*Number of clinicians who consented to study participation. Column counts sum to greater than the total counts for clinician characteristics because after the intervention was implemented in clinic 1, one clinician from clinic 1 started to also practise in clinic 4 due to operational needs. All of the clinic 4 visits for this clinician were conservatively treated as being in the control period 4 until the intervention was implemented in clinic 4. Data on clinicians are from the postintervention survey (completed by 44 of the 45 study clinicians).

†Clinicians’ self-report that more than 30% of their patients had each type of health insurance coverage. One survey participant did not complete this item.

‡The number of applicable visits across the control, intervention and follow-up periods using lists of the International Classification of Diseases, Ninth Revision (ICD-9) codes previously developed by Premera Blue Cross and the Washington State Choosing Wisely Task Force to measure rates of potentially low-value services in visits for which each respective Choosing Wisely recommendation could apply. All lists of ICD-9 codes are shown in online supplementary appendix table 1.

§Characteristics of patients in visits with an ICD-9 code from the lists of ICD-9 codes previously developed by Premera Blue Cross and the Washington State Choosing Wisely Task Force to measure rates of potentially low-value services in visits for which each respective Choosing Wisely recommendation could apply.

¶Number and per cent of patients in each clinic who were female. Data on gender were missing for 35 patients.

**Charlson scores based on ICD-9 codes used at each study visit. The weighted Charlson sum ranges from 0 to 8. The index is 0 for observations whose sum is 0, 1 for observations whose sum is 1, and 2 for observations whose sum is between 2 and 8.