Table 7 Application of the INTENTS framework to the case of the Quality and Outcomes Framework (QOF) in the UK

1

What are the expected outcomes of the intervention?

Increased primary care professional effort and therefore better performance on specified indicators of quality of care for patients with chronic conditions, resulting in improved clinical outcomes for these targeted patients.

2

At what level can spillover effects take place?

a) Who is targeted by the intervention?

All UK primary care professionals.

b) Who is expected to change behaviour as a result of the intervention?

All UK primary care professionals.

c) Whose behaviour/outcomes may change as a result of the intervention?

All UK primary care professionals and their registered patients.

3

Which spillover effects could the intervention generate?

a) Within-unit spillover effects (non-intended effect on a targeted unit)

Changes to primary care professional effort and performance on not-specified aspects of quality of care for targeted patients.

For example, the QOF directly incentivised the recording of certain risk factors (including smoking status) for targeted patients. Targeted patients were found to have experienced positive spillover effects as primary care professionals also increased their recording of other clinically effective risk factors (BMI and alcohol consumption) for which they were not financially rewarded for these patients [7].

b) Between-units spillover effects (intended effect on a non-targeted unit)

Changes to primary care professional effort and performance on specified indicators of quality of care for non-targeted patients.

For example, the QOF directly incentivised the recording of certain risk factors (including smoking status) for targeted patients. Untargeted patients (those without the specific diagnosis codes targeted) were found to have experienced positive spillover effects as general practitioners also increased their recording of specified risk factors for patients not targeted by the policy [7, 54].

c) Diagonal spillover effects (non-intended effect on a non-targeted unit)

Changes to primary care professional effort and performance on not-specified aspects of quality of care for non-targeted patients.

For example, diagonal spillovers could have occurred if untargeted patients (those without the stated diagnosis codes) were found to have experienced changes in not-specified aspects of care quality (such as the recording of clinically effective but unincentivised risk factors, including BMI and alcohol consumption).

4

What is the nature of the potential spillover effects identified in step #3?

a) Is the spillover effect really different from the intended outcome?

Yes. The QOF considers specific indicators of care for targeted patient groups. Any changes in primary care professional effort and performance either on these indicators but for non-targeted patients, or on other aspects of care quality for targeted patients, represent separate effects to those intended by the policy.

b) Is the spillover relevant and related to the goals of the intervention?

Yes. Wider changes in the quality of primary care services provided to patients are relevant and related to the goals of the QOF [51, 55].

c) Is the spillover effect consistent with the time frame of the intervention?

Yes. The spillover effects were examined and detected over the same period as the direct effects of the policy on the recording of incentivised clinical indicators for targeted patients. Trends in recruitment and retention in primary care could, at the margin, be influenced by the effects of the Quality and Outcomes Framework on the attractiveness of working in primary care, but this would be difficult to isolate from more proximal influences [56,57,58,59].

d) Is there a credible mechanism for the spillover?

The detected positive spillover effects from the QOF are consistent with the policy inducing general practices to make investments in quality that extended beyond the scheme. Complementarity in the production of healthcare appears to be a credible mechanism in this instance.

The potential for negative spillover effects due to multi-tasking concerns around effort diversion (away from untargeted patients and aspects of quality of care not specified by the incentive scheme) was hypothesised. Whilst a credible mechanism, the evidence to date does not suggest that this effect dominated in practice [60]