Evaluation of the 9-Item Patient Health Questionnaire (PHQ-9) as an Assessment Instrument for Symptoms of Depression in Patients with Multiple Sclerosis

Log in or subscribe to view full content.
Article is also available for purchase the article in one of the available formats.
Kirsten Sjonnesen; Sandy Berzins, MSc; Kirsten M. Fiest, BSc; Andrew G. M. Bulloch, PhD; Luanne M. Metz, MD; Brett D. Thombs, PhD; Scott B. Patten, MD, PhD

Table of Contents

Postgraduate Medicine:

Volume 124 No. 5


Clinical Focus

Purchase this article in one of the formats specified below:

DOI: 10.3810/pgm.2012.09.2595
Background: Patients with multiple sclerosis (MS) have a high prevalence of depression, but there are concerns regarding assessment of possible depression status using rating scales, such as the 9-item Patient Health Questionnaire (PHQ-9). The idea has been proposed that PHQ-9 scores are contaminated by the MS symptoms of fatigue and impaired concentration, decreasing the validity of measurement. Objectives: To determine the extent to which scores on the PHQ-9 are contaminated by patients reporting symptoms attributable to MS. Methods: Baseline PHQ-9 scores from an ongoing prospective cohort study of depression in patients with MS (N = 173) were compared with those of a general population sample (N = 3304). Depression prevalence estimates for the MS and general population samples were calculated using conventional algorithm and cutoff point scoring methods, as well as modified scoring methods, excluding fatigue and concentration deficits. Correlations between scores on adjusted scoring methods were analyzed. The proportion that each item contributed to total PHQ-9 scores was also calculated. A logistic regression model evaluated the relationship between symptom severity and MS status corrected for age, sex, and other depressive symptoms. Results: Conventional PHQ-9 algorithm and cutoff point scoring yielded 2-week prevalence estimates of 9.8% and 21.4%, respectively, in patients with MS, and 3.3% and 8.4%, respectively, in the general population. In both samples, conventional and modified scoring methods were strongly correlated (Spearman rank correlation coefficient > 0.9). The proportion of total scores contributed by fatigue and concentration items was not different between samples. With adjustment for other depressive symptoms, the MS sample had greater odds of endorsement for guilt (odds ratio, 2.17; P = 0.025) and fatigue (odds ratio, 1.51; P = 0.046). Conclusion: Inclusion or exclusion of fatigue and concentration items on the PHQ-9 scale does not substantially alter the performance of the scale. With use of the PHQ-9 in MS populations, we find no evidence to suggest that modified approaches to scoring are necessary.

Keywords: depressive disorders; assessment instrument; multiple sclerosis