A Reliability Generalization of the Patient Health Questionnaire (PHQ)

A reliability generalization of the Patient Health Questionnaire (PHQ) was conducted to determine the average internal consistency of the measure as well as the variability of reliability coefficients when administered to various populations.  Since reliability is a property of the scores obtained from a measure, and not an immutable aspect of a measure itself, it is useful to assess the degree of variability as well as determinants of higher reliability coefficients.  Reliability is underreported and sometimes misrepresented in research studies.  This study assessed sample and measurement characteristics relating to reliability coefficients and differences among various populations who have taken this assessment.

The PHQ is a self-report diagnostic instrument for mental health disorders used by health care professionals.  Results indicated only 161 studies (11.88%) provided reliability information for their sample; 148 were journal articles (91.93%), and 13 were dissertations (8.07%).  Correlations were run between reliability score and several sample characteristics to assess whether these variables impacted reliability.  Percentage of females was the only significant variable (r = -.255, p = .01), showing the PHQ produces scores with lower reliability when used with female samples.  An ANOVA was run to determine whether there were significant differences in reliability scores produced by the different PHQ versions; the 9-item and 15-item versions were found to produce higher reliability scores than the 2-item or original 27-item version (p = .05).  Overall, results indicated the PHQ is an effective instrument at producing scores with high reliability from diverse samples since its induction in the 1990s.