Li 2013.

Study characteristics
Methods	Controlled before‐after study
Participants	General practitioners
Interventions	Ontario, Canada Incentive group: P4P plus existing payment Comparison group: existing payment (blend of FFS and capitation) Years 1998 to 2008, intervention happened in 2003
Outcomes	Quantity of health services provided (Difference‐in‐difference with physician‐specific fixed‐effect model) Senior flu shots rate: marginal effect 2.8 percentage point increase (P = 0.007, N = 19,866), 5.1% increase over the base compliance level Toddler immunisation rate: marginal effect 1.1 percentage point increase (P = 0.012, N = 16,826) Pap smears rate: marginal effect 4.1 percentage point increase (P = 0.005, N = 19,926), 7.0% increase over the base compliance level Mammograms rate: marginal effect 1.8 percentage point increase (P = 0.005, N = 19,888), 2.8% increase over the base compliance level Colorectal cancer screenings rate: marginal effect 8.5 percentage point increase (P = 0.007, N = 19,918), 57% increase over the base compliance level
Notes	High risk of bias
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	The allocation is not made by the investigator.
Allocation concealment (selection bias)	High risk	The allocation is not made by the investigator.
Baseline outcome measurements similar	Unclear risk	No comparison of outcome measurements was reported.
Baseline characteristics similar	High risk	At baseline, the GPs in the incentive group were younger than those in the comparison group, had fewer years of practice experience, were more likely to be female, worked more days per year, practised more intensively, had larger patient populations with slightly more females and infants, were more likely to be located in urban areas, and were more homogenous (as indicated by smaller standard deviations).
Incomplete outcome data (attrition bias) All outcomes	Low risk	No incomplete outcome data
Knowledge of the allocated interventions adequately prevented	Low risk	The outcomes are objective.
Study adequately protected against contamination	Unclear risk	No more information in text
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the methods were reported in the results.
Other bias	Low risk	No more information in text