A comparison of the PRIME-MD PHQ-9 and PHQ-8 in a large military prospective study, the Millennium Cohort Study

Abstract

Background

In light of increased concerns about suicide in the military, institutional review boards have mandated increased scrutiny of the final item on the depression screening tool, the PHQ-9, which asks about suicidal thoughts. Since real-time monitoring of all individual responses in most observational studies is not feasible, many investigators have adopted the PHQ-8, choosing to remove the ninth item. This study compares the performance of the PHQ-8 with the PHQ-9 in a population-based sample of military or nonmilitary subjects.

Methods

The Millennium Cohort Study administers a self-reported questionnaire that includes the PHQ-9 at 3-year intervals to current and former U.S. military personnel. PHQ-9 responses of 143,705 Millennium Cohort members were investigated. Cross-sectional comparisons of the PHQ-9 and PHQ-8 and prospective analyses to detect a 5-unit change in these measures were performed.

Results

Greater than substantial agreement was found between the PHQ-8 and 9 instruments (kappas, 0.966–0.974 depending on survey cycle). There was similarly high agreement between the PHQ-8 and 9 in detecting a 5-point increase (κ=0.987) or decrease (κ=0.984) in score.

Limitations

One potential limitation of this study is that participants completed the PHQ-9, and PHQ-8 scores were extrapolated from the PHQ-9. In addition, the Millennium Cohort may not fully represent the U.S. military; though previous evaluations have shown the cohort to be a well-representative sample.

Conclusions

Since excellent agreement was detected between the PHQ-8 and PHQ-9 instruments, the PHQ-8 would capture nearly all the same cases of depression as the PHQ-9 in populations similar to the one in this study.

Introduction

The Primary Care Evaluation of Mental Disorders Patient Health Questionnaire (PHQ) is a standardized instrument that provides an assessment of mental health status based on scores of several health concepts (Spitzer et al., 1999, Spitzer et al., 2000, Spitzer et al., 1994). A 9-item scale from this instrument (PHQ-9) used to screen for major depressive disorder (Spitzer et al., 1999), has been shown to have high sensitivity (0.93) and specificity (0.89) (Fann et al., 2005), and correlates well with a diagnosis of depression, as outlined in the Diagnostic and Statistical Manual of Mental Disorders (2000). The ninth and final item on the scale asks about thoughts of being better off dead or hurting oneself, which is known to be related to suicidal thoughts and ideation (Corson et al., 2004). Exclusion of the final item results in the PHQ-8, which has also been shown to be a valid instrument for evaluating depression symptoms in specific populations (Kroenke et al., 2009).

The Millennium Cohort Study (Ryan et al., 2007, Smith, 2009), the largest population-based longitudinal cohort study in military history, has included the PHQ-9 as part of the standard questionnaire since the study launched in 2001. Symptoms of depression have previously been examined in this population using the PHQ-9 (Ryan et al., 2007, Wells et al., 2010), with positive screens for new-onset depression occurring in approximately 4% of men, and 8% of women at the time of the first follow-up questionnaire (Wells et al., 2010). However, in light of increased concerns about suicide in the military (Kuehn, 2009, Oquendo et al., 2005), institutional review boards have mandated increased scrutiny of this item with the intent of initiating provider referral when respondents positively endorse it. In nonclinical research settings, many investigators have adopted the PHQ-8, choosing to remove the final item that may indicate suicidal thoughts. Since real-time monitoring of more than 150,000 responses from Millennium Cohort members is not feasible, starting in 2011 the final item was removed from the questionnaire and the study currently uses the PHQ-8 to screen for depression. Although the PHQ-8 has been shown to have similar operating characteristics as the PHQ-9 (Kroenke and Spitzer, 2002), this was done using a sample of 6000 individuals seeking treatment in primary care or obstetrics-gynecology clinics (Kroenke et al., 2001). To our knowledge, no study has validated the PHQ-8 in a large, population-based cohort, military or otherwise. The objective of this study was to compare the PHQ-8 with the PHQ-9 to understand differences in depression screening capability between these two instruments in the Millennium Cohort Study, which may be generalizable to other similar population-based studies. Unique features of this study include investigating differences in a large military population that is probably healthier than a general population sample. Also, with the large sample size, subgroups such as differences by sex can be explored to determine performance of the PHQ-8 compared with the PHQ-9.