Psychometric properties of the PHQ-9 depression scale in people with multiple sclerosis: a systematic review

Abstract

BackgroundDepression affects approximately 25% of people with MS (pwMS) at any given time. It is however under recognised in clinical practice, in part due to a lack of uptake for brief assessment tools and uncertainty about their psychometric properties. The 9-item Patient Health Questionnaire (PHQ-9) is an attractive candidate for this role.\n\nObjectiveTo synthesise published findings on the psychometric properties of the 9-item Patient Health Questionnaire (PHQ-9) when applied to people with multiple sclerosis (pwMS).\n\nData sourcesPubMed, Medline and ISI Web of Science databases, supplemented by hand-searching of references from all eligible sources.\n\nStudy eligibility criteriaPrimary literature written in English and published following peer-review with a primary aim to evaluate the performance of the PHQ-9 in pwMS.\n\nOutcome measuresPsychometric performance with respect to appropriateness, reliability, validity, responsiveness, precision, interpretability, acceptability, and feasibility.\n\nResultsSeven relevant studies were identified, these were of high quality and included 5080 participants from all MS disease-course groups. Strong evidence was found supporting the validity of the PHQ-9 as a unidimensional measure of depression. Used as a screening tool for major depressive disorder (MDD) with a cut-point of 11, sensitivity was 95% sensitivity and specificity 88.3% (PPV 51.4%, NPV 48.6%). Alternative scoring systems that may address the issue of overlap between somatic features of depression and features of MS per se are being developed, although their utility remains unclear. However data on reliability was limited, and no specific evidence was available on test-retest reliability, responsiveness, acceptability, or feasibility.\n\nConclusionsThe PHQ-9 represents a suitable tool to screen for MDD in pwMS. However use as a diagnostic tool cannot currently be recommended, and the potential value for monitoring depressive symptoms cannot be established without further evidence on test-retest reliability, responsiveness, acceptability, and feasibility.\n\nPROSPERO register ID: CRD42017067814