The Four-Item Patient Health Questionnaire for Anxiety and Depression

Breast Cancer Survivors: (Weihs, Wiley, Crespi, Krull, & Stanton, 2017; n = 410 women; mean age = 56 (12) years; 75% some college; 67% married; 52% employed; stages 1, 2, 3, and 4: 44%, 39%, 11%, and 5%, respectively)

• Development sub sample has moderate agreement (K = .45)

• Validation subsample has moderate agreement where models and thresholds derived from the development process were applied to the holdout validation sample to evaluate performance (K = .50)

Predictive Validity

Breast Cancer Survivors: (Weihs et al., 2017)

• Good predictive validity (AUC = 0.79)

Chronic Migraine: (Bagley et al., 2012; n = 8726; 83.5% female; 86.1% white/Caucasian; headache pain severity mild, moderate, and severe: 1.3%, 19.8%, and 78.9%, respectively

• Excellent: Cronbach's Alpha = 0.87 to 0.96

Chronic Migraine: (Bagley et al., 2012)

• Adequate correlation with Migraine Quality of Life Questionnaire (r = 0.30 to -0.47)

Fibromyalgia: (Hauser et al., 2012; n = 1647; 94.8% female; 77.4% living with partner)

• Adequate correlation with Somatic Severity Score (r = 0.56)

• Adequate correlation with Fibromyalgia Survey Questionnaire (r = 0.48)

Primary Care: (Kroenke, Spitzer, Williams, & Löwe, 2009; n = 2149; aged 18-94; mean age = 47.2 (15.4) years; 66% female; 81% non-Hispanic white; 8% African-American; 8% Hispanic; 64% married; 7% had less than a high school education; 31% had a high school degree or equivalent; 62% had some college)

• An elevated score on the PHQ-4 of 6 or greater is not diagnostic but is, instead, an indicator for further inquiry to establish the presence or absence of a clinical disorder warranting treatment

Pre-operative Surgical Patients: (Kerper et al., 2014; n = 2852; mean age = 47; 52% female; 62.6% living with a partner; 42.9% college experience)

• PHQ-4 scores greater than or equal to 4 are significant for detecting psychological distress

General Population: (Löwe et al., 2010; n = 5036; 53.6% female; mean age = 48.4 (18) years; 59.9% living with a partner)

• PHQ-4 scores of 6 (percentile 95.7%) or greater are “yellow flags” indicating depression or anxiety disorder

• PHQ-4 scores of 9 (percentile 99.1) or greater are “red flags” indicating depression or anxiety disorder

General Population: (Löwe et al., 2010)

Author provided the percentile rank for PHQ-4 scores for individuals aged 14-92. For example, a PHQ-4 score of 3 for any man or woman of any age indicates a percentile rank of 75.5 in the total population.

Mortality Prediction: (Michal et al., 2015; n = 191 patients from N = 1384 died; deceased and alive men were 67% and 64%, respectively; age of deceased and alive men were 74.9 and 69.9 years old, respectively)

College Students: (Khubchandani et al., 2016; n = 934; mean age = 20.32 (2.72) years; 63% female; 80% Whites; 95% enrolled full time; 60% enrolled as 1^st or 2^nd year undergraduates; 9.2% had diagnosis of depression in the past year; 18.6% had a diagnosis of anxiety in the past year)

• Very good agreement for anxiety screener (K = 0.81)

• Good agreement for depression screener (K = 0.70)

Primary Care: (Kroenke et al., 2009)

• Excellent (Cronbach's Alpha = 0.85)

Pre-operative Surgical Patients: (Kerper et al., 2014)

• Excellent (Cronbach's Alpha = 0.83)

College Students: (Khubchandani et al., 2016)

• Excellent (Cronbach's Alpha = 0.81)

General Population: (Löwe et al., 2010)

• Excellent (Cronbach's Alpha = 0.78)

Pre-operative Surgical Patients: (Kerper et al. 2014)

• Excellent predictive validity (AUC = 0.88)

• Lowered cutoff point of less than or equal to 4, sensitivity was 80.5%, specificity was 80.2% for detecting clinically significant psychological distress according to the Global Severity Index (GSI)

College Students: (Khubchandani et al., 2016)

• Excellent predictive validity for anxiety (AUC = 0.835)

• Good predictive validity for depression (AUC = 0.787)

Primary Care: (Kroenke et al., 2009)

• Excellent correlation with the categories of Medical Outcomes Study Short Form General Health Survey (SF-20):

  • Excellent mental health (r = 0.80)

  • Adequate social functioning (r = 0.52)

  • Adequate general health perceptions (r = 0.48)

  • Poor role functioning (r = 0.37)

  • Poor bodily pain (r = 0.36)

  • Poor physical functioning (r = 0.36)

Pre-operative Surgical Patients: (Kerper et al. 2014)

Good discriminant validity indicated by:

• Poor age (r = 0.05)

• Poor gender (r = 0.13)

• Poor partnership status (r = -0.10)

• Poor education (r = -0.04)

• Poor physical health (r = 0.03)

Adequate to good convergent validity has moderate to strong correlations with:

• Good Brief System Inventory (BSI) depression scale (r = 0.68) 

• Good BSI anxiety scale (r = 0.63)

• Adequate BSI phobic anxiety scale (r = 0.46)

• Good BSI interpersonal sensitivity scale (r = 0.56)

• Good Global Severity Index (r = 0.69)

• Good PHG single item of subjective psychosocial impairment (r = 0.70)

College Students: (Khubchandani et al., 2016)

• Principal components analysis (PCA) using varimax rotated component-matrix for 4 screening items (2 items each for PHQ-2 and GAD-2) indicated that 83% variance was explained by the 2 components. 2 anxiety screening items had the highest loading on component 1 and the 2 depression screening items had highest loading on component 2.

General Population: (Löwe et al. 2010)

• Adequate correlation with the Rosenberg Self-Esteem Scale (r = -0.49 to -0.40)

• Adequate correlation with the Questionnaire on Life Satisfaction (r = 0.39 to -0.39)

• Adequate correlation with the Resilience Scale (r = 0.35 to -0.28)