more_vert
FINANCIAL incentives (“pay-for-performance”) are increasingly being used as a carrot to drive clinician behaviour.
It is certainly true that we often don’t do the things we ought to (such as prescribe warfarin for patients with atrial fibrillation and use beta-blockers in heart failure), and do things we ought not (such as over investigate low-risk patients with atypical chest pain, over-repeat cholesterol tests, and overuse antibiotics in acute respiratory infections).
As well, our fee-for-service system is not well designed to finance complex care for chronic conditions so anything that boosts appropriate care is worth exploring.
However, there is growing concern that financial incentives don’t always work. Experience from a number of sources suggests that there are a number of ways in which some incentives don’t just fail, but actually cause more harm than good.
At a 2-day conference last year we got together with a group of colleagues to nut out the problem, and come up with solutions. Our solution was a checklist, published last week in the BMJ — nine questions that need to be answered satisfactorily before financial incentives are set in place:
1. Do we know that the proposed clinical activity will improve patient care? The UK Quality and Outcomes Framework (QOF), designed to set targets to get GPs to lift their game, chose HbA1c outcome for patients with diabetes as one target at a time when several large trials were underway. Those trials reported that these targets may be totally inappropriate (one trial stopped early because of greater mortality among patients with lower HbA1c levels).
2. Is the financial incentive unnecessary because the change in behaviour is happening anyway? In the above UK initiative, for example, there were several examples of trends in improved care that preceded the QOF.
3. Can we measure the desired clinical behaviour? This is usually the case.
4. Are the necessary systems in place to enable the new clinical behaviour to happen? This is usually not addressed.
5. Do we know that financial incentives are better than alternative interventions? There are a huge variety of alternative strategies and it is important to check that none are more effective than money.
6. Are there unintended harmful consequences? These can include “gaming” the incentive (distorting its intent for pecuniary advantage — commonly found in hospital incentive schemes). Dropping other important activities in favour of those incentivised is also a problem (eg, the QOF resulted in UK GPs reducing teaching of medical students, and also may have harmed the GP‒patient relationship by chasing targets contrary to the patient’s best interests). More subtle is the demotivation that can occur if the doctor’s focus is switched to financial gain, rather than the altruism, professional satisfaction or commitment to patients.
If all of these questions can be answered satisfactorily, then there are operational questions:
7. Are the necessary external conditions in place? For example, when the Medicare-funded immunisation co-payment to GPs was initially introduced, there was a cold-chain failure in rural areas that immediately became apparent.
8. How much should the incentive payment be, and for how long? What happens when the incentive stops? In the QOF, some of the targets reverted when payment stopped. Or should it continue indefinitely?
9. How should the incentive payment be delivered? Who is it paid to (the practice for reinvestment, the individual GPs as income supplement, or does the corporate owner take it as extra profit)?
Financial incentives can and have contributed to driving changes for the better in clinical behaviour.
However, they are not without risk. Our checklist may help policymakers avoid unintended consequences that can leave the system worse off.
Professor Chris Del Mar is director of the Centre for Research in Evidence-Based Practice, Faculty of Health Sciences and Medicine, Bond University, Gold Coast, Queensland. Professor Mark Harris is executive director of the Centre for Primary Health Care and Equity at the University of NSW.
Posted 20 August 2012
Now retired, and in my seventies, I find myself on the opposite ‘side of the sheets’. I simply hate it — being intimidated by those who are proponents of ‘rational’ medicine, about the slightest deviation from pre-determined ‘normals’ no matter how accurate or non-accurate, having medicines prescribed for a,b,c,….ending up as a pill popper! Does no one give a red cent for the physiology of ageing??
The greatest challenge in implementing incentive systems is ensuring that you have reliable near real time access to the relevant data. As Roland pointed out in his comments, one thing lacking in Australia is the quality of record keeping. Since our fee for service model pays for time based consultations and procedures rather than for complexity of care, we have not developed the discipline we see in other countries which tie reimbursement to diagnosis. However, as we recently saw in Canberra, and is prevalent throughout the acute care sector, gaming the coding system is an unintended consequence of tying remuneration to diagnostic codes directly.
Maybe a system more like what we see in the business world of a base salary with key deliverables against it, and then incentive payments for improved productivity may make sense. That way we could link things like chronic disease management to the base pay, it is just good medicine, and the incentive to things we want to change in the short to medium term.
At the end of the day we need to ensure clinicians are properly reimbursed, and the system is sustainable.
I too thought this was a good article and the list of nine criteria seems appropriate.
I also worked in general practice in the UK at the time the Quality Outcomes Framework was introduced.
The QOF has had some very good aspects and some very bad ones. On the good side it makes how general practitioners assess and manage common important conditions more rational, and correct. I could see my treatment regimes improving (this does assume that the targets chosen are correct but this would be the case if there were no QOF as well). So I felt that it made me a better GP, or at least that I was practising better medicine.
However, what I strongly did not like was the financial overtones. I had always naively imagined that one of the favorable aspects of working for the NHS would be to be able to ignore financial considerations and get on with the medicine. What I saw with the new contract and the QOF was doctors spending substantial amounts of time counting diagnoses and assessing performance against targets to improve incentive payments.
I like performance indicators but financially rewarding targets needs to be chosen with great caution.
Coding of diagnoses raised by Dr Owen above could be very much better as he mentions but this requires that our informatics experts really create a system that works; coding in the UK with the EMIS system that most practices were using also seemed far from optimal.
A very good article. I only hope that Canberra & the powers that be take this into account to provide sustained improvements to our (great) healthcare system. Being UK trained I have witnessed the effect & implementation of QoF. I do genuinely think it did help a lot of medical targets to be hit. Some of the more “relaxed” GPs who would usually let abberant readings go would chose to treat more aggressively to hit the targets. In fact, nurse run clinics would often drive the treatment as they would follow the guidelines to the letter far better than us GPs.
One of the big problems I find over here in Australia (having worked her 6yrs as a GP now) is the medical records & diagnosis recording is suboptimal. Qof drove the improvement of recording diagnoses to then implement the targets. Good documentation of diagnosis is essential to recording & auditing healthcare. THIS was one of the biggest benefits.
Finally, let’s not go into changing the fee for service debate here… but I definitely don’t want a NHS style capitation system with no Dr input, but do think we can improve things here.