Recognising a mental disorder such as depression in patients with physical illness is important in clinical management and improving patients’ clinical outcomes. In this population of adult type 2 diabetes mellitus (T2DM) patients attending non communicable diseases (NCD) clinics in Lilongwe, Malawi, we investigated the prevalence and rate of routine detection of depression, finding that 18% of participants met criteria for major depression, none of whom were detected in routine practice by the NCD clinicians.
The prevalence of major depression in T2DM in our study is comparable to prevalence estimates of 19% and 30% found in two studies of general primary care outpatient attendees in Malawi (20, 21). There is some overlap with the populations in the previous studies as these were conducted in facilities that did not have dedicated NCD clinics, so will have included patients with T2DM. The finding of our study that one-fifth of patients with T2DM (18%) suffered from major depression is similar to the findings of a systematic review by Roy and Lloyd where the they found a prevalence of 19.1% (range 6.5–33%) (4) however the finding is considerably higher compared to that of a collaborative study carried out in 14 countries which found a prevalence of 10.6% (34). The rate of major depression in this study is also comparable to rates of depression in studies from Nigeria with a prevalence of 19.4% based on the SCID (35) and from Ethiopia with a prevalence of 17 % on the Beck Depression Inventory (BDI) (36).
In contrast to our study, several studies in other African countries have shown higher prevalence of depression among T2DM patients. A study in Morocco found that patients with T2DM had higher prevalence of major depression (33.1% ) (37)while studies in South Africa found and in Egypt found 46% and 32% respectively (35). Similarly a systematic review conducted in Ethiopia showed that the prevalence of depression among T2DM patients was 39.73% (38). In other LMICs the findings of higher prevalence of major depression are also similar. For instance a multi-centre study conducted in Pakistan found that 43.5% of T2DM patients had depression (39). A Mexican study also found a higher prevalence of 48.27% (40) compared to our study.
In our study, the majority of patients (34.4%) had mild depression, 22.6% had moderate and 2.2% had severe depression based on PHQ-9 scores. Our finding is similar to findings in other LMICs, for instance a study in Iran that showed that 38% of patients had mild depression, 30% had moderate depression, and 13% had severe depression on the BDI (41). Further, the findings are also similar to a study in Tanzania which showed that 22.1% had mild depression, and 8.2% had moderate depression while none had severe depression based on the PHQ-9(6). An Indian study also found almost similar finding which showed that 25% had mild depression, 12.5% had moderate depression, and 1.3% had severe depression on the Major Depression Inventory (MDI)(7). In contrast to our study, a Bangladesh study found that majority of patients (20.2%) had severe depression rather than mild to moderate depression (14.6%) using the Centre for Epidemiological Studies Depression Scale (42). One of the reasons could be because of the relatively small size of the data sample in the Bangladesh study compared to our study and the other reason could be the use of different tools.
Despite the high prevalence of depression, none of the patients who were diagnosed with major depression by the research team using the SCID were detected and diagnosed as such by the NCD clinician. This finding agrees with other studies conducted among patients with T2DM (8-12). The reasons for non-detection of depression in this setting could be due to lack of knowledge and skills, lack of integration of mental health in NCD care and lack of routine screening for depression.
This failure to detect and manage depression will affect management of T2DM as evidenced in other studies (11, 12). Several studies have demonstrated that depression treatment effectively reduces the severity of depression in patients with diabetes and also produces better glycaemic control (43), significantly improve HbA1C (44) restore mental health and improve medical outcome (45). As such if depression in patients with diabetes is not recognized, it will go untreated and consequently the medical and mental health outcomes will be negatively affected. Detecting depression in the clinical setting is aided by validated tools that can be easily administered with limited resources. Improved detection of depression through the use of the PHQ-9 by the clinicians would lead to the initiation of treatment or referral to mental health services for treatment which may consequently lead to improved outcomes and lower complications. Therefore, NCD clinicians have to be equipped with the necessary knowledge and skills in detection and treatment.
Based on multivariable logistic regression, the sociodemographic factors including age, gender, education, marital status and employment were not associated with depression. Related studies have gotten varying results with many similarly finding no association with this set of factors (1, 13, 46) (47-51); but, some individual studies showing associations with them (52, 53) (54, 55). In the present study, three quarters of the participants were females. However, it is unlikely that this had an influence on the results for females as there was no evidence of a difference in the odds of depression between the two genders. The over representation of females in our study is similar to other studies like the Malawi STEPS Survey (56), which may be due to the increased health care seeking behaviour by females relative to males.
One notable strength of this study is that it used a gold standard, the SCID for depression that had previously been used in different studies in Malawi. Furthermore, the quality of administration of the PHQ-9 and of the SCID was very good as the research assistants had training in administration of the research tools and received regular supervision from the principal investigator. To our knowledge, this is the first study analyzing depression among patients with diabetes in Malawian population.
The results of this study should however be interpreted in light of limitation. The limitation of this study is that the participants were selected through a consecutive sampling technique from two specialized NCD clinics in Lilongwe which may not be representative of the wider population. However, despite this limitation, the paper fills an important knowledge gap regarding the prevalence and detection of depression among patients with diabetes in Malawi, given the dearth of evidence in LMIC settings. In this regard, the paper shows a crucial gap and has the potential to stimulate policy makers and clinicians to develop interventions to improve detection of depression among patients with diabetes.