Ïã¸ÛÁùºÏ²Ê¼´Ê±¿ª½±

Older Primary Care Patients: 

(Lowe, Unutzer, et al, 2004, n = 434, mean age = 71 (7.4) years, all participants enrolled in the Improving Mood-Promoting Access to Collaborative Treatment (IMPACT), Older Primary Care Patients)

• SEM for for change due to treatment and no control of prior depression = 2.44 
• SEM for the the same number of DSM-IV depressive symptoms at both assessments = 1.32

Older Primary Care Patients:

(Lowe, Unutzer, et al, 2004, Older Primary Care Patients)

• MCID = 5 points

Chronically Ill Elderly Patients:

(Lamars, et al., 2008; n = 106; mean age = 71.4 (6.9); 51% diagnosed with DM, Chronically Ill Elderly Patients) 

• Excellent Test-retest reliability (correlation = 0.91)

Older Primary Care Patients:

(Lowe, et al, 2004; at baseline and 6 months, Older Primary Care Patients)

• Excellent test-retest reliability for change due to treatment and no control of prior depression (ICC = 0.81) 
• Excellent test-retest reliability for same number of DSM-IV depressive symptoms at both assessments (ICC = 0.96)

Older Adults: (Chen et al., 2016; n=82)

• Excellent test-retest reliability: (ICC = 0.79)

Older Adults

• Adequate internal consistency (Cronbach¡¯s alpha = 0.77) (Chen et al., 2016; n=634)
• Adequate internal consistency (Cronbach¡¯s alpha = 0.725) (Zhang et al., 2019; n= 1546)

New Nursing Home Admits (Belanger et al, 2019; n = 1,734,785; 65.09% were cognitively intact, 21.1% had mild impairment, 13.43% had moderate impairment, and .26 had severe impairment)

• Excellent internal consistency: (McDonald¡¯s omega = 0.87)

Long-Stay Nursing Home Cohort (Belanger et al., 2019; n = 122,949; 4.66% were cognitively intact, 20.7% had mild impairment, 38.95% had moderate impairment, and 32.51 had severe impairment)

•  Excellent internal consistency: (McDonald¡¯s omega = 0.87)

Concurrent Validity

Older Adults: (Chen et al., 2016; n=634)

Concurrent validity of the PHQ-9 with the Hamiltyom Rating Scale for Depression (HAMD-17) and Short Form of the Quality of Life Enjoyment and Satisfaction Questionnaire (Q-LES-Q-SF)

P Value <0.001

Criterion Validity

Older Adults (Chen et al., 2016; n=634)

• Excellent criterion validity for major depressive disorder (AUC of ROC was 0.97)

New Nursing Home Admits (Belanger et al, 2019; n = 1,734,785)

• Adequate criterion validity for depression for a PHQ-9 score of 10 or above (r= 0.54)

Long-Stay Nursing Home Cohort (Belanger et al., 2019; n = 122,949)

• Excellent criterion validity for depression for a PHQ-9 score of 10 or above (r= 0.67)

Construct Validity

Older Adults: (Chen et al., 2016; n=634)

Construct Validity of PHQ-9 individual item loadings for factor analysis (n=634)

Meta-analysis:

(Gilbody, et al, 2007; 14 validated studies reviewed; n = 5,026 participants, Meta-analysis)

• Clinicians and researchers may elect to adjust cut-scores in response to clinical population characteristics

Coronary Heart Disease:

(Thombs, et al, 2008; n = 1024, Coronary Heart Disease) 

• Cutoff score of > 6 for Major Depressive Disorder, with 83% sensitivity and 76% specificity

(McManus, et al, 2005; n = 1024, Coronary Heart Disease) 

• Cutoff score of > 10 for Major Depressive Disorder with 54% sensitivity and 90% specificity

Patients with Mental Health or Somatic Complaints:

Wittkampf, et al, 2009; n = 664; mean age = 49.8; Diagnosed as depressed based on Structured Clinical Interview for DSM-IV Axis I Disorders = 12.3%; Dutch sample, Patients with Mental Health or Somatic Complaints) 

• For screening a cutoff > 10 points demonstrated the highest sensitivity and specificity, however, the PHQ-9 was not found to be specific enough to be used for diagnostic purposes among populations at the highest risk.

General Medical Population:(Lowe et al, 2004, n = 501; mean age = 41.7 (13.8) years; participants were recruited from outpatient clinics and 12 family practices; German sample, General Medical Population) 

• The PHQ-9 was found to be more accurate in diagnosing ¡®major depressive disorder¡¯ than either the Hospital Anxiety and Depression Scale (HADS) or the Well Being Index (WBI-5) 
• Suggested cut-off to indicate depression is PHQ-9 > 11 

General Medical Population

• Suggested cut-off to indicate depression is PHQ-9 > 10 (sensitivity 0.88; specificity 0.85) (Levis et al, 2019; Studies n=29; participants n=6725; major depression n=924)
• Optimal cut-off score of 10 (Sensitivity?=?81%; specificity?=?79%; PPV?=?59% and NPV?=?92%) (Daray, F. M. et. al, 2019; n = 257)

Primary Care Patients:

(Zuithoff, et al., 2010; n = 1338; mean age = 51, Primary Care Patients) 

• Excellent Test-retest reliability (Correlation = 0.94) 

Primary Care Patients:

(Kroenke, et al, 2001; n = 3000; mean age = 46 (17) years; most common medical conditions = Hypertension (25%) and Arthritis (11%), Primary Care Patients)

• Excellent test-retest reliability over a 48 hour period (r = 0.84)

Coronary Heart Disease:

(Stafford, et al, 2007; n = 193, Coronary Heart Disease)

• Excellent internal consistency (Chronbach's alpha = 0.90) 

Outpatients:

(Lowe, Spitzer, et al, 2004; n = 501; mean age = 41.7 (13.8) years; German sample, Outpatients)

• Excellent internal consistency (Cronbach¡¯s alpha = 0.88)

Primary Care Patients:

(Kroenke, et al, 2001, Primary Care Patients)

• Excellent internal consistency (Chronbach's alpha > 0.86)

(Zuithoff, et al. 2010, Primary Care Patients) 

• Excellent internal Consistency (ICC = 0.88)

General Medical Population (Daray, F. M. et. al, 2019; n = 257)

• Excellent internal consistency (Cronbach's ¦Á value of 0.86)

Undergraduate University Students: (McCord et al., 2020; n = 231)

• Adequate internal consistency (Cronbach's ¦Á value of 0.76)\
• Poor internal consistency (inter-item correlation coefficient = 0.27)

Primary Health Care (with Full PHQ):

(L?we et al 2004, n = 501, mean age = 41.7 (13.8) years; German sample); common health complaints included: diseases of the musculoskeletal system and connective tissue (21%), endocrine, nutritional and metabolic diseases (16%), cardiovascular/circulatory diseases (10%), Primary Health Care (with Full PHQ))

General Medical Population (Daray, F. M. et. al, 2019; n = 257)

• Adequate criterion validity for major depressive episode (AUC of ROC was 0.87)

Coronary Heart Disease:

(Stafford et al, 2007; n = 193, Coronary Heart Disease)

• Moderate correlation between PHQ-9 and Hospital Anxiety Depression Scale with cutoff score > 6 for major depressive disorder (r = 0.72)

General Population:

(Martin, et al; 2006; mean age = 48.8 (18.1); 53% female and 47% male, General Population) 

• Excellent correlations between PHQ-9 and BDI (r = 0.73, p < 0.001) 
• Excellent correlations between PHQ-9 and GHQ-12 (r = 0.59, p < 0.001) 

Patients with Mental Health or Somatic Complaints:

(Wittkampf et al, 2009, Patients with Mental Health or Somatic Complaints)

• Adequate correlations between PHQ-9 total score & Hamilton Depression Rating Scale (HDRS-17) score (r = 0.52, two tailed P < 0.01)

Undergraduate University Students: (McCord et al., 2020; n = 231)

• Adequate Convergent Validity between PHQ-9 scores and Emotional/Internalizing Dysfunction subscale of Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF) (r?=?.48)
• Poor Convergent Validity between PHQ-9 scores and Thought Dysfunction subscale of MMPI-2-RF (r?=?.25)
• Poor Convergent Validity between PHQ-9 scores and Behavioral/Externalizing Dysfunction subscale of MMPI-2-RF (r?=?.24)
• Adequate convergent validity between PHQ-9 and MMPI-2-RF RCd-Demoralization (r = 0.59), RC2-Low Positive Emotions (r = 0.33), and RC7-Dysfunctional Negative Emotions (r = 0.44)

• The PHQ-9 is derived from the Primary Care Evaluation of Mental Disorders interview schedule which utilizes criteria set forth in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) to diagnose depression

• Rasch analysis suggests the PHQ-9 is a unidimensional measure of depression (Williams, et al, 2009)

• The PHQ-9 has been found to be accurate in medical settings (Thompson et al, 2011)

Depression:

(Lowe, et al, 2006; n = 1,788 (diagnosed with: Major Depressive Disorder (MDD) n = 757, minor depression n= 543, other depressive disorders  n = 438 patients); Mean age 50.3 (14.7) years, Depression)

• On the 0 to 27 scale, the mean change of for this population was 10.3 (5.6) points over a 12 week period.

Dementia:

(Hanock et al, 2009; n = 113; mean age = 69 (11.7) years, Dementia)

• Cut off point of 9 which coincides with the threshold between mild and moderate depression

Dementia:

(Hanock et al, 2009, Dementia)

• Demented group PHQ-9 scores:
  • Mode = 0
  • Median = 2
  • Mean = 4.1 +/- 5.4
• Non demented group PHQ-9 scores:
  • Mode = 0 
  • Median = 3.5 
  • Mean = 7.8 +/- 7.9

Dementia:

(Hanock et al, 2009, Dementia)

• Area under the receive operating characteristic curve (ROC) = 0.63
• Shows that this test has poor diagnostic accuracy

Dementia:

(Hanock et al, 2009, Dementia)

• Floor effects = 30%, participants reported no depressive symptoms

Parkinson¡¯s Disease:

(Williams et al, 2012; n = 229; mean age = 66.0 (10.8) years, Parkinson's Disease)

• Cut off score of ¡Ý6 for PHQ-9, with 66% sensitivity and 80% specificity 

(Thompson et al, 2011; n = 214; mean age = 72.5 (9.6) years, Parkinson's Disease)

• Cut off score of ¡Ý 5 for major depression for PHQ-9 
• Cut off score of 2-4 for minor depression for PHQ-9

Patient with Parkinson¡¯s Disease: (Chagas et al., June 2013; n=110)

• Cut-off score of 9 for major depressive disorder (sensitivity of 100% and specificity of 83.1%)

Parkinson¡¯s Disease:

(Williams et al, 2012, Parkinson's Disease)

• Active depressive disorder is probable with PHQ-9 scores of 8.9 (5.2)
• No active depressive disorder is probable with a PHQ-9 score of 3.8 (3.8)

Parkinson¡¯s Disease:

(Thompson et al, 2011, Parkinson's Disease)

• Adequate interrater reliability 95%CI = 0.4 (0.26, 0.54) between PHQ-9 and SCID

Parkinson¡¯s Disease:

(Williams et al, 2012, Parkinson's Disease)

• Excellent internal consistency (Cronbach¡¯s alpha = 0.85)

Patient with Parkinson¡¯s Disease: (Chagas et al., June 2013; n=110)

• Excellent: Cronbach's alpha of 0.83

Predictive Validity:

Parkinson¡¯s Disease: (Chagas et al., June 2013; n=110)

• Excellent criterion validity for depression for PHQ-9 (AUC of ROC was 0.94)

Concurrent validity:

• Excellent: the PHQ-9 correlated with the Zung Self-rating Depression Scale (SDS) and the 15-item Geriatric Depression Scale (GDS-15) with Spearman¡¯s coefficient of 0.630 (p<0.001) for both scales

Stroke:

(de Man-Van Ginkel et al, 2012; n = 55; mean age = 65.09 (15.03) years; mean time since stroke onset = 59.82 (77.43) days, Stroke)

• The optimum cut-off value for the PHQ-9 was 10 with 100% sensitivity and 86% specificity 

(de Man-Van Ginkel et al, 2012; n = 164; mean age = 70.6 (13.99) years; mean time since stroke onset = 6.7 (0.9) weeks, Stroke)

• Accuracy of the PHQ-9 was best at a cutoff score of greater than or equal to 10 with a sensitivity of 0.80 and specificity of 0.78

(Prisnie et al., 2016; n= 122; mean age = 60.1 (15.6) years; mean time since stroke onset = 0.3 years [0.18-0.88])

• > 13 indicates risk of depression (sensitivity 81.8%; specificity 97.1%)

Stroke:

(de Man-Van Ginkel et al, 2012, Stroke)

• Excellent test-retest reliability for the agreement between the pairs of nurses on the sum score level (ICC = 0.98)

Stroke:

(de Man-Van Ginkel et al, 2012, Stroke)

• Excellent interrater reliability (ICC = 0.98)

Stroke:

(de Man-Van Ginkel et al, 2012, Stroke)

• Excellent internal consistency (Chronbach¡¯s alpha = 0.79)

Stroke:

(de Man-Van Ginkel et al, 2012, Stroke)

• Concurrent validity between PHQ-9 and Geriatric Depression Scale (GDS-15) was excellent with r = 0.7 and P < 0.001 
• Area under the receive operating characteristic curve (ROC) = 0.87 (95%CI 0.80-0.93)
• Discriminatory power for the PHQ-9 was adequate

(Prisnie et al., 2016; n= 122; mean age = 60.1 (15.6) years; mean time since stroke onset = 0.3 years [0.18-0.88])

• Adequate criterion validity for depression at cut-off point of 13 (AUC of ROC was 0.895)

Traumatic Brain Injury:

(Fann, et al, 2005; n = 478; mean age = 42 (17.9) years; mean time since TBI = 3.8 (2.8) months, TBI) 

• PHQ-9 cutoff > 12 was the best screening criteria for Major Depressive Disorder (MDD) 

(Cook, et al. 2011; n = 365; mean age: 43 (17.7); 1 year post-injury, TBI) 

• Minimal: 0-4 
• Mild: 5-9 
• Moderate: 10-14 
• Moderately severe: 15-19 
• Severe: >20

TBI Patients: (Donders & Pendery, 2017; n = 137; mean age = 40.42 (15.03) years; Median PHQ-9 score = 9 (IQR = 0 to 25); assessed within 1 to 12 months post TBI)

• Cut off score of 10 demonstrated best sensitivity and specificity (sensitivity = .917, specificity =.602)

Traumatic Brain Injury:

(Fann, et al, 2005, TBI)

• Excellent test-retest reliability within 7 of initial assessment (r = 0.76)

TBI Patients: (Dyer et al., 2016; n = 100)

• Excellent person separation reliability (0.78)

TBI Patients: (Dyer et al., 2016; n = 100)

• Excellent internal consistency (Cronbach¡¯s alpha = 0.82-0.84)

TBI Patients: (Donders & Pendery, 2017; n = 137)

• Adequate criterion validity for diagnosis of major depression (AUC of ROC was 0.759)

TBI Patients

• Excellent convergent validity between the PHQ-9 and the Symptom Checklist-20 (r = 0.85) (Dyer et al., 2016; n = 100)
• Excellent convergent validity between the PHQ-9 and the Hamilton Depression Rating Scale (HAM-D) (r = 0.81) (Dyer et al., 2016; n = 100)
• Excellent convergent validity between the PHQ-9 and the Maier subscale of HAM-D (r = 0.82) (Dyer et al., 2016; n = 100)
• Excellent convergent validity between the PHQ-9 and the Bech subscale of HAM-D (r = 0.80) (Dyer et al., 2016; n = 100)
• Adequate convergent validity between the PHQ-9 and the Patient Global Impression scale - improvement (r = 0.49) (Dyer et al., 2016; n = 100)
• Excellent convergent validity between the PHQ-9 and the Structured Clinical Interview for DSM-IV (r = 0.78) (Dyer et al., 2016; n = 100)
• Excellent convergent validity between the PHQ-9 and the Minnesota Multiphasic Personality Inventory-2-Restructured Form Demoralization (r = 0.64) (Donders & Pendery, 2017; n = 137)
• Adequate convergent validity between the PHQ-9 and the Minnesota Multiphasic Personality Inventory-2-Restructured Form Low Positive Emotions (r = 0.48) (Donders & Pendery, 2017; n = 137)

(Donders & Pendery, 2017; n = 137)

Logistic Regression Model for PHQ-9 Elevations

Traumatic Brain Injury:

(Fann, et al, 2005, TBI)

• Excellent convergent validity between the PHQ-9 and SCL-20 (Hopkins Symptom Checklist depression subscale; r = 0.90)
• Excellent convergent validity between the PHQ-9 and HAM-D (Hamilton Rating Scale for Depression; r= 0.78)
• Excellent discriminant validity between the PHQ-9 and SCL-20 (r = 0.84)
• Excellent discriminant validity between the PHQ-9 and HAM-D (r = 0.67, p < 0.001)

SCI:

(Krause, et al, 2009; n = 727, 53.3% had cervical injuries; mean age 47.9 years, time since injury 18.2 years, Chronic SCI)

• Mean PHQ-9 score = 5.57 (5.74)
• Mean Older Adult Health and Mood Questionnaire (OAHMQ) score =  6.0 (5.0)

(Bombardier, et al, 2004; Norms based on Kroenke, et al. 2001, Chronic SCI)

• Mean PHQ-9 score = 5.48 (95% CI = 5.07¨C5.88)
• Major Depressive Disorder (MDD) is probable with PHQ-9 scores > 18.1 (3.9)

Spinal Cord Injury: (Summaka et al., 2019; n=51)

• Excellent Reliability test-retest reliability (ICC = 0.88)

SCI:

(Bombardier, et al, 2004, Chronic SCI)

• Excellent internal consistency (Chronbach's alpha = 0.87)
• First assessment completed by the patient, while the second was conducted over the phone 48 hours later

Spinal Cord Injury

• Adequate: Cronbach's alpha of 0.71 (Summaka et al., 2019; n=51)
• Excellent: Cronbach¡¯s alpha of 0.87 (Dorstyn et al., 2019; n=48 mean age= 42)
• Excellent: Cronbach¡¯s alpha of 0.82 (Williams et al., 2016; n=133)

SCI:

(Krause et al, 2009, Chronic SCI)

• Excellent correlations between the PHQ-9 and Older Adult Health and Mood Questionnaire (OAHMQ) scale (r = 0.78)
• Adequate correlations between PHQ-9 & prevalence of Major Depressive Disorder (r = 0.530)

(Bombardier et al, 2004, Chronic SCI)

• Adequate convergent validity between PHQ-9 scores and:
  • Life satisfaction (r = -0.51, p < 0.001) 
  • Subjective health (r = -0.50, p < 0.001)

Convergent Validity:

Spinal Cord Injury:

• Excellent convergent validity between the PHQ-9A total scale and the (HDRS-A) Hamilton Depression Rating Scale Arabic version (r=0.713, p<0.001). (Summaka et al., 2019; n=51)
• Excellent convergent validity between the PHQ-9 and the Hopkins Symptom Checklist-20 (r = 0.78) (Williams et al., 2016; n=133)
• Excellent convergent validity between the PHQ-9 and the Hamilton Depression Rating Scale (HAM-D) (r = 0.68) (Williams et al., 2016; n=133)
• Excellent convergent validity between the PHQ-9 and the Maier subscale of HAM-D (r = 0.67) (Williams et al., 2016; n=133)
• Excellent convergent validity between the PHQ-9 and the Bech subscale of HAM-D (r = 0.69) (Williams et al., 2016; n=133)
• Adequate convergent validity between the PHQ-9 and the Patient Global Impression scale - improvement (r = 0.66) (Williams et al., 2016; n=133)
• Excellent convergent validity between the PHQ-9 and the Structured Clinical Interview for DSM-IV (r = 0.79) (Williams et al., 2016; n=133)
•  

Discriminatory Validity

Spinal Cord Injury: (Summaka et al., 2019; n=51)

• Adequate discriminatory validity between the PHQ-9A total scale and the (HDRS-A) Hamilton Depression Rating Scale Arabic version indicating that the PHQ-9-A has greater discriminatory power (AUC of ROC was 0.83)

SCI:

(Williams, et al, 2009; n = 202; mean age = 42.6 (13.9) years; > 1 (Range = 1 to 44) years post injury, Chronic SCI) 

• 22% of participants reported no depressive symptoms

Spinal Cord Injury: (Williams et al., 2016; n=133)

• Adequate floor effects = 5% reported no depressive symptoms

Spinal Cord Injury: (Williams et al., 2016; n=133)

¡ñ  Williams et al., stated that the comparison of the PHQ-9 to the Hopkins Symptom Checklist-20 (HSCL-20), and the Hamilton Depression Rating Scale (HAM-D) in people with SCI diagnosed with major depressive disorder, the PHQ-9 detected a small effect size change over the 12 weeks of the randomized controlled trial.

• Effect Size: Small Change: effect size= -.06