Psychometric properties of MERIT: A concise tool for community health workers to screen priority mental disorders

ABSTRACT Background: Task shifting is an accepted way to address the treatment gap for mental health disorders. Community Health Workers (CHWs) can play vital roles through screening, referral, and providing basic counseling. To address the dearth of a concise screening instrument for CHWs, the “ M ental H E alth Sc R een I ng T ool for Community Health Workers in India [MERIT]” was designed and developed. In this paper, we have examined the psychometric properties of MERIT. Methods: MERIT consists of nine stems (consisting of 11 questions) covering the following five domains: (a) Substance Abuse (alcohol, tobacco), (b) Anxiety, (c) Depression and Somatoform symptoms, (d) Severe Mental Illnesses, and (e) Suicidal risk. MERIT can be applied to one key informant of the household to screen for all its adult members. Mental health professionals (MHPs) with community psychiatry expertise undertook the face and content validity exercise (for both English and Kannada versions; 17 and 11, respectively) and rated the ‘adequacy’ (ability to pick up a mental health issue) of each of the questions on a Likert scale ranging from 0 (strongly disagree) to 4 (strongly agree). For each of the domains, the item-content validity index (I-CVI) of more than 0.8 was considered adequate. Concurrent validity was examined through concordance between the screened status as per Accredited Social Health Activists (ASHAs; quintessential CHWs; n = 16; n = 116 households) and the independent clinical impression derived by seven MHPs. Inter-rater reliability (IRR) was examined by measuring the concordance between the two groups of ASHAs (first group; n = 7, second group; n = 9), who independently screened the same 116 households. Results: All experts (for both English and Kannada versions) gave a rating of 3 or more for each of the MERIT domains. I-CVI of each domain was > 0.8 (0.86 for English and 0.80 for Kannada). Concurrent validity [n = 116; k = 0.792 indicating substantial agreement between ASHAs and MHPs; P < 0.001] and IRR [n = 115, k = 0.744; P < 0.001] were excellent. The specificity (97.50%), positive predictive value (92.4%), and negative predictive value (86%) were high. The sensitivity was 62.5%. On average, 5 minutes were required to administer MERIT. Conclusion: MERIT is concise, valid, and reliable for CHWs to screen for priority psychiatric disorders among adults. While its sensitivity is comparatively lower, it can be easily integrated into public health programs as it offers the advantage of covering the entire household by interviewing one of its key informants.