An interrupted time-series analysis of the effects of withdrawal of financial incentives on diagnosis of atrial fibrillation as resolved. Does withdrawal of an incentive reverse its unintended effects?

Background: The UK introduced financial incentives for management of atrial fibrillation (AF) in 2006, after which there was an increase in the proportion of patients with AF diagnosed as resolved. Removal of incentives in Scotland provides a natural experiment to investigate the effects of withdrawal of an incentive on diagnosis of resolved AF.

Aim: To investigate the effects of introduction and withdrawal of financial incentives on the diagnosis of resolved AF.

Design & Setting: Cohort study in a large database of UK primary care records, before and after introduction of incentives in April 2006 in Scotland, England, and Northern Ireland, and their withdrawal in April 2016 in Scotland.

Method: Interrupted time-series analysis of monthly rates of resolved AF from January 2000-September 2019.

Results: A total of 251 526 adult patients with AF were included, of whom 14 674 were diagnosed as resolved AF. In April 2006 there were similar shift-changes in rates of resolved AF per 1000 in England 1.55 (95% confidence interval [CI] = 1.11 to 2.00) and Northern Ireland 1.54 (95% CI = 0.91 to 2.18), and a smaller increase in Scotland 0.79 (95% CI = 0.04 to 1.53). There were modest downward post-introduction trends in all countries. After Scotland's withdrawal of the incentive in April 2016 there was a small, statistically non-significant, downward shift in rate of resolved AF per 1000 (0.39 [95% CI = -3.21 to 2.42]) and no change in post-removal trend.

Conclusion: Introduction of a financial incentive coincided with an increase in resolved AF but no evidence was found that its withdrawal led to a reduction.