Relative importance of diagnostic delays in different head
ORIGINAL ARTICLE
and neck cancers
Teppo, H.* & Alho, O.-P.†
*Department of Otorhinolaryngology, Kanta-Häme Central Hospital, Hämeenlinna, Finland, and  Department of
Otorhinolaryngology, University of Oulu, Oulu, Finland
Accepted for publication 7 March 2008
Clin. Otolaryngol. 2008, 33, 325–330
Objectives: Clinical stage at the time of diagnosis is the
most important determinant of prognosis in head and
neck cancer. Previously, longer diagnostic delay has been
shown to worsen prognosis in cancer of tongue, pharynx
and larynx. The aim of this study was to evaluate the
relative importance of patient and professional diagnostic
delays in the prognosis of these head and neck cancers.
Design: Population-based retrospective cohort study.
Setting: Oulu University Hospital (tertiary referral
centre) district, Northern Finland.
Participants: Population-based cohort of 221 patients
with tongue, pharyngeal or laryngeal cancer diagnosed in
1986–1996.
Main outcome measures: Patient and professional
diagnostic delays, overall survival.
Even though clinical stage at the time of diagnosis is
recognised as the most important prognostic factor in
head and neck cancer,1 population screening – general or
directed to people with known risk factors – to find un-
symptomatic head and neck cancers in earlier stage is nei-
ther in routine use nor supported by literature.2 Thus,
the diagnosis is usually made after medical consultation
of the patient because of emerging symptoms. This
involves delays caused by both the patient and the physi-
cian. Previously, diagnostic delays have been shown to
have prognostic significance in certain head and neck
cancers: longer professional delay is an independent
determinant of poor prognosis in laryngeal,3 and longer
patient delay in pharyngeal cancer.4 Longer patient delay
has also been associated with more advanced stage at
presentation.5–7 In cancer of the mobile tongue, the nega-
tive impact of longer professional delay has been proved
Correspondence: Heikki Teppo, MD, Department of Otorhinolaryngol-
ogy, Kanta-Häme Central Hospital, 13530 Hämeenlinna, Finland.
Tel.: +358 3 629 3562; fax: +358 3 629 2900; e-mail: heikki.teppo@khshp.fi
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
Results: Delays were significantly longer in laryngeal can-
cer. Moreover, longer diagnostic delays worsened survival
markedly only in laryngeal cancer. Cut-off points at
which the delays showed significant adverse impact in
prognosis of laryngeal cancer were ‡3 months in patient
delay and ‡6 months in professional delay. If mirror
laryngoscopy was not performed at the initial visit,
professional delay in laryngeal cancer turned out to be
significantly longer.
Conclusions: Diagnostic delays are longer and have a
more significant impact on survival in laryngeal cancer
than in lingual or pharyngeal cancer. Thus, patients with
symptoms suggestive of laryngeal process should always be
examined properly at the initial visit in order to shorten
the professional diagnostic delay in laryngeal cancer.
indirectly: patients who were neither referred nor con-
trolled after the initial visit had both lengthened profes-
sional diagnostic delays and worse survival.8 However, we
are not aware of the relative importance of the diagnostic
delays in prognosis in different head and neck cancers, as
the above-mentioned studies have been made on only
one tumour site at a time. Results contradicting the link
between diagnostic delays and survival have also been
reported: Pera et al.9 and Allison et al.10 found no corre-
lation between diagnostic delays and prognosis in laryn-
geal cancer. Likewise, Vernham and Crowther11 and
Guggenheimer et al.12 found no correlation between
delays and clinical stage in head and neck cancer. In a
British study by McGurk et al.,13 delay had no impact on
stage or survival in oral and laryngeal cancer.
In this study, we evaluated the relative importance of
different patient and professional diagnostic delays in the
2- and 5-year survival rates of those head and neck can-
cers, in which longer delays have previously shown inde-
pendent prognostic significance (tongue, pharynx and
larynx) using a single population-based cohort of patients
325
326 H. Teppo & O.-P. Alho
diagnosed in 1986–1996. Our aim was to find out the
length of different types of diagnostic delays that signifi-
cantly increases the risk of death in head and neck can-
cers. Also, we wanted to assess the relative impact of
these delays between different tumour sites.
Methods
A population-based retrospective cohort design was used.
Population at risk
The area in which the cancer patients were identified com-
prises 87 municipalities (total population about 700 000),
which maintain one primary health centre each, and four
central hospitals and one university hospital (Oulu Uni-
versity Hospital) collectively. All head and neck cancer
patients of the area are supposedly treated in the univer-
sity hospital, as in this district secondary care units are
very small and unsuitable for treatment of head and neck
cancer. Within the whole University Hospital district,
40% of the population lives in the secondary health care
areas of the four small central hospitals and 60% in the
health care area of the University Hospital itself. However,
there is no additional diagnostic delay involved among the
former, as ENT-clinics of the secondary centres perform
the diagnostic studies (imaging, diagnostic endoscopies
and biopsies) and then refer the patients to University
Hospital for planning and execution of treatment. The
health care system in Finland is based on a general health
insurance scheme and provides equal access to medical
services for all citizens. Municipalities are responsible for
health care, which is covered by tax revenues. All patients
must first present in the primary care (municipality’s gen-
eral practitioner), and no one can be admitted to hospital
without a referral letter from a physician working in pri-
mary care except in a case of emergency. Occasionally
patients first contact a specialist working in private sector
from whom they get the referral to hospital, although this
is rare. If the referral arouses suspicion of head and neck
malignancy, the patient is admitted to the hospital within
a few weeks. The diagnostics, staging and planning and
execution of treatment are all performed in the tertiary
care referral ENT-centre. Finnish law obliges all licensed
physicians to keep medical records of each medical visit.
Cases of head and neck carcinoma
All patients with cancer of the anterior mobile tongue
(International Classification of Diseases, 9th and 10th
revisions, codes 141, C02) and pharyngeal (146–149,
C01.9, C09-13) and laryngeal cancer (161, C32) between
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
1 January 1986 and 31 December 1996, were identified
from the registers of Oulu University Hospital. Cancer of
tongue base (posterior one-third) was considered to be
oropharyngeal cancer, as it is defined in TNM classifica-
tion.14 In the following, cancer of the tongue refers to
anterior two-thirds, i.e. mobile oral tongue. Cancer of the
tongue, pharynx and larynx were chosen as examples of
different head and neck cancers with different presenting
symptoms, decisive diagnostic procedures, diagnostic
delays and prognosis. Only cases of histologically verified
squamous cell carcinoma were included (n = 318). We
have shown earlier15 that our sample was population-
based by cross-checking our cases with the entries in the
national Finnish Cancer Registry, whose files are practi-
cally complete.16 Among the 267 patients who were diag-
nosed after a referral from primary care, sufficient
primary care data were available in 221 patients (83%),
who formed our cohort.
From the university hospital charts, we collected infor-
mation on the primary site and subsite of the tumour
and clinical stage (I–IV).14 Each patient’s medical and
dental charts in the primary health centre and at the
local private practices were searched for data on the med-
ical visits made during the 12 months preceding the diag-
nosis of head and neck carcinoma. The data on the onset
of symptoms and, thus, the length of the patient delay,
were derived from these primary health care charts com-
pleted real-time at the time of the initial visit instead of
questionnaires or interviews after the true nature of the
symptoms was revealed. Patient delay was the time per-
iod between the onset of the presenting symptom and
the initial medical visit, and professional (physician)
delay was the time period between this first medical visit
and histological diagnosis of cancer. Also, data concern-
ing the examinations performed by the physician at the
initial visit in primary care were collected (i.e. whether
mirror laryngocopy, oral inspection, or posterior mirror
rhinoscopy were recorded to have been performed). Even
though mirror examination is considered difficult among
primary care physicians,17,18 in Finland it is still a part of
the curriculum in medical schools and considered a basic
skill among general practitioners. On the other hand,
fiberoptic nasoendoscopy is performed only by an ENT
resident or specialist. The dates and causes of death were
obtained from Statistics Finland. The Finnish Ministry of
Social Affairs and Health granted permission to collect
these data.
The treatment of all head and neck cancer cases at the
university hospital is planned in a joint meeting with
radiation oncologists and head and neck and plastic sur-
geons. Treatment is based on the clinical stage and loca-
tion of the tumour and follows suggested guidelines.19
 2008 The Authors
 Diagnostic delay in head and neck cancer 327

Statistical methods Table 1. Patient and tumour characteristics in a cohort of 221

head and neck cancer patients diagnosed in 1986–1996
Median delays were compared with Mann–Whitney
U-test (two groups) and Kruskal–Wallis test (three Tongue Pharynx Larynx
(n = 62) (n = 66) (n = 93)
groups). Overall (all cause mortality) survival curves for 2
and 5 years were constructed according to the Kaplan– Mean age 61 (26–86) 64 (34–89) 64 (37–86)
Meier method, starting from the date of the histologic (range), years
diagnosis,20 and the differences were tested with the log Males (%) 29 (47%) 47 (71%) 80 (86%)
rank test. The results of the Kaplan–Meier analyses are Tumour Tongue Nasopharynx Supraglottis
presented as death rate curves in figures and as mean sur- subsite 8 (12%) 33 (35%)
vival times in text. We calculated death rates for 2 and Margin Oropharynx Glottis
45 (73%) 29 (44%) 60 (65%)
5 years (starting form the date of histologic diagnosis) for
Body Hypopharynx
patient groups dichotomised according to patient and
17 (27%) 29 (44%)
professional delay cut-off points of 2 weeks, 1, 2, 3, 6 and Clinical stage*
12 months with the Kaplan–Meier method. The indepen- I 8 (13%) 1 (2%) 37 (40%)
dent effect of those diagnostic delays found to be signifi- II 22 (35%) 10 (15%) 15 (16%)
cant in univariate analyses were further evaluated with III 25 (40%) 12 (18%) 29 (31%)
Cox multivariate model,21 in which the cumulative IV 7 (11%) 41 (62%) 12 (13%)
risk of dying was calculated as a hazard ratio (HR)
*According to TNM classification.
with 95% confidence intervals (CI) adjusted for age (<
or ‡ 65 years), sex, tumour subsite and clinical stage (I– Table 2. Diagnostic delays in a cohort of head and neck cancer
III versus IV, I–II versus III–IV and all stages separately). patients diagnosed in 1986–1996
Tongue Pharynx Larynx
Results (n = 62) (n = 66) (n = 93) P*

The patient and tumour characteristics are presented in Patient delay 

Table 1 and the diagnostic delays in Table 2. Diagnostic Median 1.4 (0–46) 1.0 (0–12) 2.0 (0–61) 0.08
delays – especially professional delay – were significantly (range), months
longer in laryngeal cancer compared with those of pha- <1 months 19 (31%) 22 (33%) 22 (24%)
‡1, <3 months 23 (37%) 26 (39%) 32 (35%)
ryngeal or tongue cancer.
‡3 months 20 (32%) 18 (27%) 39 (42%)
In pharyngeal cancer, patient delay longer than
‡6 months 11 (18%) 8 (12%) 18 (19%)
2 months showed a trend towards worsened survival Professional delayà
(mean survival time 28 versus 20 months in patient delay Median 0.7 (0.1–18) 1.1 (0.1–9) 2.1 (0.2–75) <0.001
of less and over 2 months, respectively, P = 0.073). In (range), months
laryngeal cancer, the impact of delay was significant: the <1 months 35 (56%) 29 (44%) 17 (18%)
cut-off point which made the difference in survival was ‡1, <3 months 17 (27%) 25 (38%) 40 (43%)
3 months in patient delay (mean survival time 48 months ‡3 months 10 (16%) 12 (18%) 36 (39%)
versus 43 months, P = 0.039) and 6 months in profes- ‡6 months 6 (10%) 4 (6%) 22 (24%)
sional delay (22 versus 17 months, P < 0.001) (Fig. 1). In *Kruskal–Wallis test.
tongue cancer the impact of delays on survival were insig-  
From the onset of symptoms to the first medical visit.
nificant and often paradoxical: shorter delays showed a à
From the first medical visit to the diagnosis.
trend towards impaired survival. The 2- and 5-year death
rates according to different cut-off points of both patient HR of 3.5 (95% CI 1.8–6.9) in laryngeal cancer as com-
and professional delays in different tumour sites are pre- pared to the delay of <6 months. Other independently
sented in Table 3 (patient delay) and Table 4 (profes- significant determinants of survival in this multivariate
sional delay). In laryngeal cancer, patient delay of less model were older age (HR 2.3, 95% CI 1.1–4.5) and
than 3 months resulted in an absolute reduction of 24% advanced stage (stages III–IV) (HR 2.6, 95% CI 1.2–5.3).
in 5-year death rate (Table 3). The absolute difference in In laryngeal cancer, mirror laryngoscopy was per-
death rate was as high as 37% between laryngeal cancer formed in 61 (66%) of patients at the initial visit to pri-
patients with professional delay of less than or more than mary health care. Performing mirror examination of
6 months (Table 4). After adjustment for other prognos- larynx significantly shortened the following professional
tic factors, professional delay of ‡6 months resulted in an diagnostic delay (median professional delay 1.7 months

 2008 The Authors

Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
328 H. Teppo & O.-P. Alho

(a) 75 P = 0.039 lengthened professional diagnostic delay (median profes-

sional delay 1.8 months versus 1.0 month among those
with laryngoscopy performed), but the association was
Patient delay ≥ 3 months
% of patients

50 not statistically significant (P = 0.23). In nasopharyngeal

cancer (n = 8), only two patients had a posterior rhi-
noscopy executed. Oral inspection was performed practi-
25 cally to all patients with oropharyngeal (90%) or tongue
Patient delay < 3 months
(100%) cancer.

0
0 12 24 36 48 60 Discussion
Months after diagnosis
Key findings
(b) 75 P < 0.001 Physician delay ≥ 6 months
Patient and professional diagnostic delays were signifi-
cantly longer in laryngeal cancer than in cancer of the
pharynx or tongue. The length of patient and professional
% of patients

50
delays showed prognostic significance only in laryngeal
cancer. Professional delay appeared to have an indepen-
25 dent and dose-dependent effect on survival in laryngeal
cancer which supports causality. Performance of mirror
Physician delay < 6 months laryngoscopy at the initial primary care visit was an
0 important determinant of shorter professional diagnostic
0 12 24 36 48 60 delay in laryngeal cancer.
Months after diagnosis

Fig. 1. Kaplan–Meier death rate curves for overall survival in
laryngeal cancer patients (n = 93) divided by (a) patient delay
(cut-off point 3 months) (P = 0.039, log rank test) and (b) pro-
fessional (physician) delay (cut-off point 6 months) (P < 0.001,
log rank test).
versus 3.6 months among those who were not checked
with mirror laryngoscopy, P = 0.003). In hypopharyngeal
cancer (n = 29), 41% of patients were left without
mirror laryngoscopy at the initial visit. This led to
Strengths of the study
The study was population-based, and to our knowledge
the first in which the relative importance of diagnostic
delays is evaluated among different head and neck cancers
in a single cohort. Data concerning the onset of symp-
toms and the first visit to a physician and hence, the
length of the patient delay, were gathered from primary
health care charts completed real-time before the malig-
nant diagnosis was established. This decreases the risk of

Table 3. Patient delay in head and neck cancer: 2- and 5-year death rates dichotomised by different delay cut-off points

Patient delay
<3 months ‡3 months <2 months ‡2 months <2 weeks ‡2 weeks
* *
Death %* Death %* P Death %* Death %* P Death %* Death %* P*

Tongue (n = 62)
At 2 years 29 25 0.72 31 23 0.50 50 23 0.041
At 5 years 36 30 0.62 41 27 0.28 50 31 0.12
Pharynx (n = 66)
At 2 years 63 61 0.80 54 72 0.31 50 65 0.36
At 5 years 75 94 0.35 70 93 0.073 71 83 0.38
Larynx (n = 93)
At 2 years 22 31 0.44 25 27 0.96 31 25 0.67
At 5 years 30 54 0.039 32 47 0.23 39 40 0.99

*Percentages calculated with the Kaplan–Meier method, the statistical significance of the difference estimated with the log rank test.
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
 Diagnostic delay in head and neck cancer 329

Table 4. Professional delay in head and neck cancer: 2- and 5-year death rates dichotomised by different delay cut-off points

Professional delay
<6 months ‡6 months <3 months ‡3 months <2 weeks ‡2 weeks
* *
Death %* Death %* P Death %* Death %* P Death %* Death %* P*

Tongue (n = 62)
At 2 years 25 50 0.21 27 30 0.89 37 23 0.25
At 5 years 42 50 0.35 33 40 0.71 42 30 0.32
Pharynx (n = 66)
At 2 years 65 25 0.16 67 36 0.089 75 59 0.51
At 5 years 82 50 0.18 82 73 0.21 100 76 0.14
Larynx (n = 93)
At 2 years 17 55 <0.001 18 39 0.014 17 26 0.58
At 5 years 31 68 <0.001 32 53 0.023 33 40 0.73

*Percentages calculated with the Kaplan–Meier method, the statistical significance of the difference estimated with the log rank test.

bias in terms of accurate timing, when compared with
data gathered retrospectively after the patient is aware of
the true nature of his ⁄ her illness. The inclusion of Cox
multivariate analysis enhances the significance of profes-
sional delay as an independent prognostic factor in laryn-
geal cancer.
Limitations of the study
The study design is retrospective. Therefore, we were not
able to control the quality of data recorded in the pri-
mary care charts at the first visit. However, prospective
design is virtually impossible to implement when studying
diagnostic delays in cancer. Even though we present a
population-based sample, sufficient primary care data
were not available on all cases which makes selection bias
possible.
Clinical applicability of the study
In pharyngeal cancer, the most common presenting
symptom is pharyngalgia without infection4 and in ton-
gue cancer, pain and ⁄ or visible tumour or ulceration in
the tongue.8 These symptoms are threatening and proba-
bly cause the patient to seek medical advice and the pri-
mary care physician to refer the patient to a specialist
more promptly. On the other hand, hoarseness, clearly
the most common presenting symptom in laryngeal can-
cer,3 is commonplace among smokers and non-smokers
alike, especially when transient. This explains the differ-
ences in lengths of diagnostic delays and also their impact
on survival.
The public health system in Finland requires patients
to contact primary care physician – GP or private
 2008 The Authors
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
practice specialist – to get a referral to specialist care in
hospital. This can be considered a rigid, bureaucratic
arrangement but on the other hand, it helps secondary
and tertiary care to prioritise patients and thus, it ensures
that patients with need of undelayed medical attention
can be treated in due time. Accordingly, this study repre-
sents a truly population-based, unselected sample in
which patients’ economical or insurance status has no
impact on access to specialist health care. The health care
system in Finland is very similar to that in all Scandina-
vian countries and thus, the conclusions of this study are
generalisable at least in Nordic countries. However,
regardless of the health care referral pattern, the finding
of independent, significant impact of professional delay
on survival in laryngeal cancer can be generalised to all
Western countries.
According to this study, length of diagnostic delay
has significant consequences only in laryngeal cancer.
Patient delay is difficult to shorten, but there are means
to improve professional diagnostic delay. Physicians
should be trained to consider hoarseness in an elderly
smoker to be caused by malignancy until proved other-
wise, even though the symptom is common and usually
not dangerous. Thus, hoarse patients should always
have mirror laryngoscopy or some other visualisation of
larynx performed and if it cannot be reliably done, they
should be referred to specialist care without further
delay. This study suggests that mirror laryngoscopy per-
formed by a GP is beneficial in shortening diagnostic
delay in laryngeal cancer and it should be encouraged.
In pharyngeal and lingual cancer, presenting symptoms
are often dire and, therefore, diagnostic delays are not
long and they have less impact on the overall survival
of the patient.
330 H. Teppo & O.-P. Alho
Acknowledgements
We are indebted to Saara Kantola, DDS and Niko Rant-
ala, MD for collecting part of the data.
Conflict of interest
None to declare.
References
1 Lefebvre J., Lartigau E., Kara A. & Sarini J. (2001) Oral cavity,
pharynx and larynx cancer. In UICC (International Union
Against Cancer): Prognostic Factors in Cancer, Gospodarowicz
M., Henson D., Hutter R., O¢Sullivan B., Sobin L. & Wittekind
C. (eds), pp. 151–165. Wiley-Liss ⁄ John Wiley & Sons Inc., New
York
2 Franceschi S., Barzan L. & Talamini R. (1997) Screening for can-
cer of the head and neck: if not now, when? Oral Oncol. 33,
313–316
3 Teppo H., Koivunen P., Hyrynkangas K. & Alho O.P. (2003)
Diagnostic delays in laryngeal carcinoma: professional delay is a
strong independent predictor of survival. Head Neck 25,
389–394
4 Koivunen P., Rantala N., Hyrynkangas K., Jokinen K. & Alho
O.P. (2001) The impact of patient and professional diagnostic
delays on survival in pharyngeal cancer. Cancer 92, 2885–2891
5 Brouha X., Tromp D., Hordijk G., Winnubst J. & de Leeuw J.
(2005) Oral and pharyngeal cancer: analysis of patient delay at
different tumor stages. Head Neck 27, 939–945
6 Tromp D., Brouha X., Hordijk G., Winnubst J. & de Leeuw R.
(2005) Patient and tumour factors associated with advanced
carcinomas of the head and neck. Oral Oncol. 41, 313–319
7 Pitchers M. & Martin C. (2006) Delay in referral of oropharyn-
geal squamous cell carcinoma to secondary care correlates with
a more advanced stage at presentation, and is associated with
poorer survival. Br. J. Cancer 94, 955–958
Journal compilation  2008 Blackwell Publishing Ltd • Clinical Otolaryngology 33, 325–330
8 Kantola S., Jokinen K., Hyrynkangas K., Mäntyselkä P. & Alho
O.P. (2001) Detection of tongue cancer in primary care. Br. J.
Gen. Pract. 51, 106–111
9 Pera E., Moreno A. & Galindo L. (1986) Progpostic factors in
laryngeal carcinoma – a multifactorial study of 416 cases. Cancer
58, 928–934
10 Allison P., Franco E., Black M. & Feine J. (1998) The role of
professional diagnostic delays in the prognosis of upper
aerodigestive tract carcinoma. Oral Oncol. 34, 147–153
11 Vernham G. & Crowther J. (1994) Head and neck carcinoma:
stage at presentation. Clin. Otolaryngol. 19, 120–124
12 Guggenheimer J., Verbin R., Johnson J., Horkowitz C. &
Myers E. (1989) Factors delaying the diagnosis of oral and
oropharyngeal carcinomas. Cancer 64, 932–935
13 McGurk M., Chan C., Jones J., O¢Regan E. & Sherriff M.
(2005) Delay in diagnosis and its effect on outcome in head
and neck cancer. Br. J. Oral. Maxillofacial Surg. 43, 281–284
14 American Joint Committee on Cancer (2002) AJCC Cancer Stag-
ing Handbook, 6th edn. Springer, New York
15 Alho O.P., Teppo H., Mäntyselkä P. & Kantola S. (2006) Head
and neck cancer in primary care: presenting symptoms and the
effect of delayed diagnosis of cancer cases. CMAJ 174, 779–784
16 Teppo L., Pukkala E. & Lehtonen M. (1994) Data quality and
quality control of a population-based cancer registry. Experience
in Finland. Acta Oncol. 33, 365–369
17 Shaw H. (1976) Early diagnosis of cancer in the head and neck.
BMJ 1, 379–383
18 Hoare T., Thomson H. & Proops D. (1993) Detection of laryn-
geal cancer – the case for early specialist assessment. J. R. Soc.
Med. 86, 390–392
19 Vokes E., Weichselbaum R., Lippman S. et al. (1993) Head and
neck cancer. N. Engl. J. Med. 328, 184–194
20 Kaplan E. & Meier P. (1958) Nonparametric estimation from
incomplete observations. J. Am. Stat. Assoc. 53, 457–481
21 Cox D. (1972) Regression models and life-tables (with discus-
sion). J. R. Stat. Soc. 34, 187–220
 2008 The Authors