In healthcare, some things that count are hard to count…but we shouldn’t stop trying to count them
01.08.14
When healthcare managers want to change behaviour through targets, for example, they have to be aware of so-called ‘perverse incentives’ that occur because sometimes the thing you are targeting is not a perfect measure of what it is you are trying to achieve. We had a good example of this in the UK a few years ago when there was a lot of discussion on the way in which ‘waiting times in casualty department (emergency room)’ targets were being implemented. In focusing on the thing they could measure – how much time it took for a patient to be transferred out of causality – managers inadvertently created an incentive for the patient to experience even worse treatment than the target was trying to prevent. Rather than breaching the (as I recall) four-hour limit, there were stories of patients being wheeled out into corridors and left there – sometimes for hours! Not exactly what the managers had in mind…
There is a similar issue when it comes to the evaluation of health technologies, including medicines. In an ideal world, you’d want to look at all the cost and benefits of a particular technology in order to weigh up whether or not it is worth investing in it, at the price charged by the supplier. The trouble is, part of what we need to know is easy to measure – the costs. But the other part – the benefits – are often incredibly hard to quantify. The idea of a ‘cost-benefit’ analysis, then, looks pretty useless if you can only measure the costs.
The Milken Institute – a respected, non-partisan, not-for-profit economic think tank – has just published an impressive study on this issue, mostly looking at medical devices – but the principles can easily be applied to medicines. They make the point that the benefits, which although substantial, are often indirect – for example reducing hospital care. Wider economic gains – through, for instance, improving labour market participation, reducing absenteeism, improving productivity, etc. – are even trickier to quantify.
Motivated by the observation that such indirect benefits are rarely incorporated into the evaluation of medical technologies, the Milken Institute set out to estimate the size of the issue. They looked at a range of technologies in four of the most prevalent causes of death and disability: diabetes; heart disease; musculoskeletal disease and colorectal cancer. They found that once indirect benefits were taken into account the net benefit from the technologies was $23.6bn. Included in this is the fact that, by improving labour market outcomes (getting people back to work, or keeping them in work), tax revenues increased by $7.2bn.
Having estimated today’s net benefit the authors then looked at what might happen in 2035 based on three scenarios: “1. Reduced incentives to invest in improvements to technology; 2. A continuation of current trends; 3. An acceleration of innovation, associated with enhanced incentives”. The results show a cumulative gain of $1.4 trillion (in 2010 dollars) for scenario 3 (increased incentives for more innovation) and a cumulative loss of $3.4 trillion for scenario 1 (reduced incentives for more innovation).
As with all economic projections, there is obviously scope for debate about methods, assumptions, accuracy, data, etc. But, in a difficult area, it seems to me the authors have made a significant effort to come up with as accurate (and, if anything, conservative) picture of the future as possible.
The scale of benefits that are routinely ignored when evaluating technology is striking. Will it always be possible to incorporate similar analysis into all future decisions in a reliable way? Perhaps not, but we should start to try. And in the mean time, those making a final judgement on technologies should be aware of the limitations of cost-benefit analyses on which they base decisions – and perhaps adjust their decisions accordingly.
