In previous posts (here and here) the comprehensive work undertaken by Claxton et al on the returns to medical expenditure in the NHS was discussed. Claxton and colleagues estimated the average change in quality adjusted life years (QALY) that have resulted from marginal budget changes in the NHS. Their estimate was £9,000 per QALY, much lower than the current threshold of £20,000 to £30,000 per QALY, so they argued that the threshold should be reduced. While we discussed some possible criticisms of the research it nevertheless remains the best empirical work on the topic and does provide some evidence for a reduction in the threshold. Many national news outlets picked up on the story when the paper was published. Nevertheless, changes to NHS reimbursement decisions and associated policies may have effects beyond that of altering the cost-effectiveness of the portfolio of treatments provided by the NHS.
A recent paper by Koijen, Philipson, and Uhlig argues that uncertainty about government healthcare policy in the United States has reduced medical research and development (R&D) and as a consequence reduced overall medical spending as a share of GDP. The equity returns of firms in the health care sector are higher than in other sectors: they find there is a ‘premium’ of around 4-6%. This, they argue, is a consequence of there being higher risk in the health care sector.
Figure from Koijen, Philipson, and Uhlig (2016)
Koijen, Philipson, and Uhlig suggest that this higher risk is a result of uncertainty about government regulation in the health care sector. The figure above demonstrates a large drawdown in the healthcare sector, not related to other sectors, at the time when the Clinton health care reform was being discussed. However, the figure does not show an equivalent decrease specific to the health care sector for Obamacare (around 2008-10). Obamacare likely improves health care sector revenues by adding a large number of health care consumers to the market. This quantity effect outweighs any reduction in markup. Indeed, the healthcare sector spent around $150 million dollars lobbying in support of the Affordable Care Act. However, the effect of the Clinton health reform would have been to impose price controls, thus reducing returns to the health care sector.
That medical innovation is directed towards the areas with the highest returns is an uncontroversial idea. It is the reason so few new treatments are developed for tropical diseases and tuberculosis, for example. Changes in government policy that affect returns of investment are likely to affect investment decisions. A reduction in the threshold for reimbursement will reduce return on investment in the UK. Uncertainty about future health care policy may also do so for the above reasons. The health care sector is particularly sensitive to American health care policy since the US accounts for 48% of global medical spending, so such changes in the UK may not lead to large effects. But for firms whose primary customer is the NHS, this may well be an issue.
A potential solution is to increase subsidies for medical R&D in the event of a reduction in the cost-effectiveness threshold. Indeed, this is one potential solution to encouraging the development of treatments for those diseases that predominantly affect people in the global South. However, what the article above demonstrates is that if there is a large amount of uncertainty about health care policies then any R&D stimuli may not have their intended effects.
[…] as the Vioxx scandal demonstrated. We have also previously reported on how policy uncertainty reduces pharmaceutical R&D. Thus, state involvement in the industry seems to be […]
In Canada, we recently had a federal government policy of subsidizing R&D under the guise of “grants” for health research. These were highly targeted funding opportunities that required applicants to argue that their research would produce cost reductions over a reasonable timeframe (it probably wasn’t an accident that the “reasonable timeframe” roughly coincided with scheduled elections).
While I agree with the conclusions drawn in this paper – reduction of uncertainty allows us to achieve thresholds that are more in line with the empirical evidence – I am wary of the idea of increasing R&D subsidies. Not because I don’t think it’s a reasonable idea, but because I and my colleagues lived through a period where basic science funding was slashed and public-sector researchers had to start doing R&D in order to keep the lights on. If subsidies are to be increased, it’s important that we discuss how that is done.