Effect of primary health care reforms in the province of Newfoundland and Labrador: Interrupted time-series analysis.

OBJECTIVE

To examine the effects of primary health care (PHC) reforms in the Canadian province of Newfoundland and Labrador on ambulatory care-sensitive (ACS) hospitalization rates and mortality.

DESIGN

Interrupted time-series analysis of administrative data.

SETTING

All communities in the province of Newfoundland and Labrador were divided into 3 groups: rural reform (n = 69 143), rural nonreform (n = 228 914), and urban nonreform (n = 197 012). No urban communities introduced PHC reforms.

PARTICIPANTS

All residents of the province who held a valid health card and did not change their address during the 2001-2009 study period were included. Individuals were assigned to 1 of the 3 study groups based on community of residence.

MAIN OUTCOME MEASURES

Hospitalization rates for ACS conditions, hospitalization rates for control conditions, and ACS-related mortality were compared using interrupted time-series models.

RESULTS

Results are reported as rate ratio or odds ratio (OR) (95% CI). In rural reform and rural nonreform communities, there was a decreasing trend in ACS hospitalization rates that preceded reforms (rate ratio of 0.97 [0.94-1.00]) and rate ratio of 0.98 [0.96-1.00], respectively) but no change following reforms. There were no significant changes in the urban group. In all 3 groups, there was a significant increasing trend in ACS-related mortality before reforms (OR of 1.09 [1.02-1.15], OR of 1.10 [1.06-1.13], and OR of 1.09 [1.05-1.14] for rural reform, rural nonreform, and urban communities, respectively), which was reversed after the introduction of reforms ( < .01).

CONCLUSION

Primary health care reforms in Newfoundland and Labrador had no observed effect on ACS hospitalization rates, but a potential effect might have been masked by a decreasing trend that preceded the introduction of reforms. The increase in mortality rates that was reversed after the introduction of reforms cannot be attributed to the reforms because it occurred in all studied populations including those that did not introduce reforms.