We used the following search terms in our search of Medline, Embase, Cinahl; PsycLIT, and the Cochrane Controlled Trials Register:
"Health-Status-Indicators"; "Outcome-and-Process-Assessment-(Health-Care)"/ all subheadings; "Outcome-Assessment-(Health-Care)"/ all subheadings; (outcome measure*) in ti,ab; (health outcome*) in ti,ab; (quality of life) in ti,ab; measure* in ti,ab; assess* in ti,ab; (score* or scoring) in ti,ab; index in ti,ab; "outcomes-research"/ all subheadings; health outcome* in ti,ab; scale* in ti,ab; monitor* in ti,ab; assess* in ti,ab; outcome* in ti,ab; explode "Treatment-Outcomes"; explode "Psychological-Assessment"; "Quality-of-Life"; (outcome* or process*) near3 assessment*; health status indicator*; health status; health outcome* in ti,ab; quality of life in ti,ab
Table1 Randomised controlled studies evaluating use of routine outcome measures for psychiatric disorders in primary care and general hospital settings
| Author and year | Design | Population, setting and sample size | Routine outcome measure | Intervention and control conditions | Length of follow up and outcomes | Results |
| Johnstone and Goldberg 19768 | Patients randomised using alternate odd and even allocation* | Sequential attenders to a single UK doctor (n=1093); those with psychiatric morbidity (scores on general health questionnaire >5) that had not been hitherto recognised by the doctor (hidden psychiatric morbidity) were followed up | General health questionnaire (version not given) | Intervention: questionnaire administered and clinician asked about likelihood of psychiatric morbidity. Results then fed back to clinician. Those with unrecognised depression and high scores at initial interview (hidden psychiatric morbidity) were followed up (n=60). Control: questionnaire administered and clinician asked about the likelihood of psychiatric morbidity. Questionnaire folded and placed in the patient note envelope. Those with unrecognised depression and high scores at initial interview (hidden psychiatric morbidity) were followed up (n=59) | For those with hidden psychiatric morbidity the following were studied: diagnosis and severity of depression during 12 months follow up (including scores on questionnaire); length of depressive episodes; pattern of consultation over 12 months | Feedback increased the rate of detection of hidden psychiatric morbidity by 11%. Improved outcome at 12 months only for those with >severe= disorders. Feedback reduced length of illness (2.8 v 5.3 months). Feedback facilitated a more psychological, rather than somatic, pattern of consulting |
| Moore et al 19789 | Individual patients randomised | Attenders at general practice with self rated depression scores >50 (n=96) | Zung self rating depression scale18 | Intervention: questionnaire administered and score fed back (>mildly= or >severely depressed=) Control: questionnaire administered but no feedback to clinician | Notation of depression following index visit | Selective feedback increased recognition of depression for high risk patients (22% v 56%) |
| Linn and Yager 198014 | Individual patients randomised | New referrals to US medical outpatients (n=150); mean age 56 | Zung self rating depression scale18 | Intervention: five different interventions, which varied time of feedback of questionnaire, and the use of an interview to sensitise the clinician to the presence of depression Control: administration of questionnaire but no feedback to clinician | For all patients at initial interview the following were studied: whether depression noted in charts and initiation of any treatment for depression | Depression was generally under-recognised. Screening and feedback of questionnaire increased the frequency of notation of depression (8% v 25%) Increased notation of depression occurred irrespective of time of feedback (before or after consultation). Sensitisation to depression had no effect. Screening had a much smaller effect on initiation of treatment for depression |
| Hoeper et al 198410 | Individual patients randomised | Adult primary care patients in USA (n=2309) treated by 14 physicians | General health questionnaire (version 28)4 | Intervention: questionnaire administered by researcher and scores fed back to clinician, with information that a score >5 indicated mental illness Control: questionnaire administered but no feedback to clinician | Clinician diagnoses of mental illness at reference visit (information elicited as part of study) | No difference in rate of detection of mental disorders (16.0% (intervention) v 16.8% (control)). No difference in rate of detection among those with high scores (30% v 29%) |
| German et al 198715 | Individual patients randomised | Adult and elderly general medical outpatients in USA (n=1242); separate interventions for high (n=488) and low (n=754) scorers on general health questionnaire | General health questionnaire (version 28)4; administered by a research assistant | Intervention: questionnaire administered preconsultation and results fed back to clinician, together with an indication that score was high and suggested >psychiatric diagnosis= (n=165) Control: questionnaire administered but not fed back to clinician (n=323) | For all patients at initial interview the following were studied: detection of depression by clinicians; treatment initiated for depression; scores on questionnaire at 6 months | No difference in detection rate among under 65s (57% (intervention) v 58%). Greater detection of depression in over 65s with feedback (63% v 43%). No differences in management of depression in under 65s (46% v 46%), but greater proportion of over 65s receiving intervention after feedback (42% v 32%). Scores at 6 months not reported |
| Magruder-Habib et al 199016 | Individual patients randomised | Male adult veterans (mean age 60) attending a US general internal medicine outpatients clinic, with Zung self rated depression scores >50 (n=100) | Zung self rating for depression scale18 | Intervention: questionnaire administered and fed back to physicians at first clinic assessment visit–placed at front of clinic notes (n=48) Control: questionnaire administered but not fed back to clinician (n=52) | Recognition of depression; initiation of management of depression; scores on questionnaire at 3, 6, 9, and 12 months | Greater recognition of depression in feedback group (56% v 35% at 12 months). More frequent intervention in feedback group (56% v 42% at 12 months). Feedback facilitated recognition for those with a high somatic score on the subscale |
| Dowrick and Buchan 199512 | ndividual patients randomised | Consecutive attenders to general practice in Liverpool (n=116), with depression score >14 on the Beck depression inventory | Beck depression Inventory17 | Intervention: questionnaire administered preconsultation and depression scores fed back to clinician (n=52) Control: questionnaire administered but not fed back to clinician (n=64) | Rates of recognition of depression at 6 and 12 months; rates of intervention for depression (antidepressants; outside mental health referral or discussion of depression with the subject); scores on questionnaire at 6 and 12 months | Disclosure had no effect on the rate of clinician diagnosis of depression at 6 months (relative risk 0.82, 95% confidence interval 0.32 to 2.07) and at 12 months (1.71, 0.93 to 3.14). Disclosure had no significant difference on rates of intervention at 6 months (1.43, 0.75 to 2.69) and at 12 months (1.22, 0.69 to 2.11). Disclosure had no discernible effect on scores: median group difference in scores at 6 months=!1 (95% confidence interval !4 to 3) and at 12 months=0 (!3 to 11) |
| Mazonson 199611 | RCT Primary care group practices randomised | Primary care patients with hitherto unrecognised anxiety | Symptom check list-9019 (anxiety subscales only) and short form 3620 | Intervention: clinicians (n=40) given an educational package that included teaching sessions on the importance and causes of anxiety problems; they received structured feedback of anxiety scores (system check list) and functional status (SF-36) scores from 357 patients. Feedback was given at consultation at two further points in the follow up (11 weeks and 5 months) Control: clinicians (n=35) received no feedback from 216 patients who had completed both questionnaires | Recognition and treatment for anxiety problems; changes in anxiety scores at 3 and 5 months; changes in SF-36 scores at 3 and 5 months; self reported global improvement in anxiety and functional status | Increased recognition and treatment for anxiety symptoms (35.6% v 20.8%; P<0.001). Increased referral to mental health sector (9.5% v 3.2%; P<0.001), but no difference in the prescription of pscyhotropics. No differences in change for anxiety scores (P=0.89). No differences in change for SF-36 (total and mental health scores). Self reported global anxiety and functional status both improved in intervention group (46.3% v 37.0%) and reported improvement for anxiety |
| Lewis 199613 | RCT Individual patients randomised | Attenders at a single general practice with general health questionnaire (version 12) scores >2 (n=681) | General health questionnaire (version 12) and computerised assessment of psychiatric symptomatology | Intervention 1: questionnaire administered and placed in notes with no interpretation or instruction on the presence of mental disorder (n=227 patients) Intervention 2: patient asked to complete a computerised assessment; results were fed back to the clinician (n=227) Control: no feedback given (n=227) | Consultation rates and clinician attribution of encounters as due to psychological or physical problems; prescription of a psychotropic drug; rates of outside mental health referrals to outside agencies; scores on questionnaire at 6 weeks and 3 and 6 months | No differences in consultation rates, but more identified as >psychological= for general health questionnaire group (P=0.09). No differences in the rate of prescriptions for psychotropics. No differences in the rate of referral to outside agencies. Moderate improvement (5%, B3 to 14%) in general health questionnaire scores at 6 weeks for computerised feedback. No differences between groups over longer term |
*Pseudorandomised study.