Psychological Medicine, 1980,10, 231-241

Printed in Great Britain

Mental disorders in primary health care:

a study of their frequency and diagnosis in four
developing countries
T.W.HARDING, 1 M. V. DE ARANGO, 2 J. BALTAZAR, 3
C. E. CLIMENT, 2 H. H. A. IBRAHIM, 4 L. LADR1DO-IGNACIO, 5
R. SRINIVASA MURTHY 6 AND N. N. WIG 6
Investigators in a WHO Collaborative Study on Strategies for
Extending Mental Health Care

SYNOPSIS 1624 patients who were attending primary health facilities in 4 developing countries
were examined to determine how many were suffering from mental disorder. Using stringent
criteria to establish the presence of psychiatric morbidity, 225 cases were found, indicating an
overall frequency of 13-9 %. The great majority of cases were suffering from neurotic illnesses and
for most the presenting complaint was of a physical symptom, such as headache, abdominal
pain, cough or weakness. The health workers following their normal procedure correctly detected
one third of the psychiatric cases.

to be 20 %. Similar findings have been reported

INTRODUCTION
from other industrialized countries (Giel & Le
How frequent are mental disorders among Nobel, 1971).
patients presenting at the primary level of It would obviously be a mistake to assume
health care? Advances in techniques of reliable that the same result would be found in developing
case detection (Goldberg, 1974) have led to a countries. There, primary health care is very
number of studies which have shown that non- different, relying much more on simply trained
psychotic psychiatric illness is one of the most community health workers, and the incidence
common reasons for which doctors are con- and prevalence of conditions such as malnutrition
sulted. Goldberg & Blackwell (1970), in one of and infections are much higher. It would therefore
the first studies to be based on more efficient be reasonable to assume that, under such con-
research instruments than had been previously ditions, fewer patients with mental disorder
available, found the 'conspicuous psychiatric might come to primary health care facilities and
morbidity' of a London, UK, general practice that those who do would be numerically over-
shadowed by common, physical disorders such
1
Address for correspondence: DrT. W. Harding, Medical as malaria, tuberculosis, gastro-enteritis and
Officer, World Health Organization, 1211 Geneva 27,
Switzerland. Principal investigator for the study. pneumonia. Nevertheless, a few studies have
• Department of Psychiatry, Universidad del Valle, Cali, suggested that the rate of mental disorders seen
Colombia. Co-investigator and chief investigator respectively in primary health care may in fact be of the
of study area team in Colombia.
5
Institute of Public Health, Manila, Philippines. Co- same order in the developing countries as else-
investigator of study area team in Philippines. where (Giel & Van Luijk, 1969; Mbanefo, 1971;
4
Department of Psychiatry, Faculty of Medicine, Uni- Harding, 1973; Ndetei & Muhangi, 1979). Some
versity of Khartoum, Khartoum, Sudan. Chief investigator
of study area team in Sudan. further support is given to these observations by
6
Department of Psychiatry, University of the Philippines studies of mental disorders at the second level
System, Manila, Philippines. Chief investigator of study of health care (Holmes & Speight, 1975; McEvoy
area team in Philippines.
' Department of Psychiatry, Postgraduate Institute of & McEvoy, 1976). However, these are all rather
Medical Education and Research, Chandigarh, India. Co- scattered, small-scale studies which, for a number
investigator and chief investigator respectively of study of reasons, may be unrepresentative.
area team in India.
231
0033-2917/80/2828-5060 $01.00 © 1980 Cambridge University Press
Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 232 T. W. Harding and others

This paper describes a study in which con- defined study areas: (i) Union de Vivienda
secutive, adult patients1 attending primary Popular (UVP), a crowded, poor barrio on the
health facilities in defined areas of 4 developing outskirts of the city of Cali, Colombia, with
countries were screened for the presence of a population of 63757 (1974); (ii) Raipur Rani
mental disorder. The diagnostic behaviour of (RR), a rural area in the state of Haryana, India,
the existing health workers is also described. with a population of 64642 (1976); (iii) Shagara
The study was undertaken as part of the Jebel Awlia (SJA), an agricultural zone in
baseline observations in the WHO Collaborative Khartoum Province, Sudan, with a population
Study on Strategies for Extending Mental Health of 58655 (1975); (iv) Sampaloc (SAM), a densely
Care (Harding, 1978; WHO, 1979). The study's populated area in Manila, Philippines, with a
aim is to develop and evaluate low cost methods population of 75388 (1975).
of mental health care as an integral part of
general health services. After the baseline 2. Sampling
observations were completed, the health workers In each area a complete list of all health facilities
in each of the study areas participating in the was drawn up, together with information on the
study2 have received brief, task-oriented average number of consultations by adults each
training on the detection and management of week in each facility. The minimum number of
a limited number of priority mental disorders. adults to be screened was fixed as 300 per
A number of the planned interventions will 50000 inhabitants. A screening quota was then
involve community agents such as teachers, established for each health facility in proportion
policemen, communal leaders or village chiefs to the attendance rates. (Thus, in an area with
and, in some cases, traditional healers. Never- a population of 60000 with 4 health facilities
theless, the primary health care facilities will be with average daily attendance rates of 95, 60, 15
an important focus of action and it was therefore and 10 respectively, a total screening target of
essential to know, prior to the interventions, the 360 would be set with individual health facility
answers to the following. How many and what quotas of 190, 120, 30 and 20 respectively.)
type of mental disorders are already present Screening was then carried out at each facility
among patients coming for primary health care ? in turn. Consecutive attenders aged 16 or over
What proportion is correctly identified by the were screened on successive days until the quota
health workers? Which presenting complaints was reached. Screening always commenced with
are commonly associated with the presence of
the first patient attending in the morning.
a psychiatric disturbance? This paper gives the
Excluded from screening were: (i) patients who
answers to these questions from 4 developing
were so seriously ill (e.g. in coma) or required
countries.
such urgent medical care that it would be un-
reasonable to administer the research question-
METHOD naires; (ii) patients who refused to take part;
(iii) patients who had already attended once
1. The study areas
during the exercise and had therefore been
The screening exercise (for which results are screened.
reported in this paper) was carried out in 4
3. Reason for attendance and health staff rating
1
A similar study has been carried out in the same areas For each patient screened, basic identifying data
with children. The results are being published separately
(Giel et al. in preparation). A survey of community attitudes
were obtained (age, sex, marital status and
and perceptions has also been carried out (Wig et al. 1980). occupation) and the patient was asked why he
* 'First-phase' study areas are: Union de Vivienda or she had come to the health facility, the
Popular in Cali, Colombia; Raipur Rani, Haryana State,
India; Niakhar, Region of Sine Saloum, Senegal; Shagara response being recorded verbatim. The screening
Jebel Awlia, Khartoum Province, Sudan. In 1977/8 further process (see section 4) was carried out com-
teams have started work in Brazil, Egypt and the Philippines. pletely independently of the usual consultation/
Results from the team working in Sampaloc, Manila, in the
Philippines, are already available and are reported here. In treatment process, the results not being available
the Niakhar study area a modified method for case identi- to the health worker who was seeing the patient.
fication has been used with a one-stage screening of a much This health worker indicated at the end of the
larger number of individuals. The Niakhar results will there-
fore be reported separately. normal consultation process whether or not he/

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 Mental disorders in primary health care 233

she thought mental disorder was present, by to whom the questions were read by a trained
marking a cross against 1 of 5 items on a investigator until such responses were reliably
Health Staff Rating schedule (HSR). The items recorded.
were: The English version of the 20 'non-psychotic'
(1) a physical health problem only, items and the 4 'psychotic' items of the SRQ is
(2) a mental health problem only, shown in the Appendix.
(3) a physical and mental health problem, 5. Follow-up
(4) no health problem of any kind, On the basis of pilot testing, a cut-off point for
(5) no rating possible. the total score on the 20 'non-psychotic' items
of the SRQ was selected for each study area.
4. First-stage screening This selection was based on the score which
A Self Reporting Questionnaire (SRQ) with 24 was likely to yield optimal sensitivity and
items was used. The first 20 items were designed specificity, i.e. to yield as few false positives and
to detect non-psychotic disorders. They were false negatives as possible.1 All patients scoring
selected by a consensual process, comparing above this cut-off point or scoring at least one
items in 4 instruments used in a variety of positive item on the 4 'psychotic' items were
cultural settings: the Patient Self-report regarded as 'potential cases' and followed up.
Symptom Form (PASSR), an instrument de- In addition, a sample of those cases scoring less
veloped and tested in Cali, Colombia (Climent than the 'non-psychotic' cut-off point and with
& Plutchick, 1980); the PGI Health Question- no positive 'psychotic' items were also followed
naire N2 developed by Wig and his colleagues up.
in Chandigarh after they found the Cornell The follow-up procedure included:
Medical Index to be inappropriate to the Indian (i) A structured psychiatric interview: the
setting (Verma & Wig, 1977); the General shortened version of the Present State Exami-
Health Questionnaire (GHQ) used originally by nation (PSE). The PSE was administered by
Goldberg in England but subsequently validated research psychiatrists who had undergone a
in the United States (Goldberg, 1972), Jamaica recognized training in the use of the PSE in
(Harding, 1976), and many other settings; the English and had been given further experience
'symptom' items on the shortened version of in its use in local languages.
the Present State Examination (PSE) (Wing et (ii) A diagnostic assessment and formulation
al. 1974). The full version of the PSE has been (DAF) completed on the basis of the PSE
adapted and tested in a wide range of cultural ratings. Once again, research psychiatrists made
settings. these diagnoses and inter-centre reliability has
This comparison produced a list of 32 items been checked (WHO, in preparation).
which were either identical or very similar in (iii) A confidential case register established
meaning. From these, 20 items were selected by for each study area in which details of those
agreement between the chief investigators in the patients identified as psychiatric cases were
first-phase study area teams on the basis of ease entered.
of translation and cultural relevance. The 4 ad- The screening procedure is summarized in
ditional items, designed to detect psychotic Fig. 1.
conditions, were based on the items in Foulds' 6. Translation and training
Symptom Sign Inventory (Foulds & Hope, 1968)
which have been shown to be most effective in All instruments were translated by study teams
detecting psychotic illness. into local languages as necessary and appropriate,
e.g. in RR, SJA and SAM research workers
Since most attending patients were illiterate,
1
the SRQ items were read to the patient by a The details of the item selection, calibration and testing
research assistant. In most instances the research of validity are not given in this paper as they are being
published separately in a monograph describing all methods
assistants were selected from among local people. developed in the course of the collaborative study (WHO,
They underwent a 5-hour training in the ad- in preparation). Nevertheless, it is of interest to note that
ministration of the questionnaire. This included the cut-off points selected after calibration of the instrument
in the different study areas varied as follows: UVP, 10/11;
the recording of responses of several patients RR, 5/6; SJA, 3/4; SAM, 6/7.

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 234 T. W. Harding and others

All adults attending clinic

SRQ completed

|
I
Normal consultation: HSR completed independently

Potential cases according to SRQ Mental disorder unlikely

(meets none of criteria

on SRQ)

All patients 10% patients

90% patients
PSE: screening version
DAF
No further action

'Non-cases'

No further action

'Psychiatric
cases'

Entry into
register if
cases
FIG. 1. Case detection procedure: diagrammatic representation. SRQ, Self Reporting Questionnaire; HSR, Health
Staff Rating; PSE, Present State Examination (screening version); DAF, Diagnostic Assessment Form.

could use English but all questions put to detect psychiatric morbidity. The performance of
patients had to be translated into Hindi, Arabic the 24-item screening instrument has been
and Filipino. Back translations were made assessed by administering the questionnaire to
independently by workers who did not know a group of normal subjects from each of the
the original version. Where necessary, adjust- study areas and calculating the sensitivity and
ments were then made to the original translation. specificity of the instrument at the cut-off point
Research workers were trained in groups, starting selected for each study area. The sensitivity
by administering the screening schedules to each varied between 73 % and 83 % in the 4 study
other and proceeding to practising with 20 areas, while the specificity varied between 72 %
patients before the screening proper was started. and 85 %. These figures mean that the overall
misclassification rate in a population with a
RESULTS psychiatric morbidity rate of 15 % would vary
between 18 % and 24 %. The second stage of
1. The validity of the screening process the detection process was based on a Present
The results of this study depend on the reliability State Examination interview, a method which
and validity of the two-stage process used to has been widely used and reviewed in many

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 Mental disorders in primary health care 235

countries. A reliability exercise was carried out Table 1. Frequency of mental disorder among
between the psychiatrists making diagnostic adults attending primary health care facilities as
ratings. This showed a high level of agreement estimated by a two-stage screening process
in discriminating between 'cases' and 'non-
cases', but lower levels of agreement on specific Study areas
diagnostic entities at the fourth digit level of Com-
ICD-8 (e.g. between 'depressive' and 'anxiety' bined
neuroses). Full details concerning the validity UVP RR SJA SAM results
and reliability of the methods used are given in Patients screened (cases) 444 361 360 459 1624
a section of a monograph describing all the new 'Potential' cases identi- 46 61 37 66 210
methods developed and tested in the course of fied as 'definite cases'
the collaborative study (WHO, in preparation). on follow-up*
Estimate of additional 2 3 1 9 15
2. The populations screened cases missed by first-
stage screenf
The numbers screened in each area were 444 in Minimum estimate of 48 64 38 75 225
UVP, 36! in RR, 360 in SJA and 459 in SAM, cases (%) (10-8) (17-7) (10-6) (16-3) (13-9)
making a total of 1624. No patients refused to
* 'Potential cases' identified as 'possible cases' on follow-
be screened; 5 patients in UVP, 6 in RR and up are excluded from consideration.
4 in SJA could not be screened because their t Based on sample of patients who passed first-stage
conditions needed urgent attention. The male: screen not identified as possible cases, calculated as minimum
number at 95 % level of probability.
female ratios were 1:6-9 in UVP, 1:2-0 in RR,
1:1-8 in SJA and 1:3-5 in SAM. The preponder-
Table 2. Diagnoses of mental disorders (%)
ance of women in UVP and SAM reflects the
seen in primary health care
emphasis given to maternal care (antenatal, natal
and postnatal) in the health services of these All mental
areas. disorders
Diagnosis* (N = 288)
3. The frequency of mental disorders
Depressive neurosis (300.4) 48-6
The diagnostic assessment and formulation al- Anxiety neurosis (300.0) 30-9
lowed for 3 possible categories: 'non-cases', Other neurosis (other 300) 8-3
Schizophrenia (295) 31
'possible cases' and 'definite cases'. In establish- Mental retardation (310/311/312) 21
ing frequency of mental disorders, we have con- Affective psychosis (296) 1-7
Others (—) 5-2
sidered only 'definite cases'. The frequency has
been established in the following way. The * The corresponding codes from ICD-8 are given in
denominator was simply the total number of parentheses.
individuals screened. The ordinator was the sum
of (i) all 'potential cases' (by SRQ) identified as mental disorders in the 4 areas is 10-8 % in
'definite cases' in the follow-up; and (ii) an UVP, 17-7 % in RR, 10-6 % in SJA and 16-3 %
estimate of the number of cases missed by the in SAM. The combined 'minimum' frequency
first-stage screening, based on the frequency of based on 1624 adult attenders in primary health
'definite cases' among the sample of patients care facilities in 4 countries is 13-9 %.
who did not otherwise meet the criteria for The great majority of these cases were diag-
follow-up. This estimate was the minimum nosed as being neurotic disorders, as is shown
number of definite cases in the group from which in Table 2 (which is, however, based on both
this sample was drawn at the 95 % level of 'definite' and 'possible' cases). Few cases of
probability. Thus two decisions were made, repre- functional psychosis or mental retardation were
senting a conservative approach to the estimate diagnosed.
of frequency on the basis of our data. It is
4. Diagnostic ratings by health staff and their
therefore reasonable to regard the figures ob-
tained as indicating minimum estimates of the level accuracy
of mental morbidity. The detailed results are Most patients were diagnosed as having physical
shown in Table 1. The estimate of frequency of health problems, as shown in Table 3. In UVP

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 236 T. W. Harding and others

Table 3. Diagnostic ratings by health staff (%) Table 4. Comparison between health staff diag-
noses and case detection by screening procedure
UVP RR SJA SAM in 1624 consecutive adult attenders at primary
(JV = 444)(W = 36t)(W = 3 6 0 ) ^ = 459)
health care facilities in all 4 areas
Physical health 85-6 861 95-8 71-5
problem Diagnosis by health staff
Mental health 1- 44 2-5 G-9
problem Detection 'Mental ' Mental
Physical and 7-7 8-3 1-4 21-1 by screening disorder disorder not
mental health procedure present' present' Total
problem
No health problem 3-2 H — 6-5 Cases 106 182 288
Unable to rate 1-8 — 0-3 — Non-cases 97 1239 1336
Total 203 1421 1624

and RR the health workers were ready to make Frequency of mental disorder by two-stage screening
procedure = 17-7%. (This frequency is higherthan that quoted
a diagnosis of physical and mental health prob- earlier in section 3, since it includes 'possible cases' as well
lems in 7-8 % of patients, while in SJA such as 'definite cases'.)
combined diagnoses were much less frequent. Rate of diagnosis of mental disorder by existing health
staff = 6-5%, their diagnostic sensitivity = 36-8%, their
In SAM there was a strikingly higher rate of diagnostic specificity = 92-7%.
such diagnoses (21-1 %). The overall frequency
with which health staff diagnosed mental health culated. The diagnostic sensitivity is relatively
problems (either alone or in combination with low (36-8 %), while the diagnostic specificity is
physical disorders) was 9-5 % in UVP, 12-7 % high (92-7%); i.e. for every 3 cases of mental
in RR, 3-9 % in SJA and 22 % in SAM. These disorder coming to primary health care, 1 will
frequencies are lower than the frequency of be diagnosed as such by the health worker;
mental disorder found by the screening procedure while for every 100 non-cases, only between 7
in UVP, RR and SJA, but the difference is and 8 will be wrongly diagnosed as suffering
greater in SJA than the other areas. In SAM, on from mental disorder. The findings on so-called
the other hand, the health staff diagnosed more 'false positive' diagnoses need qualification,
cases of mental disorder than were detected by since it is entirely possible that the health workers
the two-stage screening procedure. were in certain cases more sensitive to mental
It is not, however, only the rate of diagnoses disorders than the research procedure (Giel &
which indicates the diagnostic skill of the health Le Nobel, 1971). The significant finding is that
worker, since both false positive and false this variety of discordance (psychiatric case
negative diagnoses may be made. Table 4 shows identified by the health worker but not by the
the relationship between diagnoses by health research procedure) occurred only half as
workers and the cases detected by the two-stage frequently as the other variety of discordance
screening procedure, with a decision by a psy- (psychiatric case identified by the research pro-
chiatrist as final criterion of 'caseness'. In this cedure but not by the health worker). The com-
and subsequent tables, both 'definite' and bined figures mask important differences between
'possible' cases are included, since we are no SAM and the 3 other study areas. The diagnostic
longer considering frequency as such, when it sensitivity in SAM was 46-9 % and the diagnostic
was reasonable to take a cautious approach. The specificity 83-3%. Thus, in SAM, nearly half
issue is now diagnostic skill, so that we would the cases were being correctly diagnosed, while
hope that health workers would be able to in the other 3 study areas less than one third
detect both 'definite' and 'possible' cases. From were diagnosed by the health workers.
Table 4 it can be seen that, of the 288 cases
detected by the screening procedure, roughly
one third were diagnosed by health workers and 5. Reasons for attendance
two thirds were missed. Based on this table, the The many reasons for attendance quoted by
diagnostic sensitivity (a measure of how few attenders could all be classified as 1 of 90 items
cases are missed) and specificity (a measure of listed under 9 headings: (a) generalized symptoms
how few non-cases are diagnosed) can be cal- (e.g. fever, weakness); (b) pains, discomforts and

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 Mental disorders in primary health care 237

Table 5. Number of reasons for attendance quoted by patients in 4 study areas

UVP RR SJA SAM All 4 areas
Number of
reasons No. 0/
No. /o No. 0/
No. 0/
No. 0/
/o
/o /o /o
1 379 85-4 267 740 206 57-2 370 80-6 1222 75-2
2 45 101 59 16-3 106 29-4 69 150 279 17-2
3 or more 20 4-5 35 9-7 48 13-3 20 4-4 123 7-6
Total 444 361 360 459 1624

Table 6. Relationship between the number of such as anxiety, depression and sleep disturbance,
reasons quoted, the diagnosis of mental disorder were quoted infrequently as reasons for attend-
by health staff, and the detection of mental ance. Table 7 shows that the proportion of
disorder by two-stage screening including a mental disorders among patients quoting 1 of
psychiatric interview the 12 most common reasons varies from 6-8 %
in the case of women attending for antenatal or
Cases (%) of mental disorder postnatal care to 35-5% and 28-9% among
diagnosed by
those quoting 'weakness' and 'dizziness' re-
(a) (b) Two-stage spectively. Health staff show, in fact, a low rate
Regular health screening of diagnosis of mental disorders among women
Number of reasons staff procedure
attending for antenatal or postnatal care, but
1 (N - 1222) 121 15-9 their rate is also low (and markedly discrepant
2 (N = 279) 12-9 201 with the rate found in the screening procedure)
3 or more (N = 123) 15-4 30-1
in the case of cough, fever and back pain. Most
patients presenting with mental disorders, there-
irritations; (c) respiratory symptoms; (d) gastro- fore, gave a physical symptom as reason for
intestinal symptoms; (e) genito-urinary symp- attendance. The majority of such cases missed
toms ; (/) psychological, neurological and sensory by the health workers were among patients
symptoms; (g) other symptoms; (h) 'diagnoses'; complaining of headache, abdominal pain, cough,
(0 other reasons. Most patients quoted one back pain and weakness.
symptom as reason for attendance, as shown in
Table 5, but a minority gave 2 or more reasons
CONCLUSIONS
(this minority being more substantial in SJA
than in the other areas). A few patients quoted Mental disorders do make up a significant pro-
not a symptom but a diagnosis as reason for portion of morbidity seen in primary health
attendance - e.g. hypertension, malaria, worms, care in the 4 communities studied. The frequency
asthma - while some patients attended for vari- recorded, using a carefully standardized two-
ous forms of routine care (e.g. family planning, stage screening procedure, was between 10-6 %
antenatal care). Table 6, which pools results and 17-7 %, a little lower than has been found
from the 4 areas, shows that those patients in industrialized countries (Kessel & Shepherd,
quoting 3 or more reasons for attendance have 1962; Shepherd et al. 1966; Goldberg & Black-
twice the chance of suffering from mental dis- well, 1970; Dilling et al. 1978). We have, how-
order (nearly 1 in 3 as opposed to 1 in 6). The ever, deliberately taken a conservative approach
pattern of health workers' diagnostic ratings in reaching these estimates, to pre-empt any
shows little variation with number of reasons criticism that the rates are inflated or do not
for attendance quoted by patients. represent 'true psychiatric morbidity'. Our
All reasons for attendance quoted by at least finding that there is indeed a significant rate of
1 % of patients are listed in Table 7. The five psychiatric morbidity in the primary health care
commonest reasons were headache, abdominal setting of developing countries is more important
pain, cough, genito-urinary symptoms and fever. than the finding that this rate is somewhat lower
At least one of these reasons was quoted by than has been reported from industrialized
over half the patients. Psychological symptoms, countries.

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 238 T. W. Harding and others

Table 7. Twelve commonest reasons for attendance ranked in order of frequency as related to diagnosis
of mental disorder by health staff and detection of mental disorder by screening procedure
Patients (%), quoting Patients (%), quoting
reason, diagnosed reason,
Number of patients by health staff as detected by screening
quoting reason suffering from procedure as having
Reason for attendance (N = 1624) mental disorder psychiatric disorder

1. Headache 208 14-4 230

2. Abdominal pain 170 12-4 21-8
3. Cough 167 13-2 180
4. Genito-urinary 129 15-5 16-3
5. Fever 128 8-6 13-3
6. Dizziness 90 28-9 28-9
7. Antenatal/postnatal care 88 3-4 6-8
8. Back pain 79 11-4 28-6
9. Chest pain 78 17-9 21-8
10. Weakness 76 22-4 35-5
11. Cold/influenza' 67 11-9 17-9
12. Family planning 54 11-1 241

The disorders detected were mainly neuroses, in the community. Meanwhile, such workers also
as has been found in earlier studies at primary need an effective means of coping with the
care level in industrialized countries (Pemberton, large numbers of patients with neurotic con-
1949; Kessel, 1960; Shepherd et al. 1966) and ditions who seek help of their own accord.
by Ndetei & Muhangi (1979) in Kenya. This In coping with patients with neurotic com-
finding raises some awkward questions. One plaints, one aim will be to avoid inappropriate
of the basic tenets in the provision of mental investigations, referral and treatment and to
health care in developing countries is the need prevent the build up of a cycle of frequent
to select priorities according to clearly established attendance and 'medicalization' of underlying
criteria (WHO, 1975; Giel & Harding, 1976). social problems. Although we have not presented
In following this principle, the teams partici- data in this paper on attendance rates, it is our
pating in this study have already established impression that patients with neurotic conditions
priorities for interventions (Harding, 1978; presenting with somatic symptoms tend to
Harding et al. 1979) in which psychotic con- consult primary care facilities on successive
ditions, presenting as acute emergencies or as occasions, to seek help elsewhere and to be
chronic disablement, figure prominently. The given inappropriate, symptomatic drug treat-
teams have also identified the primary health ment. We are collecting more systematic data
worker as the main agent for assuring simple to test this impression but the phenomenon has
mental health care (WHO, 1979). Yet our been well described elsewhere, for example by
finding is that neurotic conditions are commonly Giel & Workneh (1980) in Ethiopia and by
encountered by primary health workers in their Holmes & Speight (1975) in Tanzania. Many
everyday work, while psychotic disorders are neurotic conditions are likely to be self limiting,
rare. Indeed, so few psychotic patients were providing that maladaptive, coping mechanisms
detected in the screening exercise that it is not are not established during the early stages of
possible to draw conclusions about the health the illness. The health worker should be able
workers' ability to diagnose such patients. The to tell patients confidently when no disorder
disinclination of psychotic patients to seek help requiring medical intervention is found and to
at primary health facilities is clearly related to encourage patients to seek help within the com-
the attitudes and perceptions of mental illness munity for underlying social problems. This
which we have studied in a community survey, means building up a close relationship with
reported elsewhere (Wig et al. 1980). Reaching community leaders, traditional healers, religious
psychotic patients who do not present spon- leaders, teachers and other influential people
taneously at primary health care facilities calls who can help those with social problems. The
for active case finding by primary health workers same contacts will be useful for finding patients

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 Mental disorders in primary health care 239

with psychotic illnesses, many of whom are kept The tendency to miss psychiatric cases is
out of public view by their families. The health probably linked to another finding in the study:
worker needs to establish confidence in the the majority of patients with mental disorder
community that such patients can be helped by complain primarily of physical symptoms. Some
a combination of active medical and social of these symptoms are those that are usually
management. regarded as potentially 'psychosomatic': head-
Our results show that, with their present level ache, dizziness and weakness. There is also a
of training, health staff in the study areas diag- strong association between polysymptomatic
nose only one third of the cases of mental presentation and mental disorder, as would be
disorder among the patients they see. The expected. What is more surprising is the rate of
tendency of health workers to miss cases of mental disorder among patients complaining of
mental disorder has often been noted previously. cough and fever and among women coming
Kessel & Shepherd (1962) have pointed out the for family planning advice, all of whom have
complexity of the task of detecting psychological a rate in excess of 13%. It seems that many
disturbance in a setting of general morbidity. patients assume that a physical symptom is
Goldberg & Blackwell (1970) found that even almost a requirement in order to be seen at a
a general practitioner who was also a psychiatrist health facility. In British general practice, Kessel
'failed to detect one third of the disturbed (1962) has estimated that only 10 %of psychiatric
patients' detected by a research procedure. Giel patients present with psychological symptoms.
& Le Nobel (1971), in their studies of a Dutch Our own experience shows that the presenting
general practice, found that the practitioners complaint, once presented, often becomes in-
considered as 'likely psychiatric cases' only one significant. Many patients, for example, who
half of the 'certain or probable cases' identified complained of fever were found to have a normal
by two research psychiatrists. Studies of general body temperature. Few patients offered their
practice in Bavaria (Dilling et al. 1978) have 'psychiatric' symptoms as presentingcomplaints,
also shown that a considerable proportion of although these were often present on enquiry. In
psychiatric cases are missed. In a general hospital India, for example, people would rarely think of
study, physicians were found to miss about two a doctor as somebody to whom one would talk
thirds of all psychiatric cases among new out- about one's feelings. Only when a patient's mental
patients (Mezey & Syed, 1975). Missing psychi- status impinged on his physical sensations would
atric cases at primary care level is a world wide he seek medical advice.
phenomenon and the need for greater awareness The results reported in this paper have led us
of psychological disturbance and for improved to two main conclusions.
diagnostic skills is a general one. Our results (1) Primary health workers in developing
are therefore in no way an indictment of primary countries do see mental disorders regularly
health workers in developing countries. among their patients. They recognize only a
The rate of diagnosis of mental disorders by minority. It will be important to improve their
the health workers in patients who were not diagnostic skills so that they may recognize the
detected in the research procedure was low, but patients with mental disorders and provide
nevertheless accounted for 97 (6-0 %) patients appropriate management.
in our total sample of 1624. Giel & Le Nobel (2) Improved diagnosis will not, however, be
(1971) found 8-0% of patients falling into this sufficient to reach those patients with severe
category, while Goldberg & Blackwell's (1970) mental disorders, such as schizophrenia or de-
figures suggest a much lower rate of about 1 %. pressive psychosis, who do not normally present
These patients may represent 'false positive' themselves at primary health facilities. Contact
diagnoses by the health worker or, alternatively, with the community will also be needed to
cases missed by the research procedure and change attitudes and to demonstrate that such
correctly diagnosed by the health worker. people can be helped.
Whichever interpretation is more correct, our Each team participating in the study has
data suggest that, at the present time, the tendency drawn up a comprehensive strategy of inter-
to miss cases is operationally more important ventions, including the training of primary
than any risk of overinclusive diagnosis. health workers, a system of referral and super-

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 240 T. W. Harding and others

vision, the supply of drugs, and community tributed to this work. They include members of the
participation (Harding, 1978; WHO, 1979). The study area teams; primary health workers; WHO staff
strategy is based on a number of basic principles in country offices, Regional Offices and at Head-
put forward by a WHO Expert Committee quarters in Geneva. Miss S. Doyle has been res-
(WHO, 1975). It has also been strongly influenced ponsible throughout the study for the design of
instruments, transfer of information and the smooth
by the results reported here and by other obser-
processing of the data. Mr W. Gulbinat (statis-
vations made in the same study areas (Wig et tician), Mr Trung Ngo Khac (statistical clerk) and
al. 1980; Giel et al. in preparation). Mr Lorn Murdoch (computer programmer) expertly
Interventions have now been introduced in guided our relationship with the computer.
the study areas and we have already drawn Note: The WHO Collaborative Study on Strategies
preliminary conclusions concerning the feasibility for Extending Mental Health Care is being carried
of including mental health care at primary level out in 7 geographically defined areas in Brazil,
in developing countries (Harding et al. 1979). Colombia, Egypt, India, Philippines, Senegal and
The final phase of our collaborative work will Sudan and is designed to develop and evaluate
consist of a detailed and objective evaluation of alternative and low cost methods of mental health
these interventions. care (including training methods) in developing
countries.
The authors express their warm thanks to the very
many people in different countries who have con-

APPENDIX
Items of the Self Reporting Questionnaire (SRQ)
' Non-psychotic'
1. Do you often have headaches?
2. Is your appetite poor?
3. Do you sleep badly?
4. Are you easily frightened ?
5. Do your hands shake?
6. Do you feel nervous, tense or worried ?
7. Is your digestion poor?
8. Do you have trouble thinking clearly?
9. Do you feel unhappy ?
10. Do you cry more than usual?
11. Do you find it difficult to enjoy your daily activities?
12. Do you find it difficult to make decisions?
13. Is your daily work suffering?
14. Are you unable to play a useful part in life?
15. Have you lost interest in things?
16. Do you feel that you are a worthless person?
17. Has the thought of ending your life been in your mind?
18. Do you feel tired all the time?
19. Do you have uncomfortable feelings in your stomach?
20. Are you easily tired ?

'Psychotic'
1. Do you feel that somebody has been trying to harm you in some way?
2. Are you a much more important person than most people think ?
3. Have you noticed any interference or anything else unusual with your thinking?
4. Do you ever hear voices without knowing where they come from or which other people cannot hear?

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993
 Mental disorders in primary health care 241

REFERENCES Holmes, J. A. & Speight, A. N. P. (1975). The problem of

non-organic illness in Tanzanian urban medical practice.
Climent, C. E. & Plutchick, R. (1980). Confiabilidad, validez East African Medical Journal 52, 225-236.
y sensibilidad de los ftemes de una escala de auto-reportaje Kessel, W. I. N. (1960). Psychiatric morbidity in a London
de sintomas de enfermedad mental. Revista Colombiana general practice. British Journal of Preventive and Social
de Psiquiatria (in the press). Medicine 14, 16-22.
Dilling, H-, Weyerer, S. & Enders, I. (1978). Patienten mit Kessel, W. I. N. (1962). Conducting a psychiatric survey in
psychischen Storungen in der Allgemeinpraxis und ihre general practice. In The Burden on the Community (ed.
psychiatrische Oberweisungsbediirftigkeit. Monographien J. D.N. Hill), pp. 13-30. Nuffield Provincial Hospitals
aus dem Gesamtgebiele der Psychiatrie (Berlin) 17, 135— Trust: London.
160. Kessel, N. & Shepherd, M. (1962). Neurosis in hospital and
Foulds, G. A. & Hope, K. (1968). Manual of the Symptom general practice. Journal of Mental Science 108, 159-165.
Sign Inventory (SSI). University of London Press: London. Mbanefo, S. E. (1971). The general practitioner and psychi-
Giel, R. & Harding, T. W. (1976). Psychiatric priorities in atry. In Psychiatry and Mental Health Care in General
developing countries. British Journal of Psychiatry 128, Practice (ed. A. Boroffka), pp. 45-49. University of Ibadan:
513-522. Ibadan.
Giel, R. & Van Luijk, V. N. (1969). Psychiatric morbidity in McEvoy, P. J. & McEvoy, H. F. (1976). Management of
a small Ethiopian town. British Journal of Psychiatry 115, psychiatric problems in a Kenyan mission hospital. British
149-162. Medical Journal i, 1454-1456.
Giel, R. & Le Nobel, C. P. J. (1971). Neurotic instability in Mezey, A. G. & Syed, I. A. (1975). Psychiatric illness and
a Dutch village. Ada psychiatrica scandinavica 47, 462- attitudes to psychiatry. Social Psychiatry 10, 133-138.
472. Ndetei, D. M. & Muhangi, J. (1979). The prevalence and
Giel, R. & Workneh, F. (1980). Coping with outpatients clinical presentation of psychiatric illness in a rural setting
who cannot cope - management of persistent complainers in Kenya. British Journal of Psychiatry 135, 269-272.
in an African country. Transactions of the Royal Society of Pemberton, J. (1949). Illness in general practice. British
Tropical Medicine and Hygiene (in the press). Medical Journal i, 306-308.
Giel, R., de Arango, M. V., Climent, C. E., Harding, T. W., Shepherd, M., Cooper, B., Brown, A. C. & Kalton, G. (1966).
Ibrahim, H. H. A., Ladrido-Ignacio, L., Srinivasa Murthy, Psychiatric Illness in General Practice. Oxford University
R., Salazar, M. C. I., Wig, N. N. & Younis, Y. O. A. Press: London.
(in preparation). Childhood mental disorders in primary Verma, S. K. & Wig, N. N. (1977). Standardisation of a
health care. neuroticism questionnaire in Hindi. Indian Journal of
Goldberg, D. P. (1972). The Detection of Psychiatric Illness Psychiatry 19, 67-72.
by Questionnaire. Oxford University Press: London. Wig, N. N., Suleiman, M. A., Routledge, R., Srinivasa
Goldberg, D. P. (1974). Psychiatric disorders (a series on Murthy, R., Ladrido-Ignacio, L., Ibrahim, H. H. A. &
'screening for disease'). Lancet ii, 1245-1247. Harding, T. W. (1980). Community reactions to mental
Goldberg, D. P. & Blackwell, B. (1970). Psychiatric illness disorders: a key informant study in three developing
in general practice. A detailed study using a new method countries. Ada psychiatrica scandinavica 61, 111-126
of case identification. British Medical Journal \\, 439-443. Wing, J. K., Cooper, J. E. & Sartorius, N. (1974). The
Harding, T. W. (1973). The detection of psychiatric illness Measurement and Classification of Psychiatric Symptoms.
by questionnaire in Jamaica. West Indian Medical Journal Cambridge University Press: Cambridge.
22, 190-191. World Health Organization (1975). Organization of Mental
Harding, T. W. (1976). Validating a method of psychiatric Health Services in Developing Countries. Sixteenth Report
case identification in Jamaica. Bulletin of the World Health of the WHO Expert Committee on Mental Health. Technical
Organization 54, 225-231. Report Series No. 564. WHO: Geneva.
Harding, T. W. (1978). Psychiatry in rural agrarian societies. World Health Organization Regional Office for Africa (1979).
Psychiatric Annals 8, 74-84. Mental Health: Report of the WHO Regional Expert Panel
Harding, T. W., Climent, C. E., Collignon, R., Diop, B. S. M., on Mental Health. AFRO Technical Report Series No. 7.
Giel, R., Gueye, M., Ibrahim, H. H. A., Ladrido-Ignacio, Brazzaville.
L. & Wig, N. N. (1979). Sante mentale et soins de sante World Health Organization (in preparation). Provisional
primaires: premiers rcsultats d'une etude concertee de title: Community Mental Health in Developing Countries.
l'OMS. Psyclwpathologie africaine 15 (1), 5-28.

Downloaded from https:/www.cambridge.org/core. Centre Universitaire, on 01 May 2017 at 13:12:09, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0033291700043993