Assessing cognition in people with severe mental disorders in low- and middle-income countries: a systematic review of assessment measures.

BACKGROUND

Cognitive difficulties are common in people with severe mental disorders (SMDs) and various measures of cognition are of proven validity. However, there is a lack of systematic evidence regarding the psychometric properties of these measures in low- and middle-income countries (LMICs).

OBJECTIVE

To systematically review the psychometric properties of cognitive measures validated in people with SMDs in LMICs.

METHODS

We conducted a systematic review of the literature by searching from four electronic databases. Two authors independently screened studies for their eligibility. Measurement properties of measures in all included studies were extracted. All eligible measures were assessed against criteria set for clinical and research recommendations. Results are summarized narratively and measures were grouped by measurement type and population.

RESULTS

We identified 23 unique measures from 28 studies. None of these was from low-income settings. Seventeen of the measures were performance-based. The majority (n = 16/23) of the measures were validated in people with schizophrenia. The most commonly reported measurement properties were: known group, convergent, and divergent validity (n = 25/28). For most psychometric property, studies of methodological qualities were found to be doubtful. Among measures evaluated in people with schizophrenia, Brief Assessment of Cognition in Schizophrenia, Cognitive Assessment Interview, MATRICS Consensus Cognitive Battery, and CogState Schizophrenia Battery were with the highest scores for clinical and research recommendation.

CONCLUSIONS

Studies included in our review provide only limited quality evidence and future studies should consider adapting and validating measures using stronger designs and methods. Nonetheless, validated assessments of cognition could help in the management and allocating therapy in people with SMDs in LMICs.