A controlled before-after study of a local quality improvement programme for hypertension and AF: What benefits to patient outcomes and how long do they last?

Problem

Evidence in favour of using financial rewards for quality improvement in primary care is generally inconsistent, often short-lived, and arose from data with confounded control groups. The biggest and best-known reward scheme is QOF, which has proved unpopular with surgeries, and has some negative effects on care. East Lancashire CCG has run a local quality framework for general practice since 2016/7 designed to improve standards and diagnosis in AF and hypertension, for a fixed amount per head of population, rather than the sliding remuneration scale used by QOF. Participating surgeries appointed surgery leads for AF and hypertension, produced their own protocols, and reflected with peer groups on protocols and data at quarterly workshops. At SAPC ASM, we have previously demonstrated that this improves diagnosis rates of hypertension and management of AF, after 1 and 2 years, although benefits in AF diagnosis seen in year 1 were short lived. With latest data available for year 3, we aimed to analyse long-term effects of this continuing intervention.

Approach

This study used published QOF data to compare those surgeries taking part in the intervention (n=57) to controls in the rest of England (n=7243). It adopted a controlled before-after approach, using the primary outcome measure for AF and hypertension of change from baseline in mean recorded prevalence per surgery. Secondary outcome measures looked at indicators of quality of management. Time-series analyses at CCG and surgery level were used to determine whether the effects of the intervention appeared to be uniform with time or peaked then waned.

Findings

Welch’s t test showed a significantly higher mean increase in recorded hypertension prevalence over 36 months for the intervention group (0.82%) versus controls (0.26%), p=0.006, although the mean increase in AF prevalence (0.33%) compared to controls (0.29%) was non-significant, p=0.305. Time series analysis showed that there was a significant increase in recorded hypertension and AF prevalence in the first year (p=0.033 and 0.006 respectively) and in hypertension prevalence in the second year (p=0.011) but not in either in the third year. Opportunistic BP checks were higher in the second year (p<0.001) and AF management in the first year (p=0.001) but not in any other year.

Consequences

Earlier data demonstrated that this funded quality framework designed to meet local needs, could engage local primary care providers and improve patient outcome measures, without the need for target incentivisation. Potentially, it offers an alternative method for funding quality improvement within existing services. Similar to findings of some earlier quality-improvement schemes that demonstrated only short-lived benefits, we found that acceleration in already improving standards may be limited to one or two years, although the overall benefit gained during this time is maintained for longer. This has implications for the optimum length for improvement schemes.