Kutner 1999.

Methods	Cluster RCT ‐ with an intervention group (IG) and a control group (CG)
Participants	Cancer patients scheduled for a follow‐up visit in an ambulatory cancer clinic Country: USA Age: Physicians: IG: 41.7 ± 6.9, CG: 42.2 ± 6.3; Participants: IG: 51.5 ± 16.4, CG: 55.6 ± 13.3 Sex: Physicians: IG: 33% female, CG: 20% female; Participants: IG: 44% female, CG: 66% female Inclusion criteria ≥ 18 years of age had a scheduled follow‐up visit English‐speaking able and willing to consent and to read and complete the questionnaires Exclusion criteria N randomised: Physicians: n = 11, IG: n = 6, CG: n = 5; Participants: n = 282, IG: n = 149, CG: n = 133 N in analysis: baseline: Physicians: n = 11, IG: n = 6, CG: n = 5; Participants: n = 282, IG: n = 149, CG: n = 133
Interventions	Content of screen:CARE NEEDS: Needs assessment questionnaire adapted from published instruments, exploring needs in 13 domains: intensive care, financial, self‐care, future, symptom relief, treatment, emotional, spiritual, test, prevention, diagnosis, referral, and advance directives Interventionist: No interventionist for screening act, self‐completion of screening tool Intervention procedure:Solitary SI: Participants completed a pre‐visit needs assessment questionnaire; completed forms were attached to the patient charts prior to the clinic visit. Physicians were aware of this information, but were not instructed in use of the information provided Conditions for implementation: A person or system that gives/sends the pre‐visit questionnaire to participants and attaches it to patient files Comparative condition: Usual care: not further specified Length of follow‐up: No follow‐up
Outcomes	Primary outcomes items discussed at the clinical encounter visit‐specific patient satisfaction: 5‐item Medical Outcomes Study Patient Visit Rating Questionnaire visit‐specific physician satisfaction: Relation and Demand subscales + 1‐item satisfaction measure from Suchman's Physician Satisfaction Questionnaire physician participatory decision‐making style: a 3‐item scale Secondary outcomes: / Outcome time points: only 1, postvisit
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Unclear what method was used to randomise the physicians
Allocation concealment (selection bias)	Unclear risk	Unclear which method was used to conceal the allocation to conditions
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding of physicians or participants
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome data were collected with postvisit questionnaires in both conditions No extra person for outcome assessment aware of condition allocation
Incomplete outcome data (attrition bias) All outcomes	Low risk	Only 1 outcome time point, so no potential for missing data due to loss in follow‐up. No indication for other missing data
Selective reporting (reporting bias)	Low risk	Incomplete reporting of outcomes (only significant subscales reported for patient satisfaction, without a measure of the spread of the data)
Other bias	High risk	Adjusted P values reported everywhere to adjust for clustering, but no information on how this adjustment was done It is clear that there is a huge difference in baseline characteristics, which is the result of clustering at the physician level, but this becomes non‐significant when "clustering is taken into account". Nevertheless, this remains problematic