Вы находитесь на странице: 1из 10

INT J TUBERC LUNG DIS 4(2):123–132

© 2000 IUATLD

Gender and tuberculosis: a comparison of prevalence


surveys with notification data to explore sex differences
in case detection

M. W. Borgdorff,* N. J. D. Nagelkerke,* C. Dye,† P. Nunn‡


* Royal Netherlands Tuberculosis Association, The Hague, The Netherlands; † Communicable Diseases Prevention
and Control, World Health Organization, ‡ Stop TB Initiative, World Health Organization, Geneva, Switzerland

SUMMARY

OBJECTIVE: To explore whether lower tuberculosis The female/male (F/M) prevalence ratios were less than
notification rates among women are due to a reduced 0.5 in surveys in the South-East Asia and Western Pacific
access to health care, particularly diagnostic services, for Region, and approximately 1 in the African Region.
women. C O N C L U S I O N : In most countries the F/M sex ratio in
M E T H O D S : Age- and sex-specific tuberculosis preva- prevalent cases was similar or lower than that in notified
lence rates of smear-positive tuberculosis were obtained cases, suggesting that F/M differences in notification
from tuberculosis prevalence surveys reported to the rates may be largely due to epidemiological differences
WHO or published in the literature. Age- and sex- and not to differential access to health care. However,
specific notification rates from the same countries in available data are limited as the prevalence surveys in
1996 were used. Africa were carried out many years ago, and in Asia
R E S U L T S : Prevalence data and notifications from 29 sur- notification rates may be distorted by a large private sec-
veys in 14 countries were used. Notification rates varied tor with deficiencies in notification.
strongly among countries, but the female/male ratio was K E Y W O R D S : tuberculosis; gender; prevalence surveys;
below 1 and decreased with increasing age in almost all. notification; case detection

ANNUALLY approximately 2 million people die per se or to differences in exposure as a consequence


from tuberculosis,1 yet effective chemotherapy could of differences in the societal roles of men and women,
cure a vast majority of cases and, in combination with i.e., to gender differences. Sometimes these differ-
high case detection rates, lead to eventual elimination ences are hard to distinguish. For example, observed
from the community of the role of Mycobacterium differences in progression from infection to disease4–6
tuberculosis as a major pathogen. This has been the may be due to differences in the balance between bio-
experience in many industrialised countries. To date, logical need for (micro) nutrients and the gender-
however, low detection rates and poor cure rates have specific access to these nutrients provided by society.
prevented a repetition of this achievement in many Evidence seems to suggest that women of repro-
developing countries.2 Understanding factors which ductive age have a higher disease progression than
impact on case detection and cure rates is therefore men.4 Thus, lower notification rates for women
key in effective disease control. One conspicuous fac- appear not to be due to a lower disease progression.
tor is gender: notification rates of tuberculosis are Hence, differences in notification rates appear to con-
higher in men than in women in most countries.3 stitute prima facie evidence for differences in access to
Differences in notification rates may reflect gender health care. Such differences, if real, would form a
differences in TB epidemiology, and/or gender differ- major source of gender inequity.
ences in access to care.4–6 If the latter possibility were For various diseases other than tuberculosis, gen-
true, this would need urgent redress. One way to der differences in access to health care have been
either corroborate or vitiate this hypothesis is by described.7,8 For tuberculosis, differences in access to
studying the epidemiology of TB in relation to gender. diagnosis and treatment are found in some, but not
Gender differences in tuberculosis epidemiology all, studies. For instance, in a recent study on the
may arise either as a consequence of differences in diagnosis of tuberculosis by sputum smear micros-
biological functioning, i.e., as a consequence of sex copy in Benin, Malawi, Nicaragua, and Senegal, it

Correspondence to: Dr Martien W Borgdorff, Royal Netherlands Tuberculosis Association, P O Box 146, 2501 CC The
Hague, The Netherlands. Tel: (31) 70 416 7222. Fax: (31) 70 358 4004. e-mail: martien.borgdorff@kncvtbc.nl
Article submitted 17 May 1999. Final version accepted 14 October 1999.
[A version in French of this article is available from the IUATLD Secretariat in Paris.]
124 The International Journal of Tuberculosis and Lung Disease

Table 1 List of prevalence surveys reviewed

Cases by Notification
Survey age and data Accept/
Country year(s) Scope sex* present? Reject Ref
AFRO
Ghana 1957 National Y Y A 12
Kenya Nairobi 1958–1959 Regional Y Y A 13
Kenya Not-Nairobi 1958–1959 Regional Y Y A 14
Gambia 1958–1959 National N R 15
Liberia 1959–1960 National N R 16
Malawi 1960 National N R 17
Mauritius 1957–1958 National Y N R 18
Mozambique 1961 Regional Y Y A 19
Nigeria 1973 Regional N R 20
South Africa (Bophuthatswana) 1980 Regional N R 21
South Africa (KwaZulu) 1974 Regional N R 22
South Africa (Transkei) 1982 Regional N R 23
South Africa (Transkei) 1977 Regional N R 24
South Africa (Transkei) 1972 Regional N R 25
Sierra Leone 1958 National Y Y A 26
Somalia 1956 National N R 27
Tanzania 1957 Regional Y Y A 28
Uganda 1958 National Y Y A 29
Zambia 1980 Regional N R 30
EMRO
Pakistan 1959–1961 National N R 31
Syria 1960 Regional Y Y A 32
Tunisia 1961 Regional Y N R 33
SEARO
Bangladesh 1973 National Y Y A 34
Bangladesh 1987 National Y Y A 35
India (South) rural 1950–1955 Regional Y Y A 36
India Tumkur 1960–1961 Regional Y Y A 37
India (South) 1961–1977 Regional Y Y A 38
India Chingleput 1968–1980 Regional Y Y A 39
India Tumkur 1972–1973 Regional Y Y A 40
India Kashmir 1978 Regional N R 41
India Bangalore 1979 Regional N R 42
India (South) 1981–1983 Regional Y Y A 43
India Bangalore 1984–1986 Regional Y Y A 44
India Review TB in workers 1996 Y Y R 45
Indonesia East Kalimantan 1989–1990 Regional N R 46
(continued)

was found that the number of sputum smears exam- alence of disease, and to what extent to gender-related
ined was similar among women and men.9 The higher differences in access to public TB care, by comparing
number of positive sputum smears among men TB prevalence data with TB notification data.
reflected a higher prevalence of tuberculosis among
suspected cases.9 Whether women and men had equal
METHODS
access, given the presence of symptoms, could not be
answered by the study, as it was not determined We restrict our attention to direct sputum smear-pos-
whether the presence of symptoms was similar among itive pulmonary tuberculosis, as sputum smear micros-
men and women. A study in Nepal comparing active copy is the most widely used and available diagnostic
and passive case-finding methods found a female-to- tool. Thus, tuberculosis always refers to smear-positive
male ratio of 1:1.2 in active tuberculosis case-finding tuberculosis unless stated otherwise.
and of 1:2.6 in self-referred patients, suggesting an
access problem among women.10 Another study in Data sources
Nepal found that special ‘microscopy camps’ bring- Notification rates of new smear-positive tuberculosis
ing diagnostic services closer to the population also cases in 1996 by age, sex, country and World Health
detected a significantly higher percentage of women Organization (WHO) region have been published.3
than did routine services.11 It seems that this issue Tuberculosis prevalence surveys were obtained from
merits further investigation. the published literature and from unpublished reports
This paper aims to determine to what extent gender- available at the WHO. The list of surveys identified is
related differences in tuberculosis notification rates can presented in Table 1. Surveys carried out in a part of
be attributed to gender-related differences in the prev- a country only were included and treated as if they
Gender and tuberculosis 125

Table 1 (continued)

Cases by Notification
Survey age and data Accept/
Country year(s) Scope sex* present? Reject Ref
SEARO
Indonesia West Nusa Tenggara 1991–1992 Regional N R 47
Indonesia West Java 1992–1993 Regional N R 48
Indonesia Oost Java 1992–1993 Regional N R 49
Indonesia West Sumatra 1994–1995 Regional N R 50
Myanmar 1972 National N R 51
Sri Lanka 1970–1971 National N R 52
WPRO
China Taiwan Province 1957–1958 Regional Y Y A 53
China 1990 National Y Y A 54
China 1984–1985 National N Y R 55
Japan 1953 National N R 56
Japan 1953 National N R 57
Japan 1953-1958 National N R 58
Japan 1954 National N R 59
Japan 1958 National N R 60
Japan 1963 National N R 61
Japan 1964 National N R 62
Republic of Korea 1965 National N R 63
Republic of Korea 1970 National Y Y R† 64
Republic of Korea 1975 National N R 65
Republic of Korea 1980 National N R 66
Republic of Korea 1985 National Y Y R† 67
Republic of Korea 1990 National Y Y A 68
Republic of Korea 1995 National Y Y A 69
Malaysia West 1970 Regional Y Y A 70
Philippines 1981–1983 National Y Y A 71
Philippines 1997 National Y Y A 72
Samoa 1975 National Y Y A 73
Vietnam 1962 Regional N R 74
AMRO
Colombia 1988 Regional N R 75
EURO
Turkey 1971 Regional Y N R 76
* Smear-positive cases only.
† From Korea only the latest two surveys were included (1990 and 1995), as the prevalence survey methodology was similar in all surveys (representative national

samples) and these two were nearest in time to the notification rates used (1996).
AFRO  African Regional Office; EMRO  Eastern Mediterranean Regional Office; SEARO  South-East Asian Regional Office; WPRO  Western Pacific
Regional Office; AMRO  American Regional Office; EURO  European Regional Office.

were national surveys. Surveys were excluded from 24, 25–34, 35–44, 45–54, 55–64, 65. Regrouping
analysis if they did not provide a breakdown of prev- was done using the assumption that the population
alence rates by age and sex. If no notification data examined and cases found were distributed evenly
were available by age and sex, these countries were over the age range reported. For instance, if the sur-
excluded from the analysis as well. For Korea only the vey reported results on the age group 20–29, numbers
two most recent (1990, 1995) surveys were included as were divided equally over the age groups 15–24 and
they were national and considered adequate for the 25–34. The age group below 15 years was excluded
prevalence situation near the time of notification. All as very few smear-positive cases are found at these
other surveys were accepted (Table 1). All included ages. As the age group of 65 years and older was fre-
surveys used statistical sampling methods (e.g., cluster, quently not covered by the prevalence surveys, this
multi-stage) to ensure that the sample would be repre- group was excluded to facilitate comparisons between
sentative for the population under consideration. countries.
Though prevalence surveys covered a long period
Data analysis (from the late 1950s to 1997), for comparison with
The ratio of TB notification rates among women and the notification data it was assumed that the relative
men was calculated by age group and summarised as age and sex distribution of prevalent cases did not
age-standardised rate ratios. Prevalence surveys change over this period. However, in order to assess
turned out to be highly heterogeneous in the age the effect of the HIV/AIDS epidemic on the sex ratio
groups used. In order to allow comparison by age and of tuberculosis cases, we explored changes of the sex
sex of results from prevalence surveys and notifica- ratios in notification data over the past 6–9 years in
tion data, prevalence survey results were re-grouped Kenya, Malawi, and Tanzania. Over this time period,
into the age groups used in the notification data: 15– there has been a strong increase in HIV prevalence
126 The International Journal of Tuberculosis and Lung Disease

and in the number of notified tuberculosis cases in

15–64

0.6
0.6
0.7
0.6
0.6
0.7

0.5

0.4
0.5

0.6
0.5
0.5
0.3
8.3
these three countries. All three countries have a well-
established tuberculosis surveillance system.

65

0.2
0.3
0.2
0.3
0.2
0.4

0.5

0.1
0.2

0.2
0.4
0.3
0.3
0.0
Notified incident cases derive from the population
of prevalent cases. Prevalent TB cases are diagnosed

55–64
and notified at a certain rate, which may vary by age

0.4
0.4
0.3
0.5
0.4
0.5

0.6

0.2
0.2

0.3
0.3
0.4
0.2
0.0
Female/Male ratio
and sex. We will refer to this rate as patient detection
rate (PDR). For those countries from which preva-

45–54

0.4
0.4
0.5
0.5
0.4
0.5

0.5

0.3
0.3

0.4
0.3
0.4
0.3
n.d
lence survey reports were available for analysis, we
estimated these (age- and sex-specific) patient detec-

35–44
tion rates as:

n.d.
0.5
0.4
0.7
0.7
0.5
0.6

0.5

0.4
0.3

0.6
0.3
0.4
0.3
PDR  [notification rate]/[prevalence rate].

25–34

n.d.
The PDR among females was divided by the corre-

0.7
0.6
0.9
0.6
0.7
0.8

0.4

0.5
0.5

0.8
0.5
0.5
0.4
sponding male PDR for each age group. The resulting
rate ratios were used as a measure for age-specific

15–24
Notification rates (per 100 000) of smear-positive tuberculosis by age, sex, and country in countries with prevalence survey report*

n.d.
0.8
0.9
0.9
0.8
0.9
1.2

0.5

0.7
0.7

0.9
0.9
0.7
0.4

AFRO  African Regional Office; EMRO  Eastern Mediterranean Regional Office; SEARO  South-East Asian Regional Office; WPRO  Western Pacific Regional Office.
gender-related differences in notification and by impli-
cation as evidence for differences in access to public

15–64
TB care.

123
17
86
89
72
97

12

13
19
31

18
6
1

0
To arrive at an overall (i.e., not age-specific) rate
ratio we used a crude patient detection rate ratio and
65

15
26
23
31
46
61

14

10
51
60
2
0

1
0
two standardised rate ratios obtained by direct and in-
direct standardisation. For direct standardisation we
55–64

120
17
53
60
49
81

17

13
18
47
6
0

0
0
used, arbitrarily, the population of India in 1991 as
the reference population to which notification rates
and prevalence rates were applied. For indirect stan-
45–54

138
19
69
90
61
87

14

12
13
42

40
7
0

1
dardisation, the expected annual number notified—
Female

E(ann)—from the prevalence survey sample was cal-


35–44

culated for each age group, using age-specific notifica-


113
103
110
149
23
79

12
13
29

29
9

7
1

0
tion rates in the total population. The gender-specific
total E(ann) was then divided by the total number of
25–34

125
112

135
158
21

93

10

15
20
29

30
8
1

0
prevalent cases observed in the prevalence survey to
arrive at a gender-specific PDR. The female-to-male
15–24

ratios of these PDRs were presented as indirectly


10
76
64
49
67
74

13

13
24
25
4
1

0
0
standardised age-adjusted rate ratios.
15–64

145
129
116
170
176
31

24

15

21
38
69
1

1
2

RESULTS
Reports of 69 prevalence surveys were obtained,12–76
65

107

186
163

114
238
66
84

93

29

23

39

67
1

* Source: reference 3. Combined age groups (15-64) are standardised for age.

36 of which provided a breakdown by age and sex


(Table 1). From five countries, no notification data
55–64

121
181

212
231

134
49

92

28

28

38
72

25
2

were available by age and sex, leaving 31 surveys


from 14 countries (Table 1). As only the two most
45–54

recent surveys were used for Korea, only 29 surveys


156
179
133
211
256

105
47

25

21

28
47
2

2
0

were actually used for analysis.


Male

Notification rates of smear-positive tuberculosis


35–44

179
171
154
243
248
44

17

20

22
39
74
2

1
0

per 100 000 population by age and sex in these 14


countries showed a large variation, ranging from 0.7
25–34

to 176 among men, and from 0.2 to 61 among


202
127
156
206
206
29

26

15

19
36
63
1

1
0

women (Table 2). The female/male ratio was less than


1 in all countries except Samoa. The female/male
15–24

notification rate ratio decreased with increasing age


71
13
82

63
72
64

25

15
27
38
5
1

0
0

in all countries except Syria (Table 2).


From Table 3 it appears that in Kenya, Malawi, and
Mozambique
Sierra Leone

Bangladesh

Philippines

Tanzania the HIV epidemic, although accounting for an


Tanzania

Malaysia
Age group

Uganda

overall substantial increase in TB case-load, has led to


Samoa
Ghana
Kenya

China
Korea
Table 2

India
Syria
SEARO

WPRO
EMRO

only a minor shift in the female/male notification ratio.


AFRO

Crude prevalence rates of tuberculosis varied


Gender and tuberculosis 127

Table 3 Gender specific notification rates per 100,000 population in Tanzania (1987 and 1996),
Kenya (1989 and 1996), and Malawi (1990 and 1996)*

Country
Age group Male Female F/M ratio Male Female F/M ratio
1989 1996
Kenya
15–24 37 26 0.7 82 76 0.9
25–34 89 44 0.5 202 125 0.6
35–44 95 42 0.4 171 79 0.5
45–54 99 59 0.6 179 69 0.4
55–64 105 55 0.5 181 53 0.3
65 78 34 0.4 107 26 0.2

1990 1996
Malawi
15–24 44 57 1.3 59 93 1.6
25–34 139 114 0.8 218 179 0.8
35–44 164 104 0.6 236 164 0.7
45–54 162 91 0.6 197 114 0.6
55–64 139 62 0.4 137 69 0.5
65 108 35 0.3 96 25 0.3

1987 1996
Tanzania
15–24 47 30 0.6 71 67 0.9
25–34 130 54 0.4 206 135 0.7
35–44 147 61 0.4 243 110 0.5
45–54 160 66 0.4 211 87 0.4
55–64 137 58 0.4 212 81 0.4
65 67 21 0.3 186 46 0.2
* Sources: WHO report for 1996 notification data. Annual reports of National Tuberculosis Programmes of Kenya,
Malawi and Tanzania for earlier (1990) notification data.

strongly between surveys, ranging from 0.5 to 16.2 applied in particular to India and the Philippines,
per 1000 among men and from 0.0 to 8.7 per 1000 which had PDRs of less than 1/100 (Table 5).
among women (Table 4). The crude female/male
prevalence ratio varied considerably between surveys,
DISCUSSION
which is likely to be due in part to the small sample
size in some of the surveys. When sex ratios are sum- A striking feature of TB, which is apparent in preva-
marised by WHO region, they are found to be lence surveys from most parts of the world, is that TB
approximately 0.3 in South-East Asia (SEARO: India is more a disease of men than of women. Thus, gender
and Bangladesh), approximately 0.5 in the Western is an important aspect of TB epidemiology. The differ-
Pacific region (WPRO: China, Korea, Malaysia, Phil- ence between men and women appears to be larger in
ippines, Samoa), and approximately 1.0 in sub- Asia, especially in the SEARO region and to a lesser
Saharan Africa (AFRO: Ghana, Kenya, Mauritius, extent in the WPRO region, than in sub-Saharan
Mozambique, Sierra Leone, Tanzania, Uganda) (Table Africa (AFRO region). It is unclear why male-female
4). Age-adjusted prevalence ratios were similar to the differences in tuberculosis prevalence appear to be
crude numbers. larger in many Asian surveys than in those from Africa.
On the assumption that notified cases arise at a Whether this reflects, for example, lower (re)infection
certain rate from prevalent cases, patient detection rates of Asian women due to more restricted participa-
rates (Table 5) and their confidence intervals (Figure) tion in society, or higher breakdown rates in African
were calculated. The (directly and indirectly) standar- women due to inadequate nutrition, is unknown. To
dised female/male ratios of PDRs were well above 1 answer these questions the role of gender in TB epide-
in India (SEARO). Thus, female cases appeared more miology needs to be better studied. Epidemiological
likely to be notified than male cases. The rate ratio field studies looking at potentially confounding vari-
was somewhat higher than 1 in WPRO and below 1 ables related to exposure and immunity may give clues.
in various countries in AFRO. If the TB situation is This study suggests that overall gender differences
steady, and if all cases were eventually notified, the in TB notification rates may be largely due to gender
PDR would be the inverse of the duration of illness differences in TB prevalence. If anything, female TB
before diagnosis. A PDR of 50/100 would thus corre- cases are more likely to be notified than male cases,
spond to a duration of illness of 2 years. In many except perhaps in the AFRO region, where some
countries these PDRs were well below 50/100, imply- countries have PDRs smaller than 1. It is however dif-
ing that many cases may not have been notified. This ficult to make definitive statements on this region as
128 The International Journal of Tuberculosis and Lung Disease

Table 4 Tuberculosis prevalence rates in the age group 15–64 years by age and sex

F/M prevalence
Prev/1000 Prev/1000 ratio
Year
survey Male Age- Female Age- Age-
Country completed cases Population Crude adjusted cases Population Crude adjusted Crude adjusted Ref
AFRO 57 12 985 4.4 45 10 496 4.2 0.97
Ghana 1958 5 912 5.5 5.4 3 1016 3.0 3.3 0.54 0.62 12
Kenya 1959 6 5976 1.0 1.0 7 1809 3.9 3.9 3.85 3.94 13
Kenya 1961 17 1605 10.6 10.8 13 2404 5.2 5.3 0.49 0.49 14
Mozambique 1961 15 895 16.8 16.7 6 1352 4.4 4.7 0.26 0.28 19
Sierra Leone 1958 4 1040 3.8 3.0 3 1237 2.4 2.8 0.63 0.93 26
Tanzania 1960 5 847 5.9 5.3 3 995 3.0 3.0 0.51 0.57 28
Uganda 1958 5 1710 2.9 3.0 10 1683 5.9 5.3 2.03 1.73 29
EMRO 1 557 1.8 0 623 0.0 0.00
Syria 1960 1 557 1.8 1.7 0 623 0.0 0.0 0.00 0.00 32
SEARO 1475 159 614 9.2 447 156 352 2.9 0.31
Bangladesh 1973 24 7265 3.3 3.3 3 5879 0.5 0.5 0.15 0.15 34
Bangladesh 1987 189 13 716 13.8 11.8 92 10 783 8.5 10.0 0.62 0.84 35
India 1950 60 7470 8.0 8.0 21 6555 3.2 3.3 0.40 0.41 36
India 1960 54 8684 6.2 6.1 22 8425 2.6 2.6 0.42 0.43 37
India 1977 31 3653 8.5 8.2 5 3468 1.4 1.5 0.17 0.18 38 (1)
India 1977 30 3373 8.9 8.4 7 3263 2.1 2.2 0.24 0.26 38 (2)
India 1977 18 3285 5.5 5.2 7 3120 2.2 2.3 0.41 0.44 38 (3)
India 1977 22 3217 6.8 6.3 7 3155 2.2 2.2 0.32 0.34 38 (4)
India 1977 25 2815 8.9 6.8 10 3030 3.3 2.8 0.37 0.42 38 (5)
India 1969 856 79 471 10.8 9.9 216 82 168 2.6 2.5 0.24 0.25 39
India 1973 66 10 300 6.4 5.7 28 9929 2.8 2.9 0.44 0.50 40
India 1983 30 7422 4.0 4.0 12 7764 1.5 1.5 0.38 0.38 43
India 1986 70 8943 7.8 6.7 17 8813 1.9 1.7 0.25 0.25 44
WPRO 1128 481 635 2.3 589 491 894 1.2 0.51
China–Taiwan 1958 161 12 450 12.9 12.4 60 13 689 4.4 4.4 0.34 0.35 53
China 1990 765 410 000 1.9 1.8 441 410 000 1.1 1.0 0.58 0.59 54
Korea 1990 44 16 139 2.7 2.3 13 18 565 0.7 0.7 0.26 0.30 59
Korea 1995 28 21 728 1.3 1.0 13 25 218 0.5 0.5 0.40 0.49 60
Malaysia 1970 20 7184 2.8 2.7 5 7795 0.6 0.7 0.23 0.24 61
Philippines 1983 69 4570 15.1 16.0 40 5350 7.5 7.3 0.50 0.45 62
Philippines 1997 36 4876 7.4 7.4 11 5223 2.1 1.9 0.29 0.26 63
Samoa 1975 5 4688 1.1 1.1 6 6054 1.0 1.0 0.93 0.89 64
AFRO  African Regional Office, EMRO  Eastern Mediterranean Regional Office, SEARO  South-East Asian Regional Office, WPRO  Western Pacific
Regional Office.

all surveys from AFRO region were small in terms of gender inequities in access to this form of health care
the number of TB cases observed. Thus, generally, our are ‘invisible’ in analyses such as ours that make use
findings do not seem to support the hypothesis that of notification data as a measure of access to health
gender differences in notification rates are due to care. Therefore, if Asian women tend to seek care in
gender inequalities in access to health care. However, the public sector and Asian men in the private sector,
this conclusion cannot be generalised to individual then our results are biased. Fourth, the use of data
regions or countries, as in some countries PDRs are from prevalence surveys may also cause another form
not inconsistent with values 1. of bias. If the average duration of smear positivity in
This does not imply the absence of gender inequal- men is longer than in women, for example because
ities in access to health care. First, our findings do not they have higher defaulting rates, then prevalence
disprove gender inequalities in access to health care in may inadequately reflect the incidence of new cases,
some countries. Unfortunately, the quality of the data i.e., those eligible for notification.
does not allow firm conclusion to be drawn about A limitation of many prevalence surveys is that
individual countries. Second, access to notification they were carried out many years ago. If TB preva-
(i.e., diagnosis) does not necessarily imply access to lence has changed over time, but not equally so for
effective treatment. Third, our analysis only takes both sexes, then our approach is fallible. This limita-
into account notified incident cases. In many coun- tion applies in particular to surveys in Africa which
tries the private sector treats a substantial proportion were carried out before the widespread introduction
of TB cases and may not notify many cases that come of chemotherapy. Moreover, some of these countries
to its attention, in particular if notification is not have a substantial HIV burden, where typically HIV
mandatory. Women may have less access to this form prevalence among young women is higher than among
of health care, especially where it is expensive. Thus, men of the same age. However, data from Kenya,
Gender and tuberculosis 129

Table 5 Age-standardised notification rates, prevalence rates, and patient detection rates (PDR) in the age group 15–64 years
in 29 prevalence surveys from 14 countries

Notification rates Prevalence rates PDRs


(1996) per 100 000 per 1000 per 100 prevalent
population population cases F/M ratio of PDRs
Country Male Female Male Female Male Female Ref Direct Indirect
AFRO
Ghana 30.6 17.5 5.4 3.3 5.7 5.2 12 0.92 1.08
Kenya 144.9 86.2 1.0 3.9 147.9 22.3 13 0.15 0.16
Kenya 144.9 86.2 10.8 5.3 13.4 16.4 14 1.22 1.24
Mozambique 128.6 88.8 16.7 4.7 7.7 19.1 19 2.47 2.60
Sierra Leone 116.3 72.4 3.0 2.8 38.4 25.7 26 0.67 0.97
Tanzania 170.1 96.7 5.3 3.0 32.0 31.9 28 1.00 1.08
Uganda 176.0 122.6 3.0 5.3 58.1 23.4 29 0.40 0.34
EMRO
Syria 24.2 12.1 1.7 0.0 14.7 Undefined 32 Undefined Undefined
SEARO
Bangladesh 14.7 6.1 3.3 0.5 4.4 12.1 34 2.73 2.74
Bangladesh 14.7 6.1 11.8 10.0 1.2 0.6 35 0.49 0.63
India 1.4 0.6 8.0 3.3 0.2 0.2 36 1.13 1.14
India 1.4 0.6 6.1 2.6 0.2 0.2 37 1.07 1.10
India 1.4 0.6 8.2 1.5 0.2 0.4 38 (1) 2.57 2.61
India 1.4 0.6 8.4 2.2 0.2 0.3 38 (2) 1.75 1.84
India 1.4 0.6 5.2 2.3 0.3 0.3 38 (3) 1.05 1.09
India 1.4 0.6 6.3 2.2 0.2 0.3 38 (4) 1.34 1.36
India 1.4 0.6 6.8 2.8 0.2 0.2 38 (5) 1.10 1.09
India 1.4 0.6 9.9 2.5 0.1 0.3 39 1.84 1.81
India 1.4 0.6 5.7 2.9 0.2 0.2 40 0.92 1.02
India 1.4 0.6 4.0 1.5 0.3 0.4 43 1.21 1.18
India 1.4 0.6 6.7 1.7 0.2 0.4 44 1.87 1.75
WPRO
China–Taiwan 21.1 13.3 12.4 4.4 1.7 3.1 53 1.79 1.84
China 21.1 13.3 1.8 1.0 12.0 12.8 54 1.06 1.06
Korea 38.5 18.7 2.3 0.7 16.9 27.1 68 1.60 1.76
Korea 38.5 18.7 1.0 0.5 37.4 37.4 69 1.00 1.05
Malaysia 68.7 31.0 2.7 0.7 25.1 47.8 70 1.89 1.96
Philippines 1.1 0.3 16.0 7.3 0.1 0.0 71 0.73 0.69
Philippines 1.1 0.3 7.4 1.9 0.1 0.2 72 1.28 1.23
Samoa 2.2 18.5 1.1 1.0 2.1 19.1 73 9.24 4.16
AFRO  African Regional Office, EMRO  Eastern Mediterranean Regional Office, SEARO  South-East Asian Regional Office, WPRO  Western Pacific
Regional Office.

Malawi and Tanzania suggest that, despite substan- Another limitation is the variable size and quality of
tial increases in notification rates over the past decade, TB surveys. Some surveys were very small and detected
the female/male notification ratio has remained few prevalent cases. Most were carried out regionally,
largely unaltered. whereas we compared them to national notification

Figure Female/male ratio of patient detection rates. Indirect standardisation for age.
130 The International Journal of Tuberculosis and Lung Disease

data. While the instrument of a prevalence survey may 5 Hudelson P. Gender differentials in tuberculosis: the role of so-
be the best available for estimating the TB disease bur- cio-economic and cultural factors. Tubercle Lung Dis 1996;
77: 391–400.
den, it may not be entirely free of gender bias if, for 6 Diwan V K, Thorson A, Winkvist A, eds. Gender and tuber-
example, women with symptoms manage to avoid culosis: an international research workshop. Göteborg, Swe-
being screened by the survey from fear of being diag- den: Nordic School of Public Health, 1998.
nosed as a TB patient. An even stronger limitation of 7 Ojanunga D N, Gilbert C. Women’s access to health care in de-
TB surveys is that they have only been carried out in a veloping countries. Soc Sci Med 1992; 35: 613–617.
8 Puentes-Markides C. Women and access to health care. Soc Sci
limited number of countries. Clearly, this study could
Med 1992; 35: 619–626.
not assess gender-related problems of access to public 9 Rieder H L, Arnadottir Th, Tardencilla Gutierrez A A, et al.
TB care in countries that had not carried out a preva- Evaluation of a standardized recording tool for sputum smear
lence survey. Finally, our analysis was restricted to microscopy for acid-fast bacilli under routine conditions in low
smear-positive tuberculosis cases and did not assess income countries. Int J Tuberc Lung Dis 1997; 1: 339–345.
10 Cassels A, Heineman E, LeClerq S. Tuberculosis case-finding in
possible gender inequities in access to care for smear-
Eastern Nepal. Tubercle 1982; 63: 175–185.
negative or extra-pulmonary TB. 11 Harper I, Fryatt R, White A. Tuberculosis case finding in
In view of these limitations, additional research remote mountainous areas – are microscopy camps of any
would be instrumental in obtaining firmer conclu- value?—Experience from Nepal. Tubercle Lung Dis 1996; 77:
sions on gender bias in tuberculosis case detection. 384–388.
Gender bias may be related to health seeking behav- 12 World Health Organization, Tuberculosis Research Office.
Tuberculosis Surveys in Ghana. Copenhagen: World Health
iour and/or to steps within the health systems in sus- Organization, 1958.
pecting and diagnosing TB. Gender bias within the 13 WHO/UNICEF-assisted tuberculosis project. Tuberculosis
health system may be examined rapidly by comparing Survey of Nairobi. Nairobi: World Health Oganization/
sex- and age-specific 1) prevalence of respiratory UNICEF, 1959.
complaints among out-patients, 2) probabilities that 14 World Health Organization Regional Office for Africa. A
Tuberculosis Survey in Kenya. Brazzaville: World Health
TB suspects actually undergo sputum examination, 3) Organization Regional Office for Africa, 1961.
probabilities that TB patients being diagnosed are put 15 World Health Organization Regional Office for Africa. Tuber-
on treatment, and 4) cure rates. Such research could culosis Survey in Gambia. Brazzaville: WHO Regional Office
be combined with social science research on perceived for Africa, 1960.
gender-related barriers among patients and staff. 16 World Health Organization Regional Office for Africa. A
Tuberculosis Survey in Liberia. Brazzaville: WHO Regional
Gender-related problems of access to health care
Office for Africa, 1961.
are more difficult to assess and may need a combina- 17 World Health Organization Regional Office for Africa. Tuber-
tion of epidemiological research (population-based culosis Survey in Nyasaland. Brazzaville: WHO Regional Of-
prevalence survey) and social science research on fice for Africa, 1961.
health seeking behaviour. Alternatively, at the time of 18 World Health Organization Regional Office for Africa. Tuber-
culosis Survey in Mauritius. Brazzaville: WHO Regional Office
diagnosis, smear-positive patients could be requested
for Africa, 1959.
to bring in all other household members for sputum 19 World Health Organization Regional Office for Africa. A
smear examination. In the absence of gender bias, one Tuberculosis Survey in Mozambique. Brazzaville: WHO Re-
would expect the female/male ratio among smear- gional Office for Africa, 1962.
positive household members to be approximately 20 Pust R E, Onejeme S E, Okafor S N. Tuberculosis Survey in
equal to that among index patients. This may apply in East Central State, Nigeria: Implications for Tuberculosis Pro-
gramme Development. Trop Geogr Med 1974; 26: 51–57.
particular to spouses, who are likely to share possibly 21 Gatner E M S. Active and passive tuberculosis case-finding in
confounding characteristics (such as approximate Bophuthatswana. TBNI-TBRI Bull 1980; 1: 4–13.
age) with the index patient. Thus, if in the majority of 22 Arabin G, Gärtig D, Kleeberg H H. First tuberculosis preva-
cases the index patient in double-positive couples is lence survey in KwaZulu. South Afr Med J 1979; 56: 434–438.
the husband, gender bias, either in access to health 23 Fourie P B, Zeelie S, Mouton L, Lancaster J. Third survey of tu-
berculosis in Transkei, 1982. Pretoria: Tuberculosis Research
care or within the health system, can be suspected. Institute of the South Africa Medical Research Council, 1984.
24 Fourie P B, Gatner E M S, Glatthaar E, Kleeberg H H. Follow-
up Tuberculosis Prevalence Survey of Transkei. Tubercle 1980;
References 61: 71–79.
1 Murray C J L, Lopez A D. Mortality by cause for eight regions 25 South African Tuberculosis Study Group. Tuberculosis in the
of the world: global burden of disease study. Lancet 1997; 349: Transkei, an epidemiological survey. South Afr Med J 1974;
1269–1276. 48: 149–161.
2 Dye C, Garnett G P, Sleeman K, Williams B G. Prospects for 26 WHO Regional Office for Africa, WHO Tuberculosis Re-
worldwide tuberculosis control under the WHO DOTS strat- search Office. Tuberculosis Surveys in Sierra Leone. Brazza-
egy. Lancet 1998; 352: 1886–1891. ville: WHO, 1958.
3 World Health Organization. Global Tuberculosis Programme. 27 World Health Organization Tuberculosis Research Office.
Global tuberculosis control. WHO report 1998. WHO/TB/98- Tuberculosis Survey in the Somalilands. Copenhagen: WHO
237. Geneva: World Health Organization, 1998. Tuberculosis Research Office, 1956.
4 Holmes C B, Hausler H, Nunn P. A review of sex differences in 28 World Health Organization Tuberculosis Research Office. A
the epidemiology of tuberculosis. Int J Tuberc Lung Dis 1998; Tuberculosis Survey in Tanganyika. Copenhagen: WHO
2: 96–104. Tuberculosis Research Office, 1958.
Gender and tuberculosis 131

29 World Health Organization Tuberculosis Research Office. 47 Dinas Kesehatan Daerah Tingkat I Nusa Tenggra Barat. Lapo-
Tuberculosis Survey in Uganda. Copenhagen: WHO Tuber- ran Survei Prevalensi TB Paru Propinsi Nusa Tenggara Barat
culosis Research Office, 1959. Tahun 1991/1992. (Smear and Prevalence Survey in West Nusa
30 Watts T. Tuberculosis Infection and Disease in Three Popula- Tenggara). Tahun: Dinas Kesehatan Daerah Tingkati Nusa Teng-
tions in Zambia. Epidemiology and control of diseases in Af- gra Barat, 1992.
rica, Addis Ababa: Second African Regional Conference of the 48 Kantor Wilayah Depkes/Dinkes Provinsi Jawa Barat. Laporan
International Epidemiological Association, 1983. Hasil Pelaksanaan Survei Penyakit TB. Paru di Provinsi Jawa
31 Directorate of Tuberculosis Control (health division), Govern- Barat Tahun Anggaran 1992/1993. (Smear and Prevalence Sur-
ment of Pakistan. Report of the Tuberculosis Survey in Kara- vey in West Java). Bandung: Kantor Wilayah Depkes/Dinkes
chi, Rawalpindi & Lahore Division of West Pakistan. Proninsi Jawa Barat, 1993.
Directorate of Tuberculosis Control (health division), Govern- 49 Kantor Wilayah Departemen Kesehatan Dan Dinas Kesehatan
ment of Pakistan, 1962. Daerah Propinsi Dati I Jawa Timur.Laporan. Sigi Prevalensi
32 WHO Regional Office for the Eastern Mediterranean, Epide- TB Paru Menular Propinsi Jawa Timur Tahun 1992/1993.
miological and Statistical Centre. Tuberculosis Survey in the (Smear and Prevalence Survey in East Java). Kantor Wilayah
Syrian Arab Republic. Alexandria: WHO Regional Office for Departemen Kesehatan Dan Dinas Daerah Propinsi Dati I
the Eastern Mediterranean, WHO Epidemiological and Statis- Jawa Timur, 1993.
tical Centre, 1962. 50 Kantor Wilayah Departemen Kesehatan Dati I. Provinsi
33 WHO Regional Office for the Eastern Mediterranean, Epide- Sumatera Barat. Survei Prevalensi Penyakit Tuberkulosa Paru
miological and Statistical Centre. Tuberculosis Survey in Tuni- Se Sumatera Barat Tahun 1994/1995 (Smear and Prevalence
sia. Alexandria: WHO, 1961. Survey in west Sumatra). Padang: Kantor Wilayah Departemen
34 National Tuberculosis Control and Research Project, Govern- Kesehatan R.I. Dinas Kesehatan Dati I Provinsi Sumatera
ment of the People’s Republic of Bangladesh. Report of the Barat, 1995.
Tuberculosis Survey in Bangladesh. Dacca: WHO, 1973. 51 Ministry of Health. Tuberculosis base-line survey, Burma
35 Directorate General of Health Services, Office of the Director 1972. Rangoon: Ministry of Health, 1972.
TB & Leprosy Control Services. Report on the National Prev- 52 Radkovsky J. and other members of the team. Report on tuber-
alence Survey in Bangladesh, 1987–88. Dhaka: National TB culosis baseline survey in Sri Lanka by the tuberculosis training
Control Project Building, 1989. and evaluation team WHO project: SEARO 0113. WHO,
36 Frimodt-Moller J., A community-wide tuberculosis study in a 1973.
South Indian rural population, 1950–1955. Bull World Health 53 WHO Regional Office for the Western Pacific. Quo S K, Statis-
Organ 1960; 22: 61–170. tical Summary Report on a Sample Survey of Tuberculosis
37 Narain R A J, Geser A, Jambunathan M V, Subramanian M. Prevalence in Taiwan July 1957–June 1958. Manila: WHO,
Some aspects of a tuberculosis prevalence survey in a South In- 1959.
dian district. Bull World Health Organ 1963; 29: 641–664. 54 The Ministry of Public Health of the People’s Republic of
38 Chakraborty A K, Singh H, Srikantan K, Rangaswamy K R, China. Nationwide Random Survey for the Epidemiology of
Krishnamurthy M S, Stephen J A. Tuberculosis in a rural pop- Tuberculosis in 1990. Beijing: The Ministry of Public Health of
ulation of South India: report on five surveys. Indian J Tuberc the People’s Republic of China, 1993.
1982; 29: 153–167. 55 The Ministry of Public Health of the People’s Republic of China.
39 Tuberculosis Prevention Trial. Trial of BCG Vaccines in South Nationwide Random Survey for the Epidemiology of Tuber-
India for Tuberculosis Prevention. Indian J Med Res 1980; 72: culosis in 1984/85. Beijing: The Ministry of Public Health of
1–74. the People’s Republic of China, 1986.
40 National Tuberculosis Institute. Gothi G D, Chakraboty A K, 56 Ministry of Health and Welfare, Japanese Government. Pre-
Nair S S, Ganapathy K T, Banerjee G C. Prevalence of tuber- liminary Report on the Tuberculosis Prevalence Survey in Japan
culosis in a South Indian District - twelve years after initial sur- in 1953. Tokyo, Ministry of Health and Welfare, 1953.
vey. Indian J Tuberc 1979; 29: 121–135. 57 Yamaguchi M. Survey of tuberculosis prevalence in Japan,
41 Mayurnath S, Anantharaman D S, Baily G V J, et al. Tuber- 1953. Bull World Health Organ 1955; 13: 1041–1073.
culosis prevalence survey in Kashmir Valley. Indian J Med Res 58 Omura T, Oka H, Kumabe H, Kobayashi A. The trend of
1984; 80: 129–140. tuberculosis in Japan during the period 1953–1958, compari-
42 Chakraborty A K, Gothi G D, Isaac B, Rangaswamy K R, son of the results of tuberculosis prevalence surveys. Bull
Krishnamurthy M S, Rajalakshmi R. Chest diseases and tuber- World Health Organ 1962; 26: 19–45.
culosis in a slum community and problems in estimating their 59 Yamaguchi M, Oka H, Kumabe H, Yosano H. Survey of tuber-
prevalence. Indian J Public Health 1979; 23: 88–99. culosis prevalence in Japan: trends in tuberculosis from 1953
43 Ray D, Abel R. Incidence of smear-positive pulmonary tuber- to 1954. Bull World Health Organ 1959; 21: 145–159,1954.
culosis from 1981-83 in a rural area under an active health 60 The Ministry of Health and Welfare, Japanese Government.
care programme in South India. Tubercle Lung Dis 1995; 76: Survey of Tuberculosis Prevalence in Japan, 1958. Tokyo: Min-
191–195. istry of Health and Welfare, 1958.
44 Chakraborty A K, Suryanarayana H V, Krishna Murthy V V, 61 Wakamatsu E. Tuberculosis prevalence survey in Japan in
Krishna Murthy M S, Shashidhara A N. Prevalence of tuber- 1963. Kekkaku 1964; 7.8: 257–265.
culosis in a rural area by an alternative survey method without 62 Konishi H. Tuberculosis prevalence survey in Japan in 1964.
prior radiographic screening of the population. Tubercle Lung Kekkaku 1965; 10: 445–452.
Dis 1995; 76: 20–24. 63 Ministry of Health & Social Affairs, Korean National Tuber-
45 Chakraborty A K. Prevalence and incidence of tuberculous in- culosis Association Report on the Tuberculosis Prevalence Sur-
fection and disease in India: a comprehensive review. Banga- vey in Korea, 1965. Seoul: Ministry of Health & Social Affairs,
lore: WHO, 1996. Korean National Tuberculosis Association, 1965.
46 Kantor Wilayah/Dinas Kesehatan Propinsi Kalimantan Timur 64 Ministry of Health & Social Affairs, Korean National Tuber-
Samarinda. Laporan Hasil Pelaksanaan Survei Prevalensi Pen- culosis Association. Report on the 2nd Tuberculosis Prevalence
yakit TB Paru di Propinsi Kalimantan Timur Tahun Anggaran Survey in Korea 1970. Seoul: Ministry of Health & Social Af-
1989/1990 (Smear and Prevalence Survey in East Kalimantan). fairs, Korean National Tuberculosis Association, 1970.
Samarinda: Kantor Wilayah/Dinas Kesehatan Propinsi Kali- 65 Ministry of Health & Social Affairs, Korean National Tuber-
mantan Timur, 1990. culosis Association. Report on the 3rd Tuberculosis Prevalence
132 The International Journal of Tuberculosis and Lung Disease

Survey in Korea, 1975. Seoul: Ministry of Health & Social 71 National Institute of Tuberculosis. National Tuberculosis Prev-
Affairs, Korean National Tuberculosis Association, 1975. alence Survey in the Republic of the Philippines, 1981–1983.
66 Ministry of Health & Social Affairs, Korean National Tuber- Manila: Ministry of Health, 1989.
culosis Association. Report on the 4th Tuberculosis Prevalence 72 Tropical Disease Foundation. Final Report 1997 National
Survey in Korea, 1980. Seoul: Ministry of Health & Social Tuberculosis Prevalence Survey, Philippines. Makati City:
Affairs, Korean National Tuberculosis Association, 1980 Tropical Disease Foundation, 1997.
67 Ministry of Health & Social Affairs, the Korean National 73 WHO Regional Tuberculosis Control Team and WHO Lep-
Tuberculosis Association. The 2nd Report on the 5th Tuberculo- rologist. Report on the Tuberculosis/Leprosy Prevalence Survey
sis Prevalence Survey in Korea. Seoul: Ministry of Health & So- in Samoa, 26 June–24 October 1975. Western Samoa: WHO,
cial Affairs, Korean National Tuberculosis Association, 1985. 1977.
68 Ministry of Health & Social Affairs, Korean National Tuber- 74 World Health Organization Regional Office for the Western
culosis Association. Report on the 6th Tuberculosis Prevalence Pacific. A Tuberculosis Prevalence Survey in Saigon, Vietnam
Survey in Korea.1990. Seoul: Ministry of Health and Social 1962. Manila: WHO, 1963.
Affairs, Korean National Tuberculosis Association, 1990. 75 Zuluaga L, Betancur C, Abaunza M, Londoño J. Prevalences
69 Ministry of Health & Welfare, Korean National Tuberculosis of tuberculosis and other respiratory diseases among people
Association. Report on the 7th Tuberculosis Prevalence Survey over age 15 in the north-east sector of Medellín, Colombia.
in Korea, 1995. Seoul: Ministry of Health & Welfare, Korean Bull PAHO 1992; 26: 247–255.
National Tuberculosis Association, 1995. 76 Atlamaz T. A study of mass X-ray survey in Istanbul. Istanbul:
70 Regional Tuberculosis Advisory Team WHO/WPRO. Chia M, Tevfik Saalam International Antituberculosis Training and
Huang J J. Report on a prevalence survey in West Malaysia Demonstration Center of Istanbul, 1971.
1970. Malaysia: World Health Organization, 1971.

RÉSUMÉ

O B J E C T I F : Explorer dans quelle mesure les taux de ratios F/M de prévalence furent inférieurs à 0,5 dans
déclaration plus bas de tuberculose observés chez les les enquêtes d’Asie du Sud-est et de la Région du Paci-
femmes sont dus à une accessibilité réduite des soins de fique Occidental ; ils sont approximativement de 1 dans
santé et en particulier des services de diagnostic. la Région Afrique.
M E T H O D E S : Les taux de prévalence de tuberculose C O N C L U S I O N : Dans la plupart des pays, les ratios de
spécifiques pour l’âge et le sexe et concernant les tuber- sexe F/M des cas prévalents sont similaires ou plus bas
culoses à bacilloscopie positive ont été obtenus à partir que ceux des cas déclarés, ce qui suggère que les dif-
d’enquêtes de prévalence de la tuberculose signalées à férences F/M dans les taux de déclaration peuvent être
l’OMS ou publiées dans la littérature. Les taux de décla- dues largement à des différences épidémiologiques et
ration spécifiques pour l’âge et le sexe pour les mêmes non à un accès différentiel aux services de santé. Toute-
pays en 1996 ont été utilisés. fois, les données disponibles sont limitées, puisque les
R É S U L T A T S : On a utilisé les données de prévalence et de enquêtes de prévalence en Afrique ont été conduites il y
déclaration provenant de 29 enquêtes dans 14 pays. Les a de nombreuses années, et puisqu’en Asie les taux de
taux de déclaration varient fortement entre pays, mais le déclaration peuvent être perturbés par l’existence d’un
ratio femmes/hommes (F/M) est inférieur à 1 et diminue vaste secteur privé où les déclarations sont déficientes.
avec les progrès de l’âge dans presque tous les pays. Les

RESUMEN

O B J E T I V O : Investigar si la notificación más baja de aumento de la edad en casi todos. Las relaciones de pre-
tuberculosis en las mujeres es debida a un acceso redu- valencia mujer/hombre (M/H) eran inferiores a 0,5 en
cido a los centros de salud, en especial a los servicios de encuestas en el Sud-Este Asiático y en la región del Pacífico
diagnóstico. Occidental y aproximadamente 1 en la Región Africana.
M É T O D O S : Se obtuvieron las tasas de prevalencia de C O N C L U S I Ó N : En la mayoría de los países la relación
tuberculosis, sexo y edad, de los tuberculosos con esputo M/H en los casos prevalentes fue similar o inferior a la
positivo obtenidas de las encuestas de tuberculosis de los casos notificados, sugiriendo que las diferencias
comunicadas a la OMS o publicadas en la literatura. Se M/H en las tasas notificadas son más bien debidas a
utilizaron las tasas de notificación de edad y sexo de los diferencias epidemiológicas que a diferencias en el
mismos países, de 1996. acceso a los centros de salud. Sin embargo, los datos
R E S U L T A D O S : Se usaron los datos de prevalencia y noti- accesibles son limitados y las encuestas de prevalencia en
ficaciones de 28 encuestas en 14 países. Las tasas de noti- Africa fueron llevadas a cabo hace muchos años y las
ficación variaron fuertemente entre los países, pero la re- tasas de notificación en Asia pueden ser influenciadas por
lación mujer/hombre fue inferior a 1, y disminuyó con el un gran sector privado con deficiencias en la notificación.

Вам также может понравиться