• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

Appendix G Table 1Evidence table of screening trials-KQ1 and KQ1a.


USPSTF Quality
Study design
Unit of randomization
Point of randomization Setting
Intervention and control conditionsDepressive disorders identified or targeted

% Currently treated for depression

Population targeted
CONSORT numbersPopulation characteristics

Baseline depression scores
Inclusion and exclusion criteriaScreening instrument
Mode of administration
Who administered and scored
How results sent to clinician
Diagnostic work-up
Depression outcomes
Other health outcomes
Adverse events
Group selected for analysis

Similarity to all randomized

Similarity at baseline
P values
NS=p>0.05, *p< 0.05, **p<0.01
KQ1 trial
Williams et al 199910


Consecutive patients in study clinics recruited on designated days

Patients randomized before screen to longer screen, brief screen, or no screen. Diagnostic phone interview after visit.

Several different outpatient medical clinics
I1: Completed CES-D before scheduled appointment and results placed in chart on bright orange form(for everyone--not only positive screens)

I2: Same as above but used 1-item depression screen

C: Usual care (no screen)

All providers received guide for managing depression in primary care and continuing education session on interpreting case-finding questionnaire and diagnosing depression

All patients had post-visit diagnostic interview via phone
MDD, Dysthymia, Minor Depression

% Currently treated:
San Antonio, TX and Washington, D.C.

General outpatient primary care
1,083 approached
1037 eligible
969 consented
969 screened
969 randomized

I1= 323
I2= 330
C= 316

863 complete baseline data
I1= 296
I2= 291
C= 276
Mean Age:
I= 59
I= 58
C= 56

I1= 74%
I2= 68%
C: 71%

I= 61%
I2= 59%
C= 58%
I1= 9%
I2= 12%
C= 10%

Yrs educations:
I1= 10
I2= 11
C= 11

% income <$7,200:
 I1=42%; I2=40%;
% income ≥ $16,800
 I1=24%; I2=25%;

% any depression diagnosis: I1=12.4%;
I2=13.8%; C=13.8%
Inclusion: scheduled appt at one of study clinics

Exclusion: No telephone; no stable address
I1: CES-D; I2: “Have you felt depressed of sad much of the time in the past year?”

Paper-and-pencil, in medical clinic before appointment

Who administered/scored:

Results placed in chart on bright orange sheet of paper

Patient contacted by phone by the researcher after visit to complete diagnostic interview for DSM-IIIR diagnosis
3-month, San Antonio clinics only. Only followed up patients with diagnosis confirmed by post-visit phone interview + random sample of non-depressed per phone interviewDepression: %
depressed; % ≤1
DSM-IIIR depression symptoms

Other: None

Only those in follow-up group (see follow-up time frame)


Combined I1 & I2;
Compared I vs. C on % depressed, % ≤1
DSM-R, # depressive symptoms. Specific test NR.
% Depressed @ 3-mo (of all followed-up):
I=37% (56/153)
C=46% (30/65)

% Recovered (of those with depression diagnosis per post-visit phone interview)*:
I=48% (32/67)
C=27% (8/30)

Mean symptom count reduction:
I=1.6, C=1.5 p=0.21

Results similar when only analyzed those who had followup (not sure how missings handled above)
San Antonio clinics only. Only followed up patients with diagnosis confirmed by post-visit phone interview + random sample of non-depressed per phone interview
KQ1a trials
Bosmans et al 2006115

RCT, cluster randomized at clinic level

Randomized before screen, enrolled if positive screen + diagnosis confirmed by PRIME-MD

General practice
I: Screening; further evaluation by provider to determine if meet MDD criteria using PRIME-MD; providers attend 4-hr training session on screening, diagnosis, and treatment of late-life depression. Two treatment phases: every 2 wks for 2 months; monthly for 4 months.

C: Screening, further evaluation by study staff to determine if meet MDD criteria using PRIME-MD. Practitioner remained blinded to results; no training.

0% currently using antidepressants
Amsterdam, The Netherlands

Older (age 55+) general practice patients
3,937 screened
579 pos screen
339 agree to diagnostic interview
178 MDD diagnosis
145 consent to rest of study

I=70; C=75

I=18 clinics; C=16 clinics
Mean Age:
I=66.4; C= 64.7

I=66%; C=54%

Race: NR


100% not on antidepressants
History of depression:
I=85%; C=82%
Inclusion: age 55+; visit with general practitioner; GDS-15 score ≥5; PRIME-MD diagnosis of MDD;

Exclusion: current use of antidepressants; current psychosis, bipolar, or drug abuse diagnosis; severe social dysfunction; inability to communicate in Dutch; impaired cognitive functioning
GDS mode NR, PRIME-MD interview

I: GP assistant admin and score GDS, GP admin/score

C: research assistant admin and score both GDS and PRIME-MD. Result given to GP, GP administer PRIME-MD if screened positive.

Validation through GP or research assistant administration of PRIME-MD
12-monthsDepression: MDD diagnosis from PRIME-MD; depression severity from MADRS

Other: Quality Adjusted Life Years from EuroQol (EQ-5D)

Completing follow-up assessment (86% of enrolled)


Enrolled group similar on age, sex, marital status, previous history of depression, and baseline depression severity
t-test for depression severity and QALYs, Chi-sq for percent recovered

Intention to treat on all who completed all followup.
% recovered:
I=43%; C=48%, p=0.60
Mean change in MADRS from baseline:
I=−7.8; C=−7.2, p=0.70
Mean QALYs gained:
I=0.65; C=0.70, p=0.20
Analysis on I=58; C=67
Bergus et al 2005113

RCT; randomized at patient levelRandomized using random number table after positive screen, all randomized screenedTwo private rural family practice clinics with 10 physiciansI: Provider asked to review PHQ-9 results, educated about PHQ-9C: Providers educated about PHQ-9, did not receive screen resultsAccepted all positive screens, used screen cut-offs to categorize as Major and Minor depression for sub-analyses 38% on medications for depression or anxietyIowa; ruralGeneral outpatient family practice951 approached861 completed screen119 positive screen59 enrolledI=27; C=33Mean age: I=38.2; C=43.4
Female: I=62%;C=70%
Caucasian: I=92%; C=96%
Some college: I=58%, C=44%
History of depression
treatment: 60%
Currently treated for depression: 38%
PHQ: I=12.0; C=12.7
All baseline differences-NS
Inclusion: English-speaking; adult; patient at study clinicExclusion: dementiaPHQ-9Mode:
NRAdmin/score: NR
1Provider asked to review completed PHQ-9Validation:
provider review
4-, 10-, 24-weekDepression: Change in PHQ-9 score; remission per PHQ-9Other: noneAE: NRCompleted ≥1 followup (inferred).
Those unable to be contacted were dropped from the studyNRyes
t-tests; repeated measures ANOVA for continuous measuresChi-sq for % remissionPHQ9 scores carried forward for missing data10 week followupAll subjects:
Change in PHQ: I= −5.8; C= −5.8 (p=0.45)%
Remission: I=54%; C=37%
Patient with Major Depression: Change in PHQ:
I= −7.3; C= −9.1 (NS)%
Remission: I= 36%; C= 38% (NS)
24 week followup
All subjects:
Change in PHQ: I= −5.7; C= −5.0 (p=0.45)%
Remission: I=52%; C=38%
Patient with Major Depression:
Change in PHQ: I= −8.5; C= −8.2 (NS)%
Remission: I= 54%; C= 31% (NS)
Small N, probable contamination, blinding NR
Jarjoura 2004114

RCT; randomized at patient level

Randomized by permuted blocks of 40 using sealed envelopes after positive screen; all randomized screened

Indigent primary care clinic
I: Screening nurse gave results + treatment protocol to provider. Protocol includes: explore symptoms to confirm diagnosis; rule out other explanations for positive screen; educate patient & give materials; offer appointment for behavioral counseling; prescribe antidepressant if acceptable to patient; schedule return visit in 4 weeks

C: provider not informed of results; patient told may have problem with depression and that there are effective treatments
Not specified-just positive screen

0% currently being treated for depression

Internal medicine residency clinic
1,095 screened
497 positive screen
101 not currently treated for depression
61 randomized*
I=33; C=28

*24 of those not randomized were ineligible due to suicidal ideation
Mean Age: I=45; C=46

Female: I=76%; C=61%
Race: NR
SES: NR, but all were required to be below federal poverty level.

100% not current treatment for depression

BDI: 28 (I) 23 (C)

All characteristics-NS
Inclusion: age 18+; Medicaid or low-income & no private insurance; positive screen for major depression episode; not receiving treatment for depression; not seeking help for depression or other emotional problems; could read and respond to questionnaire

Exclusion: suicidal ideation

Presume mode paper-and-pencil, based on reading requirement

Nurse administered/score

Nurse gave results to clinician

Provider asked to validate screen results
6-, 12-monthDepression: BDI, PRIME-MD (1-yr only)

Other: utilization of medical & behavioral services, costs of care, SF-36

Complete ≥1 followup (inferred)


Repeated measures mixed model.

Adjusted for baseline
% showing ≥33% decline in BDI (averaging 6- and 12- mo data): 39% (I), 21% (C) (p NR)

6-mo: I-group BDI change 7.6 pts > C-group change (p NR)
12-mo: I-group BDI change 6.5 pts > C-group change (p=0.03)

“Intervention effect”=3.6 (p=0.27)

Health Care costs NS
(p-value range 0.26–0.93)
Small N, probable contamination, control group told test results, possibly capitalizing on regression to mean since I-group 6 patients higher on BDI at baseline (though NS)

Only newly-detected depression
Rost et al 2001109

Rost et al 2000190

Rost et al, 2006119

Rost et al, 2005174

RCT, cluster randomized

Randomized 12 matched primary care practices, 6 to each condition one control-group practice did not meet recruitment goals and so was replaced with another practice, participant enrolled after 2-step screen process, if patient positive on both screens
I: 2 physicians, 2 nurses, 1 administrative person trained; administrative staff recruited participant before index visit with physician; if physician confirm diagnosis, patient scheduled return visit in 1 week; re-assessed immediately before 1- week visit, patient educated about preferred treatment; ask patient to complete homework assignments; arranged further follow-up; up to 9 weekly visits with same pattern

C: Administrative staff administered screening, but physicians were not informed which patients were participating and nurses did not meet with depressed patients.
Major Depression

I: 48%; C: 40% recently (past 6 months) treated for depression
CO, MI, MN, NJ, NC, ND, OK, OR, VA, WI, urban and rural11,006 approached
9,555 completed 1st step screen
2,082 positive 1st step screen
653 positive 2nd step screen
479 patients enrolled
I=239; C=240

Patients with 6-mo followup: Previously-known cases N=243, Newly-identified cases N=189
Mean age: I=41.4;

Female: I=84%;
Caucasian: I=84%;
High school: I=79%;
C= 79%

PHQ: I=12.0; C=12.7
all baseline differences NS
Inclusion: routine- length appt with participating provider; age 18+; sufficient English literacy/cog. function to complete questionnaire; no acute life-threatening physical condition; access to telephone; positive score on both screening instruments;

Exclusion: pregnant, breastfeeding, or< 3 mos post-partum; bereavement; did not intend to receive on-going care in target clinic during next 12 months; lifetime history of mania; use of lithium; current alcohol dependence
(1) 2 WHO-CIDI items
(2) 9-item Inventory to Diagnose Depression

Mode: NR

Office staff hand-scored
Note placed in front of chart informing provider that the patient had screened positive for MDD and agreed to be in the study

Provider asked to evaluate depression diagnosis and begin study protocol of agreed with diagnosis
6-month on everyone;
12-month, 18-month, 24-month only on those with newly-identified depression episode
Depression: CES-D collected via phone interview

Other: SF-36

AE: None
Completed the follow-up interview. Report separately those who had been treated for depression in past 6 months and those who had not.


Intention-to-treat hierarchical models stratified by whether patient had recently been treated for depression, controlling for all baseline differences that were p<0.20Decrease in CES-D baseline to 6-mo:
Recently treated:
I=14.5; C= 11.0

Newly identified depression: I=21.7; C=13.5*

% Remission at 24 mos (CES-D≥15) in newly identified depression only: I=74%; C=41%
No overall results reported, only by subgroups of whether they had been in treatment for depression at baseline or not. Overall differences WERE significant, per personal communication (Rost, 2006)
Whooley et al 2000111

RCT, randomized at clinic level

Randomized before screen, all randomized screened

Primary care
I: provider notified of screening results + handout with score interpretation and general recommendations; group psycho-education classes offered to patients and family; 1-hour provider training on depression management

C: screening & no feedback, 1-hour provider training on depression management
Not specified-just positive screen

% on antidepressants in past 12 month:
I=23%, C=17%
Oakland, CA

Geriatric (65+) outpatient primary care
# approached NR
2,896 eligible for Screening
2,346 screened
2,346 randomized
331 positive screen

Mean age: I=75.7; C=75.9

Female: I=59%; C=62%
African American: I=28%; C=37%
Caucasian: I=49%;
Completed High School: I=87%; C=76%
Inclusion: age 65+, had medical appointment at one of study clinics

Exclusion: NR

Mode: NR

Administered/scored: research assistant

Results + explanatory hand-out put in chart

Validation at discretion of provider
2-yearDepression: GDS; % depressed

Other: # visits; hospitalization

Baseline GDS≥6: N=331
(I=162; C=169)
2-year GDS only avail for I=97, C=109

Those completing 2 yr followup were more likely to be divorced/separated and had fewer clinic visits in previous 12 mo.

Similar on all variables reported except % completed high school (p=0.04, fewer in C group) and income category (p=0.002, C group lower income, fewer unknown income)
Mean change based on stepwise regression-include all baseline demographic/social/heal th variables, Keeping variables with p≤0.05 in model;
Chi-sq of % improved; logistic of # improved, controlling for baseline differences;
Compare differences in mean # clinic visits and hospitalizations
GDS: both groups significant change over time; no group differences

% depressed: no group differences
I=42% (41/97); C=50%

Adjusted mean decrease in GDS score from baseline:
I=1.8; C=2.2 (p=0.41)

Health care utilization: no group differences
# clinic visits: I=1.8; C=1.6
# hospitalizations: I=1.1; C=1.0
Only analyzed those who screened positive at baseline

12% attended group session.
Wells et al 2000110

Wells et al 2004118

Sherbourne et al 2001189

Rubenstein et al, 1999240

RCT, cluster randomized

Randomized managed care organization regions; regions in matched clusters of 3, each cluster had a region assigned to one of 3 conditions; individual provider consented to participate or not.

Consecutive patients in study clinics screened, only positive screen enrolled

Primary care
I1: (QI-Meds) Screening; institutional monetary commitment; staff and clinician training (1- or 2-day workshops); clinician manuals; monthly training lectures; academic detailing as needed; numerous materials for clinicians, staff, patients; trained nurse specialists for followup assessment and on-going adherence support (medication adherence, presumably)

I2: (QI-Therapy) Screening; same QI elements as above except trained therapists to provide individual or group manualized CBT and reduced co-pay for therapy rather than nurse med specialists

C: Screening; participant told they could inform their providers of screen results; medical directors mailed the AHRQ depression practice guidelines + quick reference guides for clinicians
MDD and Dysthymia-specific screener used

% Currently treated: NR
7 geographic regions-sites chosen to oversample Mexican Americans

Outpatient primary care
44,052 approached
37,452 consent to screening
27,332 screened*
3,918 positive screen
2,176 eligible for enrollment
1,356 patients enrolled (46 clinics randomized)

I=913 (30 clinics); C=443 (16 clinics)
*most of those not screened were ineligible because they were not patients of study providers
Mean age: I=44.5; C=42.2

Female: I=71.6%;
Hispanic: I=29.1%;
Caucasian: I=57.9%;
C= 55.3%
Completed college: I=22.2%; C=15.0%

% depressed per baseline CES-D: I=75.4; C= 75.7%
Inclusion: positive screen; intended to use clinic as source of care for next 12 months

Exclusion: <18; acute medical emergency; did not speak English or Spanish; no insurance or public-pay arrangement that covered mental health care
CIDI-MDD and dysthymia sections (based on DSM)

Mode: NR

Administered/scored: study staff

Intervention clinics provided lists of study participants. Control clinics not notified.
6-month, 12-month, 18-month, 24-month, 57-month

6–24-month follow-up from mailed questionnaire with phone follow-up, 57-month followup by phone
Depression: CES-D; % probably MDD or dsythymia diagnosis per CIDI

Other: SF-12 Mental health and physical summary scales

Those completing follow-up qx
6-mo N=1,156 (85%), 12-mo N=1,126 (83%)
24-mo N=NR
57-mo N=924


I slightly older and more likely to be married than C
Patient-level intention-to-treat, using multiple imputation for missing data at item-level. Intraclass orrelations were lose to zero so analyses were not adjusted for cluster effects.

Multivariate egression models
% depression per CES-D cut-off
Baseline: I=75.4; C= 75.7
6-month: I=55.4; C=64.4**
12-month: I=54.5; C=61.4*

Still positive on CIDI:
6-month: I=39.9%; C=49.9**
12-month: I=41.6%; C=51.2**
24-month: I=est 35%; C=34%
57-month: I=37.0%; C=43.6%*

Mean mental health summary score:
Baseline: I=35.6; C=36.1
6-month: I=41.6; C=39.8**
12-month: I=40.9; C=39.3*
57-month: I=44.8; C=42.6

Mean physical summary score
Baseline: I=45.2; C=44.6
6-month: I=43.9; C=43.7
12-month: I=44.1; C=44.6

Depression diagnosis per CIDI interview (24-month)
I1=39%; I2=31%; C=34%
Fairly extensive baseline data collection procedures for all participants, plus mailed followup questionnaire every 6 months
Callahan et al 1994112


Clinical practice sessions randomized, patients scheduled during selected practice sessions and who screened positive in two separate interviews were enrolled.

Primary care
I: 2-step screen; 3 appointments with primary provider over 3 months to address symptoms of depression; letter for provider including HAM-D results, medications associated with depression, and treatment recommendations placed in chart; educational materials for patient included in chart; post-visit questionnaire

C: 2-step screen; no screening results given to provider; further appointments at discretion of provider; post-visit provider questionnaire for baseline interview visit

Other: N=97 randomly selected for extensive psychiatric interview for other purposes, not included in these results

All providers given talk on late-life treatment of depression; any patient positive for suicidal ideation at baseline interview referred for immediate psych evaluation
Not specified-just positive screen

% on antidepressants:
I=10.0%; C=13.5% (NS)
Indiana, multi-specialty ambulatory care clinic associated with urban county hospital

Target age 60+
4,413 approached
3,767 screened during primary care visit for depression, alcohol, & dementia
612 positive CES-D
515 retained
(randomly selected 97 for other purposes, not included in this study)
254 consented to 2nd interview
175 positive on HAM-D
175 enrolled
I=100; C=75
Mean age I=65.5; C=65.1

Female: I=76.0%; C=75.7%

African American:
I=50.0%; C=52.7%

Years of education:
I=8.6; C=9.1

Mean HAM-D: I=22.0; C=21.8
Mean SIP: I=33.0; C=29.9

No statistical differences
Inclusion: age 60+; regularly-scheduled primary care visit;
score ≥16 on CES-D;
score ≥15 on HAM-D

Exclusion: prisoners; patients residing in nursing home; unable to speak English; hearing impaired

In-person interview; CES-D performed during regularly scheduled visit; HAM-D at special visit within 2 wks of first visit

Research assistant
1-month, 3-month, 6-month, 9-monthDepression: HAM-D, how assessed NR

Other: SIP, how assessed NR

Completers at 6 mo.


Groups similar on sex, age, race, education, cognitive impairment, alcohol dependence, % depression diagnosis, % on medication associated with depression (e.g. narcotics), on antidepressants, mean # of diagnosis, HAM-D, SIP
Paired-comparison t-tests at 6 monthsCES-D: all patients improved over 6 months**, no group differences in amount of improvement

HAM-D: %
I=13%; C=12% (NS)

SIP: All patients improved over 6 months (p-value NR), no group differences in amount of improvement
Rubenstein et al 2007116

Subgroup analysis of CCT

Clinics assigned to treatment group, participants enrolled after screening positive for one of 5 conditions

Primary Care
I: Structured phone assessment with case manager, then referrals as needed to (1) multidisciplinary geriatric assessment clinic; (2) home-based primary care program for homebound elders; (3) primary care provider; or (4) other specific services (e.g. mental health). Also health education and promotion, written summaries of recommendations, appointments, etc.. Called again after 1 month, then quarterly for 3 years.

C: Usual Care
Not specified, just positive depression screen

Curr tx: NR
Los Angeles, CA outpatient VA clinic

Age 65+ at risk for one or more of five common geriatric conditions
(falls/balance problems, urinary incontinence, depression, memory loss, and functional impairment)
2,646 mail questionnaires
2,382 returned questionnaires
1001 met criteria as high risk
792 enrolled in study


Among those with positive screen for depression:
Mean age I=74.6

Female: I=3.7%
HS or beyond:

No statistical differences
Inclusion: age 65+, 1+ clinic visits in past 18 mos; screen positive for one of five conditions

Exclusion: live outside of 30 mile radius of clinic; already enrolled in geriatric services; living in a long-term care facility


Case manager completed more extensive work-up, referred to primary provider, specialty care as needed
1-yr, 2-yr, 3-yr (only 1-yr results reported on subgroup screening positive for depression)Depression: GDS

Other: None on subgroup screening positive for depression

Those screening positive for depression at baseline


Repeated measures ANOVAMean GDS Score:
I=8.7; C=5.0
I=8.8; C=6.1
(Time*treatment interaction p=0.05)

USPSTF=United States Preventive Services Task Force; AHRQ=Agency for Healthcare Research Quality; RCT=Randomized Control Trial; CES-D=Center for Epidemiologic Studies Depression Scale; I=Intervention; C=Control; MDD=Major Depressive Disorder; NR=Not Reported; AE=Adverse Events; PRIME-MD=Primary Care Evaluation of Mental Disorders; SES=Socio-Economic Status; GDS=Geriatric Depression Scale; GP=General Practitioner; GPSS=Geriatric Postal Screening Survey; PHQ-9=Patient Health Questionnaire; NS=Not Specified; BDI=Beck’s Depression Inventory; SF-36=Short Form; WHO-CIDI=World Health Organization Composite International Diagnostic Interview; QI=Quality Improvement; CBT=Cognitive Behavioral Therapy; HAM-D=Hamilton Rating Scale for Depression; SIP=Sickness Impact Profile; MADRS=Montgomery Asberg Depression Rating Scale

From: Appendix G, Evidence Tables

Cover of Screening for Depression in Adults and Older Adults in Primary Care
Screening for Depression in Adults and Older Adults in Primary Care: An Updated Systematic Review [Internet].
Evidence Syntheses, No. 75.
O'Connor EA, Whitlock EP, Gaynes B, et al.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.