43 11.7 the sclera) for whom her hospital chart
Registered Same neoplasm Different neoplasm 2* 0.5
Registered Same neoplasm No mention 6 1.6 stated “carcinoma of lung.” Further com-
Registered Different neoplasm Different neoplasm 2* 0.5 parison of the hospital chart abstract and
Registered No mention Same neoplasm 3 0.8
Registered Not abstracted Same neoplasm 22 6.0 the SCA record for this patient indicated
Registered Not abstracted Different neoplasm 3* 0.8 that the correct chart had been abstracted.
Registered Not abstracted No mention 1 0.3
Registered Same neoplasm Alive 40 10.9 The SCA had correctly recorded the pri-
Registered Different neoplasm Alive 2* 0.5 mary cancer (choroid melanoma), whereas
Registered Chemotherapy for unspecified neoplasm Alive 1 0.3
Registered No mention Alive 1 0.3 a secondary site (lung cancer) diagnosed
Not registered Adenocarcinoma of the lung No mention 1 0.3 shortly before the hospitalization in ques-
Not registered No mention No mention 88 23.9
Not registered No mention Not found 6 1.6 tion had been entered in the hospital chart
Not registered No mention Alive 147 39.9 as what appeared to be another primary
* Details of these nine discrepant cases are provided in Table IV cancer.

TABLE IV DISCUSSION
Details of the Nine Discrepant Cases
This evaluation of the concordance
Case Age/ Cancer Registry Hospital Chart Death Registration
No. Sex between cancer registry data, hospital
charts and death registrations in
1 68/M Acute leukemia, NOS Acute leukemia, NOS Myelodysplasia
2 31/M Erythroleukemia Erythroleukemia Red cell aplasia Saskatchewan is an opportunistic one and,
as such, has limitations. Only a relatively
3 83/F Myeloproliferative disorder Acute lymphoblastic Acute lymphoblastic
and “cancer of uncertain leukemia leukemia small number of patients were included, all
behaviour” of whom had a blood dyscrasia, which may
4 78/F Cancer of the kidney Cancer of the bladder, Cancer of the bladder
rectum, cervix and limit the generalizability of the results,
peritoneum although there was no other obvious influ-
5 72/M Leukemia, NOS Not abstracted Pre-leukemia ence in selecting the patients. No hospital
6 84/F Leukemia, NOS Not abstracted Myelodysplasia abstract was sought for 25 patients who
7 83/F Leukemia, NOS Not abstracted Myelodysplasia
had a death registration mentioning cancer
8 5/M Rhabdomyosarcoma Rhabdomyosarcoma Alive (Table III) because this was sufficient
of the epididymis of the testicle
9 67/F Choroid melanoma Cancer of the lung Alive information to remove them from the
aplastic anemia and agranulocytosis
NOS: Not otherwise specified
study.18 A range of cancers were included,
noma of the lung but was not in the reg- tion (cases 1, 2, 5, 6 and 7); cases 5-7 were but there were few patients with some
istry, and no neoplasm was recorded in his all “death certificate only” registrations. types of cancer. Finally, the level of agree-
death registration when he died two The hospital chart for case 4 indicated ment was limited to the general cancer
months after being discharged. There were widespread cancer of the “uterine endo- type or site. Nevertheless, the findings pro-
also two patients recorded in the cancer cervix involving the urinary tract, bladder, vide information about the reliability and
registry who had either a death registration rectum and pelvic peritoneum,” but the accuracy of the Saskatchewan cancer reg-
that did not indicate cancer and no hospi- SCA had recorded cancer of the kidney, istry data, which has been sorely lacking.
tal chart abstract (a 78-year-old male), or a which may have been a secondary site. The Concordance on the occurrence of can-
hospital chart abstract that did not men- discrepancy in site between registry and cer between the data sources was 98% and,
tion cancer and was alive (a 74-year-old hospital chart in case 8 was minor and for 91%, the registry and hospital chart or
female). unsurprising in a 5-year-old child. The 83- death registration agreed on the type or site
Details of the nine cases in Table III year-old female (case 3) whose hospital of the cancer. These figures are impressive,
with a hospital chart and/or death registra- chart and death registration both recorded especially considering the high proportion
tion disagreeing about the cancer type or acute lymphoblastic leukemia was regis- of hematological cancers, which can be dif-
site recorded by the SCA are provided in tered by the SCA as having myeloprolifera- ficult to classify.
Table IV. The most common discrepancy tive disorder and “cancer of uncertain However, high concordance should be
was leukemia being recorded in the registry behaviour” because insufficient data were expected with cancer being a reportable
and a pre-leukemic or leukemia-associated available to record more precisely. These disease in Saskatchewan and the require-
condition reported in the death registra- differences are regarded as minor. ment for registration before physicians,

392 REVUE CANADIENNE DE SANTÉ PUBLIQUE VOLUME 91, NO. 5


who are almost entirely remunerated on a
fee-for-service basis, can receive payment
for services with an associated cancer diag-
nosis. These conditions are also reflected in
the fact that only 6 of the 127 cancer
patients (4.7%) were registered posthu-
mously by the SCA from their death
records (the rate of “death certificate only”
registrations is usually around 3%). These
proportions are much lower than corre-
sponding figures (>20%) reported from
England,31-33 where cancer registration is
non-statutory, less complete 10,11,16 and
often inadequately funded.34 In our data,
there was only 1 patient with a hospital
chart that recorded a neoplasm who was
not registered and 2 registered patients
without a chart or a death registration that
mentioned cancer (Table III). The single
apparent major discrepancy (Table IV),
which resulted from the recording of the
primary site in the registry and the impre-
cise reporting of a secondary site in the
hospital chart, was probably due to the
SCA and the hospital having different
recording priorities.
Although there may be some limitations
on the generalizability from results based
on cases with blood dyscrasias to all
patients, this evaluation suggests a high
degree of concordance between the cancer
registry, hospital charts and death registra-
tions in Saskatchewan. However, because
many patients with cancer do not have a
secondary diagnosis recorded in the hospi-
talization discharge datafile (127 in this
analysis), this file alone is inadequate to
identify cancer patients and linkage with
the more reliable SCA data is essential.
REFERENCES
1. Wilson S, Prior P, Woodman CBJ. Use of cancer
surveillance data for comparative analyses. J
Public Health Med 1992;14:151-56.
2. Canadian Council of Cancer Registries, Health
and Welfare Canada, and Statistics Canada. The
Making of the Canadian Cancer Registry: Cancer
Incidence in Canada and Its Regions, 1969 to
1988. Ottawa: Minister of Supply and Services
Canada, 1993.
SEPTEMBER – OCTOBER 2000
CONCORDANCE ON RECORDING OF CANCER IN SK
3. Brewster D, Crichton J, Muir C. How accurate
are Scottish cancer registration data? Br J Cancer
1994;70:954-59.
4. Glass S, Gray M, Eden OB, Hann I. Scottish val-
idation study of cancer registration data child-
hood leukaemia 1968-1981. Leuk Res
1987;11:881-85.
5. Gulliford MC, Bell J, Bourne HM, Petruckevitch
A. The reliability of cancer registry records. Br J
Cancer 1993;67:819-21.
6. Nwene U, Smith A. Assessing completeness of
cancer registration in the north-western region of
England by a method of independent compari-
son. Br J Cancer 1982;46:635-39.
7. Schouten LJ, Jager JJ, van den Brandt PA.
Quality of cancer registry data: A comparison of
data provided by clinicians with those of registra-
tion personnel. Br J Cancer 1993;68:974-77.
8. Swerdlow AJ, Douglas AJ, Vaughan Hudson G,
Vaughan Hudson B. Completeness of cancer reg-
istration in England and Wales: An assessment
based on 2,145 patients with Hodgkin’s disease
independently registered by the British National
Lymphoma Investigation. Br J Cancer
1993;67:326-29.
9. Vickers N, Pollock A. Incompleteness and
retrieval of case notes in a case note audit of col-
orectal cancer. Qual Health Care 1993;2:170-74.
10. Villard-Mackintosh L, Coleman MP, Vessey MP.
The completeness of cancer registration in
England: An assessment from the Oxford-FPA
contraceptive study. Br J Cancer 1988;58:507-11.
11. Warnakulasuriya KAAS, Acworth P, Bell J,
Johnson NW. Incompleteness of oral cancer reg-
istration in south-east England, 1971-87. Br J
Cancer 1994;70:736-38.
12. Seddon DJ, Williams EMI. Data quality in pop-
ulation-based cancer registrations: An assessment
of the Merseyside and Cheshire Cancer Registry.
Br J Cancer 1997;76:667-74.
13. Benson DL, Robson DL. Saskatchewan Cancer
Atlas 1970-1987. Regina: Saskatchewan Cancer
Foundation, 1988.
14. Strand LM, Downey W. Health databases in
Saskatchewan. In: Strom BL (Ed.),
Pharmacoepidemiology 2nd ed. Chichester: Wiley,
1994;217-29.
15. Haines CS, Wang PP, Cao Y. Wilms’ tumours in
Saskatchewan, 1932-1990. Chron Dis Can
1994;15:97-101.
16. Rushton L, Romaniuk H. Comparison of the
diagnosis of leukaemia from death certificates,
cancer registration and histological reports:
Implications for occupational case-control stud-
ies. Br J Cancer 1997;75:1694-98.
17. Rawson NSB, D’Arcy C. “Validity” and reliabili-
ty: Idealism and reality in the use of computerized
health care databases for pharmacoepidemiologi-
cal research. Post Market Surveill 1991;5:31-55.
18. Rawson NSB, Rutledge Harding S, Malcolm E,
Lueck L. Hospitalizations for aplastic anemia and
agranulocytosis in Saskatchewan: Incidence and
associations with antecedent prescription drug
use. J Clin Epidemiol 1998;51:1343-55.
19. Malcolm E, Downey W, Strand LM, et al.
Saskatchewan Health’s linkable data bases and
pharmacoepidemiology. Post Market Surveill
1993;6:175-264.
CANADIAN JOURNAL OF PUBLIC HEALTH 393
20. Rawson NSB, Malcolm E. Validity of the
Recording of Cholecystectomy and
Hysterectomy in the Saskatchewan Health Care
Datafiles. Pharmacoepidemiology Research Unit
Technical Report #3. Saskatoon: University of
Saskatchewan, 1995.
21. Rawson NSB, Malcolm E. Validity of the record-
ing of ischaemic heart disease and chronic
obstructive pulmonary disease in the
Saskatchewan health care datafiles. Stat Med
1995;14:2627-43.
22. Edouard L, Rawson NSB. Reliability of the
recording of hysterectomy in the Saskatchewan
health care system. Br J Obstet Gynaecol
1996;103:891-97.
23. Rawson NSB, Malcolm E, D’Arcy C. Reliability
of the recording of schizophrenia and depressive
disorder in the Saskatchewan health care
datafiles. Soc Psychiatry Psychiatr Epidemiol
1997;32:191-99.
24. Rawson NSB, D’Arcy C. Assessing the validity of
diagnostic information in administrative health
care utilization data: Experience in Saskatchewan.
Pharmacoepidemiol Drug Saf 1998;7:389-98.
25. Tennis P, Andrews E, Bombardier C, et al.
Record linkage to conduct an epidemiologic
study on the association of rheumatoid arthritis
and lymphoma in the province of Saskatchewan,
Canada. J Clin Epidemiol 1993;46:685-95.
26. Risch HA, Howe GR. Menopausal hormone
usage and breast cancer in Saskatchewan: A
record-linkage cohort study. Am J Epidemiol
1994;139:670-83.
27. World Health Organization. Manual of the
International Statistical Classification of Diseases,
Injuries and Causes of Death 9th rev. Geneva:
WHO, 1977.
28. Percy C, Van Holten V, Muir C (Eds.).
International Classification of Diseases for Oncology
2nd ed. Geneva: World Health Organization,
1990.
29. Landis JR, Koch GG. The measurement of
observer agreement for categorical data.
Biometrics 1977;33:159-74.
30. Altman DG. Practical Statistics for Medical
Research. London: Chapman & Hall, 1991;404.
31. Pollock AM, Vickers N. The impact on colorec-
tal cancer survival of cases registered by “death
certificate only”: Implication for national survival
rates. Br J Cancer 1994;70:1229-31.
32. Pollock AM, Vickers N. Why are a quarter of all
cancer deaths in south-east England registered by
death certificate only? – factors related to death
certificate only registrations in the Thames
Cancer Registry between 1987 and 1989. Br J
Cancer 1995;71:637-41.
33. Pollock AM, Benster R, Vickers N. Why did
treatment rates for colorectal cancer in South
East England fall between 1982 and 1988? – the
effect of case ascertainment and registration. J
Public Health Med 1995;17:419-28.
34. Day NE, Davies TW. Cancer registration:
Integrate or disintegrate? BMJ 1996;313:896.
Received: October 27, 1999
Accepted: February 15, 2000
C O M M E N T A R Y
Immigration, Women and
Health in Canada
Jacqueline Oxman-Martinez, PhD,1 Shelly N. Abdool, MA,2
Margot Loiselle-Léonard, PhD(C)3
Canadian statistics show that one in six
people are foreign-born comprising 17.4%
of the total population and, with the
exception of refugees, when immigrants
arrive in Canada they are healthier than
the Canadian-born population.1 Two fac-
tors are attributed to the healthy immi-
grant effect: 1) Canadian screening proce-
dures disqualify people with serious med-
ical conditions, and 2) healthy people are
more likely to emigrate.2 Immigrant health
deteriorates with length of stay in Canada.
Current approaches to health have not
adequately addressed this problem.3
While conscious that immigrant and/or
refugee men also face declines in health
status, we explore immigrant and/or
refugee women’s health as they are an
invisible, isolated population within
Canadian health interventions. Focus is
placed on cultural, socio-political and eco-
nomic issues in order to reflect upon future
avenues for research. No one foreseeable
solution to this growing problem has been
identified but the experiences, lives and
voices of immigrant and/or refugee women
are necessary to lead us towards appropri-
ate avenues.
Women, migration and health
Relations and perceptions are influenced
by gender, race, ethnicity, nationality, reli-
gion, level of education, class, physical
1. Adjunct Professor, School of Social Work;
Research Associate, The Centre for Applied
Family Studies, McGill University
2. Faculty of Nursing, Université de Montréal
3. École de service social, Université de
Montréal/McGill University
Correspondence and reprint requests: Jacqueline
Oxman-Martinez, The Centre for Applied Family
Studies, School of Social Work, McGill University,
3506 University Street, Suite 106, Montreal, QC
H3A 2A7
394 REVUE CANADIENNE DE SANTÉ PUBLIQUE
ability, sexual orientation and social loca-
tion. Both men and women participate in
the global phenomenon of migration,
however they experience this differently;
immigrant women face triple spheres of
oppression as women, workers and foreign-
ers in their new country.4,5
Women’s health is perceived as a contin-
uum that extends throughout the lives of
women, critically and intimately related to
their life conditions. Included among the
determinants of health outlined by Health
Canada are gender and culture. 6
Nonetheless, pre- and post-migratory expe-
riences remain unexplored as factors of
women’s health. Gender must be taken
into simultaneous consideration with race,
culture and social location within the con-
text of an intersectional approach.
Controversial issues and immigrant
and/or refugee women’s health
Cultural, socio-political and economic
environments impede immigrant and/or
refugee women in maintaining their physi-
cal and mental health, affecting them pro-
foundly at both family and individual lev-
els. Certain mainstream perceptions of
health, wellness and illness do not accom-
modate nor respect the cultural and reli-
gious beliefs of immigrant and/or refugee
women.7-10 Consequently, health interven-
tions should be formulated and imple-
mented in ways that respect these differ-
ences while maintaining a standard of
quality care. Culturally sensitive approach-
es must be developed for health and social
service professionals in specific situations.
Training for health professionals must pre-
pare and equip them with necessary
knowledge for providing quality care to a
multicultural population.
Language barriers also create distance
between the health system and immigrant
and/or refugee women. Illiterate women,
or those who speak neither English nor
French, are powerless to personally access
information on, as well as services from,
the health system. They remain dependent
upon others for information related to
their own bodies. Using family or commu-
nity members as interpreters is not always a
solution as taboo subjects (sexuality) may
be misinterpreted; professional interpreters
should be used to ensure the accuracy of
information.
Isolation and loss of pre-existing social
support systems also affect immigrant
and/or refugee women’s health. Early hos-
pital discharges (either after childbirth or
illness) might cause fatigue and/or further
ill health in immigrant women. In addi-
tion, refugee families have often survived
great pre-migratory losses, long family sep-
arations, ruptures and traumatic events due
to which their physical and mental health
have already become fragile.11-14 Refugee
women have often cited experiencing rape,
abduction, sexual abuse, harassment
and/or the obligation to grant sexual
favours in return for food or necessary legal
papers before or during their migration
processes, which exacerbate post-traumatic
stress disorders.3,11,13
Some sponsored women live happy and
healthy lives while others face controlling
and abusive husbands who threaten to
cease sponsorship, abandon or return them
to their country, ignoring Canadian law
and the status and rights of immigrant
women. Bound through multiple fears to
their husbands and extended family, anx-
ious for the well-being of their children,
immigrant women are forced to accept vio-
VOLUME 91, NO. 5

lence, leaving them bruised psychologically
and physically and with a feeling of moral
emptiness.*14,15 Furthermore, immigrant
and/or refugee women from visible minori-
ty groups suffer additional prejudice and
discrimination.3-5,12,14
The reality of the double day is rein-
forced for immigrant and/or refugee
women living in already tenuous situations
upon arrival into a new country. Many
immigrant and refugee women are vulnera-
ble to excessive workloads, low wages, con-
trol, devaluation and abuse within their
homes and places of work. Those that find
employment before their husband risk fac-
ing repercussions caused by the perception
of women’s work within traditional gender
role ideology. Further, they lose control of
their earnings, which are handed over to
the male head of the household.
Immigrant and/or refugee women in these
situations are more at risk of conjugal vio-
lence from a husband who perceives him-
self unable to maintain his traditional sta-
tus as the family breadwinner. Finally,
immigrant and/or refugee women might
also encounter unemployment due to a
lack of professional accreditation or educa-
tion, or language barriers. These factors
increase stress and depression and lead to
low self-esteem.7-9
Where are we? Where do we go from
here?
The Centres of Excellence for Women’s
Health Program (CEWHP) and the four
Metropolis Centres of Excellence that span
the country, work towards strategic inter-
ventions and research regarding women’s
health and migration respectively.
Metropolis and CEWHP researchers work-
ing together have filled some of the gaps in
research on immigrant and/or refugee
women. Much more remains to be done in
the area of immigration health research,
such as: 1) evidence for considering the
immigration experience itself as a health
determinant3 and 2) a need for more inter-
sectional analysis in this area of health
research.
* In refugee couples and families, the preceding
descriptions are also true, except that a husband
cannot threaten his wife with abandonment and
deportation since the family is considered a
refugee unit and receives protection under
Convention provisions.
SEPTEMBER – OCTOBER 2000
IMMIGRATION, WOMEN AND HEALTH IN CANADA
CONCLUSION
The growing immigrant and refugee
population of Canada has largely earned
research and policy attention in terms of
demographic studies and economic inte-
gration. Yet, the gradual loss of immigrant
health and well-being is related to multi-
dimensional risk factors that are unknown
prior to immigration. This lack of prepara-
tion erects hurdles with regard to the
maintenance of the physical and mental
health of immigrants and/or refugees.
Immigrant and/or refugee women face
additional barriers to maintaining their
health and well-being compared with
immigrant/refugee men and Canadian-
born women. Despite this well-documented
fact, immigrant and/or refugee women are
expected to change portions of their identi-
ty in order to “fit” into current health and
policy interventions. Policies and/or pro-
grams are designed for immigrants in gen-
eral rather than for immigrant women as a
specific group; immigrant and/or refugee
women’s culture, heritage and ethnicity are
necessarily compromised to fit into policies
and/or programs developed for Canadian
women, often treated as a homogeneous
group.
Women’s health strategies have been
developed without consideration of cultur-
al diversity, thereby hiding the faces of the
many immigrant and refugee women who
enrich the fabric of Canadian society.
Culturally appropriate strategies must be
developed to take into account the varied
experiences and difficulties immigrant and
refugee women face. Inappropriate models
and programs, designed for a non-immigrant
or male clientele, constitute an act of negli-
gence by the Canadian health system
whose mandate is to provide adequate care
and services to the entire population. In
erasing immigrant and refugee women
from agendas, policies and programs, we
impose detrimental alterations to their
identities and encourage a continued dete-
rioration of their health status.
CANADIAN JOURNAL OF PUBLIC HEALTH 395
REFERENCES
1. Samuel TJ, Schachhuber D. A portrait of
Canadian diversity. In: Nancoo S, Ramcharan S
(Eds.), Canadian Diversity 2000 and Beyond.
Mississauga: Canadian Educator’s Press, 1995.
2. Chen J, Wilkins R, Ng E. Health expectancy by
immigrant status, 1986 and 1991. Health Reports
1996;18(3):29-37.
3. Kinnon D. Canadian Research on Immigration
and Health: An Overview. Ottawa: Health
Canada, 1999.
4. Agnew V. Resisting Discrimination: Women from
Asia, Africa, and the Caribbean and the Women’s
Movement in Canada. Toronto: University of
Toronto Press, 1996.
5. Agnew V. In Search of a Safe Place: Abused
Women and Culturally Sensitive Services. Toronto:
University of Toronto Press, 1998.
6. Horne T, Donner L, Thurston WE. Invisible
Women: Gender and Health Planning in
Manitoba and Saskatchewan and Models for
Progress. Winnipeg: Prairie Women’s Health
Centre of Excellence, 1999.
7. Clarkson M, Dahan I. La famille comme protec-
tion. Enquête Santé Québec auprès de la com-
munauté du Magreb et du Moyen-Orient.
Rapport de la phase 1 (entrevues de groupe).
Montréal: Santé Québec, 1997.
8. Clarkson M, Eustache R. La santé c’est la
richesse. Enquête Santé Québec auprès de la
communauté haïtienne. Rapport de la phase 1
(entrevues de groupe). Montréal: Santé Québec,
1997.
9. Clarkson M, Nghi TT. Au confluent de deux
médicines. Enquête Santé Québec auprès de la
communauté chinoise. Rapport de la phase 1
(entrevues de groupe). Montréal: Santé Québec,
1997.
10. Spitzer S, Young D. More than disease:
Definitions of health and healing from new
Canadians. Notes from the Prairie Centre
2000;4(1):2-3, 8.
11. Bibeau G, Chan-Yip AM, Lock M, et al. In:
Gaëtan Morin (Ed.), La santé mentale et ses
visages : un Québec pluriethnique au quotidien.
Boucherville, QC: Gaëtan Morin, 1992.
12. Guindon N. Femmes réfugiées en attente de
statut. Montréal: Centre des femmes de
Montréal, 1995.
13. Johnsson AB. The international protection of
women refugees. In: Kelley N (Ed.), Working
with Refugee Women. Geneva: International
NGO Working Group on Refugee Women,
1989.
14. MacLeod L, Shin M. Isolées, apeurées et
oubliées : les services aux immigrantes et aux
réfugiées : besoins et réalités. Ottawa: Santé et
bien-être social, Division de la prévention de la
violence familiale, 1990.
15. Krane J, Oxman-Martinez J, Wehbi S.
Responding to conjugal violence across cultures:
Exploring an intersectional approach. Canadian
Ethnic Studies/Revue canadienne d’études ethniques
1999 (submitted).
Received: July 18, 2000
Accepted: September 7, 2000