Table 1.
Description and Results of Included Studies
| Study | Administration | Gold standard | Setting | Participants | Biometric properties | Cronbach's αa | Blindingb/limitations | Risk of bias |
|---|---|---|---|---|---|---|---|---|
| PHQ-9 | ||||||||
| Wang et al. (2014)49 | Self | MINI; within 2 weeks | Mainland China; urban community; 2011–2012 | 1045; adults 16 and over | Ideal cutoff value: 7 Sensitivity: 0.86 Specificity: 0.86 AUCb: 0.92 [0.86–0.97] Test/retest correlation coefficient (2 weeks): 0.86c |
0.86 | Blinding is unknown. | L |
| Yuan (2019)43 | Self | MINI | Mainland China; hospital; 2013–2015 | 782; adult patients in the coronary care unit with diagnosis of Acute Coronary Syndrome | Ideal cutoff value: 10 Sensitivity: 0.869 Specificity: 0.847 AUC: 0.842 [0.806–0.894] Test/retest correlation coefficient: not reported |
0.837 | Blinded. Administration of gold standard relative to index is unknown. |
I |
| Yeung et al. (2008)25 | Self | bCB-SCID-I/P; within 4 weeks | United States; Urban outpatient clinics; 2004–2005 | 1940; adults 18 and over; Chinese Americans | Preset cutoff value: 15 Sensitivity: 0.81 Specificity: 0.98 AUC: 0.97 Test/retest correlation coefficient: Not reported |
0.91 | Not blinded. | L |
| Zhang et al. (2013)42 | Self | MINI | Hong Kong; outpatient; 2010–2011 | 586; Chinese patients between 25 and 75 years of age with type 2 diabetes | Ideal cutoff value: 7 Sensitivity: 0.826 Specificity: 0.737 AUC: 0.85 [0.76–0.94] Test/retest correlation coefficient (2–4 weeks): 0.70d |
0.86 | Blinding is unknown. Administration of gold standard relative to index is unknown. |
L |
| Ye et al. (2020)44 | Self, with personnel guidance if needed | American DSM-V; immediately | Mainland China; outpatient and inpatient; 2018 | 148; patients with psoriasis | Ideal cutoff value: 9 Sensitivity: 0.98 Specificity: 0.908 AUC: 0.979 [0.968–0.991] Test/retest correlation coefficient (1 week): 0.955c |
0.938 | Blinding is unknown. Exclusion criteria are unclear. |
I |
| Chen et al. (2013)46 | Self | SCID; within 2 weeks | Mainland China; primary care clinics; 2009–2010 | 2639; adult patients 18 and over | Ideal cutoff value: 10 Sensitivity: 0.87 Specificity: 0.81 AUC: 0.91 [0.87–0.94] Test/retest correlation coefficient (2 weeks): 0.76c |
0.89 | Blinded. 7% dropout rate. |
L |
| Du et al. (2017)48 | Self (online) | MINI; within 48 h | Mainland China; university | 230; adult university students 18 and over | Ideal cutoff value: 10 Sensitivity: 0.74 Specificity: 0.85 AUC: 0.897 [0.823–0.970] Test/retest correlation coefficient (2 weeks): 0.78e |
0.80 | Blinded. | L |
| Peng et al. (2020)45 | Self | SCID | Mainland China; outpatient; 2019 | 258; adults with acne between ages 18 and 24 | Preset cutoff value: 9 Sensitivity: 0.957 Specificity: 0.886 AUC: 0.973 [0.956–0.990] Test/retest correlation coefficient (1 week): 0.824c |
0.851 | Blinding is unknown. Administration of gold standard relative to index is unknown. 8% dropout rate. |
I |
| PHQ-2 and PHQ-9 | ||||||||
| Liu et al. (2016)40 | By postgraduate students and psychiatrists | SCID-I; immediately | Mainland China; rural community; 2010–2011 | 839; adults 60 and over | Ideal cutoff value for PHQ-9: 8 Sensitivity: 0.97 Specificity: 0.89 AUC: 0.97 [0.96–0.98] Test/retest correlation coefficient for PHQ-9: Not reported Ideal cutoff value for PHQ-2: 3 Sensitivity: 0.90 Specificity: 0.90 AUC: 0.94 [0.90–0.97] Test/retest correlation coefficient for PHQ-2: Not reported |
PHQ-9: 0.82 PHQ-2: 0.76 |
Not blinded. | L |
| Xiong (2015)47 | Self | MINI | Mainland China; outpatient clinics; 2011–2012 | 491; adults 18 and over with multiple somatic symptoms | Ideal cutoff value for PHQ-9: 10 Sensitivity: 0.77 Specificity: 0.76 AUC: 0.82 [0.77–0.86] Test/retest correlation coefficient for PHQ-9: not reported Ideal cutoff value for PHQ-2: 3 Sensitivity: 0.77 Specificity: 0.74 AUC: 0.81 [0.77–0.86] Test/retest correlation coefficient for PHQ-2: not reported |
0.90 | Blinding is unknown. Administration of gold standard relative to index is unknown. |
I |
| Zhang et al. (2013)50 | Self | SCID; within 2 weeks | Mainland China; University | 959; Chinese college students | Ideal cutoff value for PHQ-9: 11 Sensitivity: 0.89 Specificity: 0.97 AUC: 0.977 [0.966–0.988] Test/retest correlation coefficient (4 weeks) for PHQ-9: 0.873f Ideal cutoff value for PHQ-2: 3 Sensitivity: 0.81 Specificity: 0.96 AUC: 0.939 [0.911–0.967] Test/retest correlation coefficient (4 weeks) for PHQ-2: 0.829c |
PHQ-9: 0.854 PHQ-2: 0.727 |
Blinded. 13.8% dropout rate. |
L |
| Xia (2019)51 | Self | MINI; within 1 day | Mainland China; outpatient, 2018 | 213; Chinese adult patients with epilepsy | Ideal cutoff value for PHQ-9: 7 Sensitivity: 0.8286 Specificity: 0.8427 AUC: 0.888 [0.838–0.927] Test/retest correlation coefficient for PHQ-9: Not reported Ideal cutoff value for PHQ-2: 2 Sensitivity: 0.7714 Specificity: 0.7528 AUC: 0.802 [0.742–0.853] Test/retest correlation coefficient for PHQ-2: not reported |
PHQ-9: 0.86 | Blinded. | L |
| Chen (2010)53 | Self, with assistance from nurse as needed | SCID | Mainland China; primary care clinics; 2008 | 364; adult patients 60 and over | Ideal cutoff value for PHQ-9: 9 Sensitivity: 0.86 Specificity: 0.85 AUC: 0.92 [0.88–0.96] Test/retest correlation coefficient for PHQ-9: not reported Ideal cutoff value for PHQ-2: 3 Sensitivity: 0.84 Specificity: 0.90 AUC: 0.92 [0.87–0.97] Test/retest correlation coefficient for PHQ-2: not reported |
PHQ-9: 0.91 | Blinded. Administration of gold standard relative to index is unknown. 20% dropout rate. |
L |
| Liu et al. (2011)41 | Self | SCAN | Taiwan; community-based primary care clinics; 2007–2008 | 1954; adult patients 18 and over | Ideal cutoff value for PHQ-9: 10 Sensitivity: 0.86 Specificity: 0.939 AUC: 0.96 [0.93–0.98] Test/retest correlation coefficient (2 weeks): 0.87g Ideal cutoff value for PHQ-2: 2 Sensitivity: 0.88 Specificity: 0.82 AUC: 0.90 [0.85–0.95] Test/retest correlation coefficient for PHQ-2: not reported |
PHQ-9: 0.80 | Blinded. Administration of gold standard relative to index is unknown. 21.6% dropout rate. |
L |
| Studies published in Chinese language journals | ||||||||
|---|---|---|---|---|---|---|---|---|
| PHQ-9 | ||||||||
| Jin et al. (2011)39 |
Self |
SCID |
Mainland China; community |
1275; adult patients 60 and over |
Preset cutoff value: 10 Sensitivity: 0.8831 Specificity: 0.825 AUC: not reported Test/retest correlation coefficient: not reported |
0.767 |
Blinding is unknown. Administration of gold standard relative to index is unknown. |
I |
| Xu et al. (2007)38 |
Self, with guidance from family or research staff if illiterate |
SCID; within 1 week |
Mainland China; community |
622; adult patients 60 and over |
Preset cutoff value: 15 Sensitivity: 0.88 Specificity: 0.99 AUC: not reported Test/retest correlation coefficient for PHQ-9 (1 week): 0.934 |
0.8325 |
Blinding is unknown. |
I |
| Yang et al. (2015)37 |
Self |
SCID |
Mainland China; outpatient |
258; psychosomatic outpatients between 16 and 65 |
Preset cutoff value: 10 Sensitivity: 0.98 Specificity: 0.67 AUC: not reported Test/retest correlation coefficient: not reported |
Not reported |
Blinded. Administration of gold standard relative to index is unknown. |
L |
| Zhi et al. (2013)36 |
Self |
MINI; immediately |
Mainland China; outpatient |
2009; adult patients between 18 and 65 |
Ideal cutoff value: 8 Sensitivity: 0.856 Specificity: 0.802 AUC: 0.903 [0.869–0.937] Test/retest correlation coefficient: not reported |
0.855 |
Blinding is unknown. 0.6% dropout rate. |
L |
| Chen et al. (2015)35 |
Self, with personnel guidance if needed |
MINI; immediately |
Mainland China; outpatients Department of Psychiatry; 2012–2013 |
319; adult patients 18 and over |
Cutoff values reported: 6 (mild); 12 (moderate); 15 (severe) Sensitivity: not reported for mild depression; 0.94; 0.83 Specificity: not reported for mild depression; 0.82; 0.90 AUC: 0.94 Test/retest correlation coefficient: not reported |
0.89 |
Blinding is unknown. |
L |
| PHQ-2 and PHQ-9 | ||||||||
| Wang et al. (2015)33 | Self | CIDI; immediately | Mainland China; outpatients from Psycho-cardiological Department; 2013–2014 | 201; Adult patients 18 and over | Ideal cutoff value for PHQ-9: 10 Sensitivity: 0.871 Specificity: 0.835 AUC: 0.877 Test/retest correlation coefficient for PHQ-9 (1 week): 0.882h Ideal cutoff value for PHQ-2: 2 Sensitivity: 0.857 Specificity: 0.692 AUC: 0.806 Test/retest correlation coefficient for PHQ-2 (1 week): 0.813e |
PHQ-9: 0.809 PHQ-2: 0.785 |
Blinded | L |
a
As a measurement of internal consistency.
b
“Blinded” refers to the interviewers not knowing the PHQ-2 and PHQ-9 scores before conducting the clinical interview.
c
Conducted after 2 weeks.
d
Conducted after 2 weeks on 6.82% of the participants (40/586).
e
Conducted after 2 weeks on 65.2% of the participants (150/230).
f
Conducted after 4 weeks on 10% of the participants (121/959).
g
Conducted after 2 weeks on 13.1% of the participants (256/1954).
h
Conducted after 1 week on 24.9% of the participants (50/201).
ACS, Acute Coronary Syndrome; AUC, area under the curve; bCB, Chinese Bilingual version of the PHQ-9; CIDI, Composite International Diagnostic Interview; DSM-V, American Diagnostic and Statistical Manual of Mental Disorders, fifth edition; I, indeterminate; L, low; MINI, Mini International Neuropsychiatric Interview; PHQ, Patient Health Questionnaire; SCAN, Schedule for Clinical Assessments in Neuropsychiatry; SCID, Structured Clinical Interview for Diagnostic and Statistical Manual Disorders.