Marshall 2007

Methods	Controlled ITS with monthly data points
Participants	Seniors ≥ 65 years (provincial drug plan beneficiaries) in Quebec (control), Ontario and British Columbia (interventions)
Interventions	Effects of different reimbursement policies on the use of COX‐2 inhibitors, NSAIDS, and gastrointestinal protective agents
Outcomes	Drug use
Notes	Quebec: no restrictions, Ontario: submit indication with prescription, British Columbia: request PA
Risk of bias
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias) All outcomes	Low risk	Our study spanned a period of 5 years. This provides a sufficient number of observations in both the pre‐coxib and post‐coxib time periods to capture meaningful trends and avoid spurious spikes in the data.
Selective reporting (reporting bias)	Low risk	All outcomes reported
Other bias	Low risk	No other bias detected.
Knowledge of allocated interventions adequately prevented during the study	Low risk	Outcomes objective: how do these different prescribing restrictions affect the uptake and expenditures on COX‐2 inhibitors? Is Ontario’s Limited Use policy an effective deterrent to the prescribing of COX‐2 inhibitors? To what extent are increased coxib expenditures in jurisdictions which do not restrict their use (Quebec) offset by reductions in gastro‐protective agent (GPA) expenditure?
Intervention independent of other changes	Low risk	"The Therapeutics Letter was found to have shifted prescribing an average of 30% from baseline according to an aggregate analysis of 12 letters using randomised controls [24]. Four letters were published during the period between 1999 and 2004 [25–28], the first of which pointed to the lack of published clinical data on coxib effectiveness and the other three which focused on serious adverse events associated with COX‐2 inhibitors (fewer gastrointestinal events, but more myocardial infarctions)." Nonetheless, the strong temporal correlation between the introduction of administrative restrictions for COX‐2 inhibitors, and the lack of other trends to explain the change in NSAID utilisation, support our findings. In all three jurisdictions, the policy environment for pharmaceuticals as stable over the observation period. There were no significant changes in the insurance environment that would be expected to have had an effect on uptake of NSAIDs.
Intervention unlikely to affect data collection?	Low risk	Aggregate data on anti‐inflammatory and GPA use between April 1997 and December 2002 were obtained from each of the three provincial drug plans (Regie de l’assurance maladie du Quebec (RAMQ), ODB, and BC PharmaCare). These data reported the number of prescriptions filled and the total reimbursement by the drug plan on a monthly basis.