This paper used a stratified random sample of offices in the North West Region of Trinidad and showed that the prevalence of depression among consecutive patients attending these offices was 12.8%. Also, this paper determined the commonest reasons for the encounter (RFE) in Trinidadian family practices. Chi-square analysis revealed no statistically significant associations between the presence of depression and either age, gender, ethnicity, level of education achieved, level of occupation, or number of presenting complaints among these family practice patients. Results of binary logistic regression indicated that the independent variables associated with depression were age, with younger persons (aged 39 or less) more likely to be depressed; level of education, with persons achieving only secondary education or below, more likely to be depressed; and marital status.
This paper adds to a growing literature that illustrates the mental health burden facing this population. Among adolescents in Tobago, the sister isle to Trinidad, , adolescents in Trinidad [13, 14], among attendees at chronic disease clinics in the south-west region  and general practice attendees with fatigue  the prevalence of depression was 10%, 28%, 28% and 66% respectively.
The international literature
Depression in primary care populations has been well recognized in developed countries. For example, among over 75 000 patients in the US primary care populations and using a cut-off of 55 on the Zung scale the researchers found the prevalence of significant depressive symptoms to be 20.9% . The lower cut-off point used in that study (55) may account for the differences in prevalence rates found between this study and the present one. Further, it was found that those who perceived their health as poor were more likely to have severe depressive symptoms than patients who perceived their health as excellent. Women, those in older age groups, and those with lower levels of education were more likely to have clinically significant depressive symptoms than men, those in younger age groups, and those with higher levels of education. When classified by marital status within each sex, married men and women were the least likely to have clinically significant depressive symptoms . Similarly in Trinidad among patients attending chronic disease primary care clinics, women, the older patients, those with poorer educational status and those with many presenting complaints were more likely to be depressed .
The RFE can be a useful clinical marker for depression, for example, Gerber et al. , at a teaching medical centre in the USA, studied 1042 consecutive patients screened for depression. The presenting complaints or RFE that discriminated between depressed and non-depressed patients (p < 0.05 level) were sleep disorders, fatigue, the presence of multiple complaints (3+), non-specific musculoskeletal complaints, back pain, patients with amplified complaints, shortness of breath and patients with vaguely stated complaints. While the international research has found clear predictive factors for depression, in the Trinidadian population this lack of statistical associations between depression and common RFE has importance for family practice. This result implies there are no simple markers, except age and educational level achieved, which the Trinidadian family physician can use as a 'red flag' for depression. The family physician has to maintain a high index of suspicion, evaluating each patient on individual merit.
The study is limited by several factors, firstly, by the use of the Zung scale, which is not recognized as a diagnostic, but as a screening scale. This scale was adapted for use in this present study as a brief diagnostic tool because of availability, the researcher's previous experience with the scale and patient's easy understanding of the language. The scale's language had been modified to local idiom prior to testing, and by comparing with a psychiatric interviewer using DSM IV-TR criteria, we come close to a diagnostic scale. The Zung, does not identify manic depressive episodes or dysthymia and so these diagnoses are possibly all part of the spectrum of mood disorders that his study may have detected but not distinguished between. Further work is needed to elucidate these. Secondly, there is always uncertainty as to the representativeness of the sample for the population in question. There are differences between the proportions of the ethnic groups in this study versus the general population. The Central Statistical Office in Trinidad reports that the ethnic mix consists of 40.3 per cent of East Indian descent, 39.5 per cent African, 18.4 per cent Mixed, European 0.6 per cent, Chinese and Other 1.2 per cent. Table 1 illustrates that this study has a larger proportion of persons who describe themselves as 'Mixed'. Thirdly, every effort was made to ensure that the lists used to create the sample frame of practices were up to date and comprehensive. However, since there is no compulsory enforced registration of practices it is difficult to be certain of the representativeness of the family practices selected.
Strengths of this study
Two components of this research build confidence in the results, firstly, the validity of the scale for the high cut-off point used, and secondly, the random sample of offices used through the north-west region of Trinidad. Another mechanism by which we might determine the quality of this work is to measure it against standard criteria for quality when considering prevalence studies. One such set of criteria suggests that 11 of 12 criteria were met . There is no information on non-responders among the patients. In fact this component was not built into the study. It is possible that non-responding patients were depressed and found the questionnaire too intimidating. The research assistants reported informally that 1 patient was so affected. If all the patients responded and no more cases were identified the lowest proportion of depressed patients would be 10.8% (i.e. 65 cases/599 total patients).
Reasons for the encounter (RFE)
The major reason for the encounter (RFE) was the 'check-up'. In an audit published locally, the premier reasons for the encounter were complaints due to the respiratory system (13.3%), gastrointestinal system (12.5%), obstetric and gynaecological problems (11%). Check-ups were 10th on the list accounting for 3.4% of encounters . Similarly research from Barbados found that among the RFE the 'health examination' accounted for 3.6% of health problems . However, US studies suggest that as many as 14.5% of patients attended for a 'General medical examination' . So there are precedents for the results seen in the present study. It may be that 'check-up' may have been an easy response for patients who were uncomfortable discussing personal or difficult issues with a researcher in a public setting of a doctor's waiting room. What evidence do we have for this supposition that many Trinidadian patients have potentially embarrassing problems? Two recent studies may clarify this issue. In 2002, a cross-sectional study of 346 men in 37 Trinidadian family practices found that 53% had some level of erectile dysfunction . Additionally a survey carried out by medical students, also in 2002, revealed that among females attending Trinidadian family practices, their experience of intimate partner abuse was as high as 27% .
The information from this study can be used in several ways. Aside from "check-ups" and joint pains, depression represents the next most common problem in Trinidadian Family Practice. See Table 3. This has important implications for undergraduate medical education, postgraduate training in Family Practice locally and for the delivery of mental health care to the population. Additionally, the scope for future research appears substantial. Questions for future consideration include, what exactly is the nature of this 'check-up'? What is the prevalence of the other mood disorders, such as dysthymia, adjustment disorders or manic-depression which might be subsumed under the diagnosis of depression as detected by this scale? What is the detection rate by physicians? Does the presence of depression have a negative consequence for patients in family practice and primary care and does treatment improve the outcomes for patients in a Trinidadian setting? This has already been shown to hold true in other populations [32–36].
One study from 1999 suggests that dedicated mental health professionals do a better job in caring for these patients . So, should the job of the primary care physicians be recognition and referral? In a developing world setting such as Trinidad, there may never be sufficient mental health professionals to adequately address this issue and so the role of the primary care physician and primary health care worker becomes crucial.
As we move from detection to interventions for depression a recent systematic review of the literature is instructive. This review suggests that the most successful guideline and educational interventions were accompanied by complex organisational interventions. The most impressive results apparently arise from changes within health care systems and 'those with complex interventions that incorporated physician education, an enhanced role of the nurse (nurse-case management), and a greater integration between primary and secondary care (consultation-liaison)'. Telephone medication counselling delivered by practice nurses or trained counsellors was also effective. Simple guideline implementation and educational strategies were found to be generally ineffective . This review is an important one and should guide future interventions and research in the field of depression, both in Trinidad and worldwide.