Research analysis: Payment by Results and hospital admissions
The implications of pay-for-performance models of addiction treatment funding for policy and practice. This original research article (1) by Dr Thomas Mason and colleagues was published in Addiction in 2021. Commentary by Rob Calder.
Click here to listen to an interview with first author, Dr Mason.
This is a very timely article, although, it is also 10 years too late and simultaneously forward looking.
It is timely because it comes as the Lansley reforms are being reversed and quietly ushered out of the building. It is out of date because it highlights problems with a payment system from 10 years ago. It is forward looking because policies have a tendency to come round again.
- Areas that operated pay-for-performance models had higher levels of drug-related emergency hospital admissions than those that didn’t
- Previous research has demonstrated that pay-for-performance models have not been effective in improving performance; this study indicates that those models can be detrimental for people who use drugs
- Across the healthcare system, pay-for-performance models do not necessarily save money
So, let’s go for timeless. This timeless piece of research looks at hospital admissions in UK local authorities – eight that operated a pay-for-performance model compared with 141 that didn’t. Those that operated a pay-for-performance model had higher levels of drug-related emergency hospital admissions than those that didn’t.
The study estimated that there were 3,352 more emergency hospital admissions in authorities that piloted pay-for-performance than in those that did not. The cost of these extra admissions was estimated to be around £2.1m, enough to fund a reasonably sized drug treatment centre for a year. This irony is not explored in the discussion (in fairness, irony and academic articles are a bit like oil and water).
The paper suggests that pay-for-performance funding discouraged admissions to addiction treatment, that this in turn made treatment harder to access and therefore there was an increase in emergency hospital admissions.(2,3) There is good evidence that pay for performance models reduce admissions to treatment; however, as a line of causality (from funding model to A&E), it’s slightly too long for comfort.
“Real-world” policy effectiveness
The study analysed a huge dataset, nearly 4 million observations relating to over half a million people. It’s an impressive piece of work; and, we need more research about the ‘real-world’ impact of policy, particularly research that relates to funding.
But then we lurch onwards towards the reason why there isn’t as much real-world policy research as we would like. It’s all just so complex. There are so many outcomes, so many unpredictable people, and so many potential variables and confounders.
Even with such a big dataset, there were just 8 authorities in the intervention group, which is not a big enough sample of local authorities to attribute differences to the policy alone.
It is possible, for example, that local authorities introducing pay-for-performance were authorities that were keen to save money (in fairness not so rare between 2012 and 2014) and will have cut funding to other health services used by people who use drugs.
To implement a new funding model, most authorities will have had to retender their treatment services. The disruption of retendering and the associated TUPE has been highlighted by, among others, the ACMD.(4,5) So, to isolate the effect of pay-for-performance from the recommissioning process, you would need a comparison local authority that had been retendered with absolutely no resulting change. A placebo borough? It would almost certainly be unethical for researchers to put a workforce at risk of redundancy for the sake of creating a comparator group. I’m fairly confident that there’s an irony lurking somewhere here too.
Along with the tendering process comes the effect that a different manager or provider organisation can have on a treatment service. What else happened in those 8 authorities? Did they offer naloxone? Were outreach services cut? Again, I’m drifting into causal chains that are just too long for comfort.
Payment by results was intended to improve treatment by attaching increased funding to the achievement of specific outcomes. Contingency management for a treatment system if you like. You get more people registered as having completed treatment (preferably drug free) you get more money. The Pay for Performance pilot, on which this study is based, ran from 2012 to 2014 and focused on three outcomes, drug reduction, crime reduction and improvements in health and wellbeing.
And this is the problem with ‘real world’ research into addiction treatment – there’s just so much of the real world to control. It is, however, the same challenge faced by funders. It is important to use money effectively, efficiently, and to incentivise improved treatment. The findings of this study suggest that making funding dependent on just three outcomes is not only ineffective but could also be harmful.
This study demonstrates that addiction treatment funding isn’t as simple as pay-for-performance funding processes might assume. Addiction treatment is not a simple economic transaction. Not just because it involves people’s lives, but also because there are so many variables, so many confounders and so many outcomes. This makes it daunting, but ever more important, to study.
Optimise payment by performance
The authors note that this was a pilot, and that it might be possible to improve the effectiveness of pay-for-performance models. It might be that this specific flavour of pay-for-performance doesn’t work rather than the principle of incentivising outcomes itself. Whilst this might be possible, one would hope for a few more positive indicators before pursuing it too far. There is, for example good evidence for performance-based funding around specific processes such as BBV testing and treatment. But for a whole treatment system? It seems that it’s a bit like running a pay-by-weight buffet and hoping that this will spur the chef towards Michelin stardom.
1: Mason T, Whittaker W, Jones A, Sutton M. Did paying drugs misuse treatment providers for outcomes lead to unintended consequences for hospital admissions? Difference-in-differences analysis of a pay-for-performance scheme in England. Addiction 2021 https://doi.org/10.1111/add.15486
2: Freeman, M., O’Neil, M., Rose Relevo, M.L.I.S. and Kansagara, D., 2017. The Effects of Pay-for-Performance Programs on Health, Health Care Use, and Processes of Care. Annals of Internal Medicine, 166(5), pp.341-353.
3: Mendelson, A., Kondo, K., Damberg, C., Low, A., Motúapuaka, M., Freeman, M., O’neil, M., Relevo, R. and Kansagara, D., 2017. The effects of pay-for-performance programs on health, health care use, and processes of care: a systematic review. Annals of internal medicine, 166(5), pp.341-353.
4: Advisory Council on the Misuse or Drugs (ACMD) 2017. Commissioning impact on drug treatment: The extent to which commissioning structures, the financial environment and wider changes to health and social welfare impact on drug misuse treatment and recovery. London: Advisory Council on the Misuse of Drugs (ACMD).
5: MacMillan, R. 2010. The third sector delivering public services: An evidence review. Birmingham: Third Sector Research Centre.
The opinions expressed in this post reflect the views of the author(s) and do not necessarily represent the opinions or official positions of the Society for the Study of Addiction.
The SSA does not endorse or guarantee the accuracy of the information in external sources or links, and accepts no responsibility or liability for any consequences arising from the use of such information.