Cooper 1975.

Methods	Controlled clinical trial with follow‐up at 12 months. Collaborating partners Lead agency: Primary care Strategic involvement (policy making and service planning): Primary health care. Commissioning (implementing strategy taking account of resources available): Primary health care, social services. Operational (providing services directly): Primary health care, secondary health care, social services. Set in United Kingdom.
Participants	189 primary care patients living in a metropolitan area with chronic neurotic illness were followed up at 12 months. Intervention group 92 patients (86.8% of patients enrolled), males 26.1%, mean age 42.1 years. Control group 97 patients (84.3% of patients enrolled), males 22.7%, mean age 45.5 years.
Interventions	Attachment of a social worker to a primary care practice, and involvement of research psychiatrists. The GPs, health visitors, social worker and research psychiatrists attended fortnightly meetings to discuss new referrals and progress of cases. Once experiment was established this evaluation was set up to assess the therapeutic value of the service. Patients in intervention group had usual care plus one or all of the following: 1.      Recommendations to GP 2.      Referral to local psychiatric or social services 3.      Social support within practice 4.      Consultation with research team psychiatrist
Outcomes	Change in psychiatric rating (scale now known as GHQ 30) Change in social adjustment score (author scale)
Notes	No power calculation was reported. Additional resources used included a social worker allocated to GP practice. Involvement of two research team psychiatrists. Overall risk of bias was high.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	No randomisation. Control patients selected from other practices.
Blinding (performance bias and detection bias) All outcomes	High risk	It does not appear that there was any blinding. Although assessors who treated patients did not assess those patients at follow‐up there was no indication that psychiatrist was blind to study group.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Full outcome data.
Selective reporting (reporting bias)	Unclear risk	This study was conducted in the early 1970s and the protocol is not available
Other bias	High risk	Possible selection bias. Experimental cases were put forward in the hope that expert help might be given while the control patients were selected purely for research purposes.
Randomisation adequately described/protected?	High risk	Non‐randomised controlled study, not possible to randomise.
Protection against contamination?	Low risk	Contamination unlikely. Patients for the control group were drawn from separate practices without access to the experimental facility.
Follow‐up rate adequate?	Low risk	Follow‐up rate >80% and balanced across both arms. Whilst the authors state that 86.8% of intervention and 84.3% of controls were successfully followed up, it is impossible to confirm this from the way the data are presented. It is equally impossible to say whether the authors conducted an ITT analysis.
Reliable primary outcome measure?	Low risk	Standardised psychiatric interview which had been tested for inter‐rater reliability. The psychiatric outcomes are measured using what has become a validated self‐administered checklist: the General Health Questionnaire (GHQ‐30). Social outcomes appear to be measured by a scale developed by one of the authors.
Groups measured at baseline?	Low risk	Only sociodemographic baseline data were assessed for statistical differences between groups. Differences noted though none reached significance (but relatively small sample sizes).  Relative excess of 60‐65 year olds and social classes I and II in controls and small excess of retired persons and social classes IV and V in experimental group.  Groups appeared well matched for psychiatric ratings.
Appropriate choice of controls (CBA studies only)?	High risk	Control patients identified up to 18 months ahead of the intervention group. Cannot be certain that treatment/prognosis did not change during period though authors say no reason to believe that this was the case.
Contemporaneous data collection (CBA studies only)?	Low risk	Data collected at one year in both groups.
IS THE STUDY AT LOW RISK OF BIAS?	High risk	OVERALL RISK OF BIAS WAS HIGH