Accuracy of Community Informant Led Detection of Maternal Depression in Rural Pakistan.

Maternal depression is a global mental health and a public health priority. Despite the priority its active detection is still a challenge. We tested the accuracy of an adapted version of Community Informant Detection Tool for Maternal Depression (CIDT-MD) in rural settings of Pakistan. Using a single stage design, trained community informants (lady health workers and lay peers) identified women (pregnant and/or with children) with symptoms of probable depression using CIDT-MD. This was immediately followed by diagnostic interviews of all the women using the Structured Clinical Interview for the Diagnostic and Statistical Manual (SCID-V) for current major depressive episode by trained assessors, blinded to the outcome of CIDT-MD. Data were analyzed using Statistical Package for Social Sciences (Version 25.0, IBM Corp., Armonk, NY, USA) and FACTOR software (Version. 10.3.01, Virgili University, Tarragona, Spain). Descriptive statistics, factor analysis, validity, reliability and known group validity was conducted to evaluate the psychometric properties of the adapted CIDT-MD. In all, 425 women, with mean age of 28 years (SD = 4.7), participated. Nearly 10% were illiterate, while the rest (90%) had an education ranging from eight to 15 years of schooling. The majority (73.2%) of the participants had 1-3 children while only 17.4% had >3 children. The sensitivity and specificity of CIDT-MD in detecting depressive symptoms was 97.5% (95% CI: 94.2-99.1) and 82.4% (95% CI: 77.8-86.4) respectively. It's positive predictive value (PPV), 77.3% (95% CI: 72.9-81.2) and the negative predictive value (NPV) was 98.17% (95% CI: 95.7-99.2). While factor analysis revealed high inter-item correlation for most items (0.62-0.77) with an adequately fair Kaiser-Meyer-Olkin (KMO) sampling adequacy (0.73), significant Bartlett's test of sphericity (p < 0.001). Uni-dimensionality for the CIDT-MD based on one-dimensional congruence (0.97), explained common variance (0.85), excellent internal consistency (0.90), good criterion validity (Area Under Curve = 81%), tester-test reliability (0.87-0.89) and statistically significant known group analysis (p < 0.001). The adapted version of the Community Informant Detection Tool for Maternal Depression is a valid and a reliable tool for active case detection of maternal depression in rural settings of Pakistan.