BACKGROUND AND PURPOSE:

• The Patient Health Questionnaire-9 (PHQ-9)
  • Nine-item depression screening questionnaire
  • Designed to be used in primary care and other screening settings
  • Standard cut-off positive score: ≥10
• Levis et al. (BMJ, 2019) sought to determine the accuracy of the PHQ-9

METHODS:

• Individual participant data meta-analysis
  • Study included all cut-off scores from all studies (not just published cut-off scores)
• Inclusion Criteria
  • Diagnostic classification for current major depressive disorder or major depressive episode (using validated interviews)
  • Diagnostic classification made within two weeks of PHQ-9 screening (DSM and ICD codes specify symptoms must be present within previous 2 weeks)
  • Participants ≥18 years and not recruited from psychiatric settings or have symptoms of depression
• Three different diagnostic interviews were used
  • Administered by clinicians: Semistructured diagnostic interviews
  • Lay administration: Fully structured interviews
  • Brief structured interview: Mini International Neuropsychiatric (MINI) (which may be overly inclusive)
• Statistical Analysis for sensitivity and specificity were calculated for
  • Cut-off scores 5-15 against the 3 different interview methods
  • Excluding participants who may have received treatment for a mental health issue at time of PHQ-9 administration
  • Subgroup studies based on confounders such as age, sex, country human development index and recruitment setting

RESULTS:

• 58 studies included
  • 17,357 participants | 2,312 with major depression
• Combined sensitivity and specificity (29 studies, 6725 participants) using semi-structured interviews for diagnosis
  • Maximized at a cut-off score of ≥10 or above
  • Sensitivity 0.88 (95% CI, 0.83 to 0.92)
  • Specificity 0.85 (95% CI, 0.82 to 0.88)
• With cut-off scores 5-15
  • Sensitivity was 5-22% higher with semi-structured interviews compared to fully structured interviews (MINI excluded) ,
  • Sensitivity was 2-15% higher compared to the MINI
  • Specificity was similar regardless of interview type
• Subgroup analysis
  • Younger vs older participants: PHQ-9 is similarly sensitive but less specific
  • A cut-off score of ≥10 or higher can be used regardless of age
  • Accuracy higher in primary care settings (sensitivity 0.94, specificity 0.88)

CONCLUSION:

• PHQ-9 performance is dependent on the interview method used to diagnose depression
  • More sensitive than previously reported when diagnosis made using semi-structured interviews administered by professionals
• In primary care settings, the PPV would be 52%, using standard ≥10 threshold
  • Consideration should perhaps be given to using probabilities across all scores rather than a sharp threshold

Learn More – Primary Sources:

Accuracy of Patient Health Questionnaire-9 (PHQ-9) for screening to detect major depression: individual participant data meta-analysis

Related ObG Topics: