Globally, it is estimated that 5% of adults suffer from depression, with a high burden in low- and middle-income countries (LMICs) (World Health Organisation, 2021). Social, psychological and biological factors all contribute to the development of depression, through complex interactions. After the COVID-19 pandemic, the disability adjusted life years for depressive disorder was 49.5 million globally, highlighting a very high burden of disease (Santomauro et al, 2021).
However, in high income countries, more than 50% of potential depression cases go undetected (Mitchell et al, 2009). Without a diagnosis of depression, patients are unable to access appropriate treatment, which may in turn prevent them from recovering and regaining their quality of life. Therefore, the detection of depression, particularly in primary care settings, is crucial to reducing the global burden of depression.
This blog summarises a recent systematic review exploring detection of depression in primary care settings in LMICs (Fekadu et al, 2022).
Fekadu and colleagues searched 6 databases, alongside a manual search, to identify papers written in any language which explored detection of depression by clinicians in primary care settings, in LMIC, compared to a “gold” standard diagnosis. Their review focussed on adults and studies were included from the inception of the databases up to December 2020.
Data from studies that met the inclusion criteria were extracted by two independent reviewers. Risk of bias was assessed using the Effective Public Health Practice Project (EPHPP) and quality assessed by using the STROBE checklist.
A meta-analysis was performed, stratified by the different diagnostic approaches taken in the included papers. Estimates of depression prevalence were also pooled.
The database searched retrieved 5,577 articles, which reduced to 3,159 titles after duplicated removed. 85 articles were included in the full text review, with a total of 9 different publication types included. Two multi-country studies were included resulting in 12 individual country level reports that had relevant data. Two reports originated from Malawi, Nigeria and India, whilst Ethiopia, Palestine, Nepal, South Africa, Uganda and Turkey had one.
Three studies were assessed by the EPHPP as being of strong quality, 5 of moderate quality, and 3 as weak. The STROBE checklist assessed all studies as either moderate or high.
The authors found that in 4 of the reports, the detection rate was 0%. The pooled detection from PHQ-9 with a cut-off of 5 was 3.9%, compared to 7% when a cut-off of 10 was used for the same tool. One study used CIDI as gold standard and found a detection rate of 28.4%, whilst a study in male participants found 69% detection. A study in pregnant women found a lower detection rate of 1.4%.
In relation to the prevalence of depression, the results varied by tool used to assess depression. Prevalence in PHQ-9 with a cut-off of 10, were lower than a cut off of 5, as would be expected. Pooled prevalence was as follows:
- PHQ-9 (cut off 10) – 13.2% (95% CI 8.2% to 18.2%)
- PHQ-9 (cut off 5) – 38.2% (95% CI 35.7% to 40.7%)
- SCID – 21.6%
There were 2 studies that looked at specific populations (diabetes and pregnant women): the prevalence for these studies were 40.8% and 7.3% respectively. A study using CIDI estimated the rate as 11.6%.
This review highlights that there is a paucity of evidence exploring detection of depression in primary health care in LMICs (low- and middle-income countries). The evidence that is available indicates that detection in primary care settings is low, with clinicians often missing diagnoses of depression in their patients.
Strengths and limitations
The review employed an extensive search strategy across several relevant databases which should have ensured that most of the available published evidence was captured by this review.
Similarly, the authors stratified their results by the different screening tools used to confirm the presence of depression. By doing so, the differences between tools can be identified and ensures that direct comparison between tools does not occur, due to different questions being asked, or different sensitivities, which would invalidate the comparisons made.
A limitation of this paper is the lack of evidence available. Of course, the authors are unable to mitigate this, but it does present difficulties when interpreting the findings. LMICs are diverse in many different ways and therefore, the results found in some LMICs cannot necessarily be applied to others. There were no studies from South America, Europe or Oceania. This issue may be a particular issue in the countries studied, and therefore this would overestimate the rates of under detection. It reiterates the need for further evidence in LMICs.
Acknowledged in the paper as a limitation is that the authors only searched for peer reviewed papers, which could lead to publication bias.
The authors do not state that they followed the PRISMA guidelines, so it is not known if they did. They do not state their screening process for full text articles, and therefore it is not known how this was done.
Implications for practice
For me, the key implications from this study are:
- There needs to be a greater focus on detecting depression in LMICs, as without it, individuals are prevented from getting the relevant and necessary treatment
- Although it is often undetected, there are high prevalence rates of depression in LMICs
- The lack of research identified in this study reveals that there has been little attention to this research area in published papers.
This research is important for researchers, clinicians and policymakers alike in both LMICs and High Income Countries (HICs). Clinicians should be provided more training on detecting mental health problems, but should also themselves be more aware of this as a possible diagnosis and ask the appropriate questions. Researchers should take note of the paucity of evidence on this topic, and try to fill this evidence gap, if possible, as this would allow policy-makers to truly understand the extent of the problem. The problem of depression, and other mental health conditions in LMICs is considerably underestimated; policy-makers need to acknowledge this and introduce policies that can reduce the stigma about being diagnosed with depression, and also ensure clinicians know how to detect this ever-increasing public health issue.
Detection in HICs has shown to be much higher, but this is still considered to be an under-detection of the true rates of depression. Policy-makers in the UK could use this review to evaluate their own mental health policies, and whether detection of depression is as good as it could be, whilst recognising what could be done better.
However, I do think it is important to consider that depression can be difficult to diagnose and if patients come to primary care settings with other illnesses, this makes it even harder. The review does not discuss if the papers included were exploring individuals who were attending the health setting with just depressive symptoms or whether it was not the main focus of their visit.
Statement of interests
No conflicts of interest.
Fekadu A, Demissie M, Birhane R. et al. (2022). Under detection of depression in primary care settings in low and middle-income countries: a systematic review and meta-analysis. Systematic Reviews, 11(1), 21. doi.org/10.1186/s13643-022-01893-9
Mitchell AJ, Vaze A, Rao S. Clinical diagnosis of depression in primary care: a meta-analysis. Lancet. 2009:374(9690):609–19.
Santamauro DF et al. Estimating the global prevalence and burden of depressive and anxiety disorders in 2020 due to the COVID-19 pandemic. The Lancet. 8 October 2021. doi: 10.016/S0140-6736(21)02143-7.
World Health Organisation. 2021. Depression. (accessed 26/03/23).
Hi – nice blog but few typos? Malaria or Malawi?
Thanks! We’ve corrected that now 😊