Mathias 1994.

Study characteristics
Methods	Randomised trial, USA
Participants	75 physicians and 573 primary care patients with unrecognised and untreated anxiety at TakeCare, a mixed‐ model health maintenance organisation (HMO) in central Colorado. Mean age of participants was 41.5 years for the demonstration group and 43.6 for the control and were mainly female (61.1% for the demonstration group and 54.6% for the control)
Interventions	Participating physicians were randomised to either the demonstration or the control arm, and patients were assigned to a study arm based on the randomisation of their physicians.The patients were followed for change in outcome measures during the five‐month study period. The physician intervention was to providing an educational demonstration of anxiety in the primary care setting and to provide a reporting system for summarising the anxiety symptom levels and functioning status of the patients enrolled in the study.   Intervention features Multiple complex feedback (multiple PROMs at multiple times)  PROM(s) used as intervention: Global Anxiety Score (GAS), Global Severity Index (GSI), Highest Anxiety Subscale Score (HASS) Constructs measured: Symptoms, Functioning Instrument categories/domains: Generic, Domain/Disease specific (mental health)   Administration features Where PROMs administered: Clinical setting (e.g. waiting room, office, etc) How administered: Self‐administered Format of PROMs questionnaire(s): Electronic and Paper   Feedback features Format of PROMs feedback: Unclear How often information fed back: 3 times Who information fed back to: Clinicians Information fed back: Scores
Outcomes	Main outcomes: anxiety symptoms (GAS, HASS) Other outcomes: psychological distress (the Global Severity Index (GSI), functioning and well‐being (SF‐36), global improvement (perceived changes in anxiety level, functioning and well‐being, and perceived changes in communication with their physicians since the baseline survey). The study period was not reported. Conflicts of interest were not reported.
Notes	The study was supported by Upjohn Company, Kalamazoo, MI; TakeCare,CO.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Physicians were randomised by call group but no other information available about how that was done
Allocation concealment (selection bias)	High risk	Due to cluster‐randomised design not possible to conceal allocation from clinicians.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to nature of intervention not possible to blind patients and personnel.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Due to nature of the intervention blinding of outcomes not possible: PROM used for feedback also used to assess outcome, patients were aware they received the intervention.
Baseline outcome measurements similar	Low risk	None apparent
Baseline characteristics similar	Low risk	Table of baseline characteristics provided and no significant differences identified.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Baseline characteristics of those lost to follow‐up were compared with those who completed the study
Was study protected against contamination	Low risk	None apparent
Selective reporting (reporting bias)	Unclear risk	Unclear whether selective reporting took place