Depression is a major public health problem that affects patients and society [1, 2]. The clinical condition named as unipolar depression, is characterized by depressed mood, hopelessness, helplessness, intense feelings of guilt, sadness, low self-esteem, thoughts of self-harm and suicide [3]. The estimates of global disease burden ranked depression as the fourth leading cause of disease burden in the year 2000, accounting for 4.4% of total Disability Adjusted Life Years (DALYs) [4]. Depression is known as one of the most prevalent yet treatable mental disorders presenting in general medical as well as specialty settings [5, 6]. About one in ten patients seen in the primary care settings suffers from some form of depression [7, 8]. According to a systematic review, depression substantially increases the risk of death, and gives rise to other chronic disease conditions such as cardiovascular disease [9].

The World Mental Health Survey reported that 15% of the population from high-income countries compared to 11% from low and middle-income countries were likely to get depression over their lifetime [10]. Globally, 5.5% reported having an episode of depression in the previous year. Studies from the South Asian region show varying prevalence of depression possibly due to differences in the study settings and instruments used to measure it. For example, the age-adjusted prevalence of depression in an urban South Indian population was 15.9% according to Patient Health Questionnaire-12 (PHQ-12) [11], in contrast to 45.9% in urban Pakistan according to depressive symptom questionnaire [12]. It was found to be 29% in rural Bangladesh using the Montogomery and Aasberg Depression Rating Scale (MADRS) [13]. In Sri Lanka, there is a scarcity in mental health research, and only few studies have reported depression and its correlates [14]. According to the national mental health survey of Sri Lanka conducted in 2007, depression was assessed using the Patient Health Questionnaire-9 (PHQ-9), and the prevalence of major depression was 2.6% [15].

The effects of war on mental health have been documented in previous literature, and the common war-related mental health conditions included post-traumatic stress disorder (PTSD), anxiety and depression [16–18]. Studies in the post-conflict populations showed a definite increase in the incidence and prevalence of mental disorders, with women being more affected than men. Prevalence rates were associated with the degree of trauma, and the availability of physical and emotional support [19]. A study in Sudan after 20 years of war reported a high prevalence of mental disorders, for example, over one third of respondents met symptom criteria for PTSD, and half of respondents met symptom criteria for depression [20].

In Sri Lanka, civilians in the Northern Province were affected by 30-year long armed conflict that ended in May 2009. A household survey in 2009 among residents of Jaffna district in the Northern Province revealed a substantially high prevalence of symptoms of war-related mental health conditions, which were significantly associated with displacement status and underlying trauma exposure. In this survey, the overall prevalence of PTSD, anxiety, and depression were 7.0%, 32.6% and 22.2% respectively [21]. A high prevalence of common mental disorders (18.8%) was reported among internally displaced persons, particularly where displacement was prolonged, with major depression at 5.1% and other depressive syndromes at 7.3% [22]. A study in school children in areas affected by the armed conflict reported that the great majority of children experienced severely traumatizing events such as combat, bombing, shelling, or witnessing the death of a loved one, and their performance and functioning were related to the total load of traumatic events experienced [23]. According to a recent qualitative study in Northern Sri Lanka, complex mental health and psychosocial problems at the individual, family and community levels in a post-war context were found to impair recovery [24]. These studies indicate that more efforts related to psychological health are needed to re-establish the normalcy in the region despite the re-settlement of the internally displaced persons in the Northern province, together with infrastructure and livelihood development programmes by the state and non-state sectors.

Improving services of primary health care facilities would provide easy access to a wider population for their basic health needs, and reduce overcrowding of larger hospitals. The Primary Medical Care Units and Divisional Hospitals (PMCU/DH) in the Northern Province have been strengthened after the conflict, with the aim of providing first contact care to treat minor illnesses and screen and refer patients for health care institutions at higher level. We assume that the prevalence of depression and other psychosocial issues would be higher in a post-conflict population that seeks care from these facilities in the Northern Province than in the rest of the country. The aim of this paper is to report on the burden of depression and describe the factors associated with depression in a post-conflict population, using a large sample of adult patients accessing primary medical care settings in the Northern Province in Sri Lanka.

Using a large sample of 12,841 individuals, we report that the prevalence of major depression is 4.5% in adult patients attending primary care settings in the Northern Province, four years after the end of a protracted 30-year armed conflict. Prevalence of mild depression is almost 3-fold higher in the study population (13.3%). Furthermore, the study reveals that older individuals, women and persons with disability are at a greater risk for depression.

To our knowledge, this is the first study that assessed depression within a population accessing primary care settings in Sri Lanka. The study design therefore yielded towards a public health approach to understand depression [34, 35]. From the health services point-of-view, knowledge on burden of depression in a population that seeks care from primary care settings than in the general population would be more beneficial to provide targeted services. Further, the post-conflict nature of the study population provides a unique situation and critical evidence for policy and action.

It is noteworthy that the study finds higher prevalence of major depression (4.5%) in primary health care settings in the Northern province than the national estimate of 2.6% of major depression reported in 2007 [15]. This may be due to the fact that the national survey excluded conflict affected districts in the Northern province (due to inability of research teams to access those areas due to ongoing conflict), and that it was a community-based survey rather than a study among those seeking health care. Our estimate of major depression is somewhat closer to findings of a recent study which found major depression at 5.1% (95% CI: 3.2-7.7) in an internally displaced community due to war in the Northern province [22]. However, contextual difference and lower precision of the estimate in the latter study may limit the comparability of findings.

Depression can aggravate existing illnesses, signs and symptoms and vice-versa. Therefore, it is important to identify patients at high risk for depression so that these patients could be targeted for depression screening and treatment. Knowledge about risk factors or predictors of depression can ease identification of these patients. Regarding factors associated with depression, advancing age represents the strongest factor in our analysis, showing a linear trend with age. Although, an incremental increase of odds for depression with increasing age is observed in the univariable analysis, this stepwise pattern is not reflected in the multivariable analysis. Instead, the multivariable analysis revealed a greater but approximately equal odds (each age group showed an OR of around 5), for depression relative to the youngest age group of 18–24 years. This could be interpreted as younger age group being relatively at less risk to all other age groups in the context of other risk factors such as sex, and disabilities. Previous findings support the evidence that age contributes significantly to the prediction of depression [10, 36, 37]. While older adults may face widowhood, loss of function, or loss of independence, depression is not a ‘normal’ symptom of aging [38]. Studies show that depression that initially appears later in life is linked to a more chronic course of illness [39, 40]. Living with untreated depression presents a serious public health problem since it may complicate chronic conditions such as heart disease, diabetes, and stroke; often accompanies functional impairment and disability and leads to increased health care costs [41, 42]. Depression among older adults can be addressed through better community-based approaches for identifying and treating depression, and through more public awareness programmes [40, 43, 44].

Females are at a higher risk for depression than males; however the odds are marginal in our results. Many previous studies found a similar pattern [10, 30], while some studies could not reveal significant sex differences [22, 37]. The district differences of depression could be attributed to many external factors that were not included in the present study. There are differences across districts in the extent of re-settled population, livelihood development programmes, employment opportunities and access to health and other social services, that would affect the risk of depression [32]. Our study found a strong positive association between major depression and disabilities, however the cross-sectional nature of the study design limits conclusion whether disability led to depression or vice-versa. A cohort study of 6247 subjects 65 years and older in USA, who were initially free of disability has revealed that depression in elder patients caused limitation in activities of daily living and mobility after 6 years of follow-up [45]. This excess risk is partly explained by depressed persons’ decreased physical activity and social interaction. Further, there is evidence that improved depression by treatment reduces disability days and disability scores in depression persons [46]. Since depression and disabilities go hand-in-hand, further evidence through randomized trails would be needed to see effects of reduction of one on another.

Trauma and potential exposure to traumatic events due to protracted civil conflict appear to be associated with adverse mental health symptoms [47]. A study conducted among residents in Jaffna district in Sri Lanka in the aftermath of war revealed that the prevalence of symptoms of war-related mental health conditions was substantial and significantly associated with displacement status and underlying trauma exposure [21]. The same study found that approximately 68% of Jaffna residents experienced at least 1 trauma event and most individuals experienced multiple traumas. Furthermore, a dose–response relationship between the number of trauma events and psychiatric morbidity was evident, and chronic exposures to trauma events corresponded with higher levels of PTSD, anxiety, and depression symptoms. A qualitative inquiry into the psychosocial situation among internally displaced persons concluded that the collective trauma, i.e., traumatic psychological effect shared by community, can be profound [48]. However, the present study failed to collect data on the previous trauma exposure among individual participants. Despite, we can confirm that all communities which were serviced by the selected primary care facilities were affected by collective trauma.

The major implication of our findings on the health system is the importance of ensuring support to primary health care services for early detection and referral of common mental health conditions. Thus, training and sensitization of primary health care personnel at PMCU and DH can indeed make a difference for early detection of common mental health disorders, especially major depression. Capacity of the health personnel at primary care level should be enhanced in the area of mental health. These personnel should be trained to diagnose depression at primary health care level and to conduct initial basic management including counseling. It is also necessary to explore the referral system and continuity of care which were not addressed by the present study. Previous literature highlights that treatment gap and stigma are major barriers for communities to seek care at the primary health care settings for mental and psychological illnesses [49, 50]. Therefore, further research is needed to address these issues such as treatment gap and stigma on mental illnesses. Since the prevalence of any form of depression including the mild forms, is high, we recommend that community-based mental health programmes be strengthened to increase knowledge and skills of community level workers to deal with common mental health and psychosocial issues and psychosocial problem solving. Our findings support the recommendations of a recent qualitative study aiming to rebuild family and community agency and resilience [24].

A number of limitations of our study has to be noted. First, though many studies demonstrate that the PHQ-9 has proven to be a sensitive and specific measure, the final diagnosis needs to be confirmed by a clinical assessment. Our study adopted a health systems perspective where the assessments occurred in routine primary care settings. Obtaining a definitive diagnosis for all those entering primary health care centres for treatment would be highly resource intensive, and warrant dedicated health programs from health system. Second, we did not cover all potential predictor variables due to the fact that the data collection was done during the history taking time of each patient at the out-patient department, consecutively on all patients. Our aim was to reveal the best determinants out of few easily obtainable parameters such as age, sex, occupation and disability etc. A major limitation of our study was the exclusion of the trauma scales. Due to the sensitive nature of the questions, administration of the trauma exposure scale was considered as ethically unsound and not approved by the health and administrative authorities of the Northern Province. Finally, this may not be representative of the population of the Northern Province, since the sample was obtained from primary health care facilities. Despite the mentioned weaknesses, hitherto this study is the largest study that has assessed depression among those seeking primary health care in Sri Lanka covering a large segment of the population. The quality of data was maintained at a high degree, especially with the PHQ-9 as described by Kroenke and Spitzer [26].

In conclusion, this study reports that the prevalence of major depression is 4.5% (95% CI: 4.1-4.9) in adult patients attending primary care settings in the Northern Province, and that the older individuals, women and persons with disability are at a greater risk for depression. Prevalence of mild depression is almost 3-fold higher in the study population. The results indicate that the services at primary health care settings should be strengthened to screen patients for depression and advise accordingly. Strengthening community mental health services is necessary to detect psychological issues early and manage such issues promptly.