We are now firmly into the ‘pause and listen’ mode on the great NHS reforms. Competition is a key area being discussed. But what does competition mean for the NHS? And is it a healthy thing? These are perhaps two of the most frequently asked questions.

My academic endeavours – running in parallel with my day job as finance director – have given me an opportunity to look for answers to these questions among the existing academic literature.  

It would be great if we could determine how much of the change in healthcare outcomes in the last 20 years is due to: increased funding; approaches to managing the system; advances in medical understanding and technology; and external population factors (rising obesity levels for example). Unfortunately determining causality in a healthcare system is difficult.

In fact being able to link a change in outcome to a specific change in input or practice is a bit like trying to find a politician who recognises the contribution of NHS managers – it’s rare.

Two articles by economist Carol Propper stand out in reviewing competition in healthcare.  One reviews the responses of hospitals to the more competitive market models[i], by looking at what factors determine prices.  The other assesses the impact of competition on the quality of care[ii], by evaluating three years of death rate data.  Both take the original 1989 Working for patients reforms[iii] as their context.

The first paper, which is itself dated given it stretches back to 1996, uses regression analysis to evaluate the relative importance of potential drivers of price setting, including individual hospital costs, average prices, market size, bargaining power of seller and hospital characteristics.  The paper concludes that ‘the results offer some support to the view that competition will result in lower prices in the NHS internal market.  Clearly, these results are partial[iv].

But of course efficiency in the healthcare system is only an element of the requirement to deliver taxpayer value. We also need to consider effectiveness.  A cheap system that saves no lives would not be regarded as optimal.  Hospital death rates provide one obvious way of measuring effectiveness, although they can be misleading.  In her second and later article (2004), along with colleagues, Ms Propper attempts to evaluate whether the competitive system set out in Working for patients led to better outcomes for patients, as measured by death rates after patients had been treated for a heart attack.

Published data on individual hospital death rates for acute myocardial infarction (AMI) within 30 days of admission to hospital were analysed by type of hospital (specialist heart, teaching specialist heart, London hospitals and all hospitals). Two proxies for the level of competition were factored in – the number of hospitals serving the catchment population and the number of hospitals divided by the population they serve. (This latter measure could be the more sensitive as the lower the population and higher the hospitals, potentially the higher the competition).

The research concluded: ‘We find the impact of competition is to reduce quality.  Hospitals located in more competitive areas have higher death rates, controlling for hospital characteristics, actual and potential patient characteristics.  The estimated effect of competition is small, but is robust to different measures of competition and hospital volume[v]. This research is fascinating but it can clearly be challenged.  First, is competition really linked to hospital concentration? You need more than a monopoly provider before you can have a competition. But they need to do more than just exist, they have to actually compete and that would require a marketing strategy and spare clinical capacity.  Also the choice of clinical area provides some difficulties. How much choice or competition is involved when the patient is lying in the back of a blue light ambulance having a heart attack? 

In these circumstances, the ‘choice’ of hospital will be a function of the shortest journey time and the most appropriate unit with an available coronary care bed or emergency staffed facilities.

Inconclusive at best. The researchers clearly found differences in outcomes, but had they really proved correlation and causality. I don’t think so. These studies struggle to conclude what the precise impacts, and cause and effect are, of competition in healthcare, it is therefore unsurprising that politicians and the service itself is struggling too.

 In my view there are probably areas of service delivery where competition keeps the provider sector its toes and other areas where these incentives don’t work.  The trick is how to draw the line. Any regulator trying simultaneously to determine whether the healthcare marketplace is adhering to potential duties to compete and collaborate is going to have its work cut out.

As I said before, if we are to use competition, there is a clear issue around providers or services being able to fail.  If ‘access to essential public services’ is shorthand for all services, in their current provider locations, competition may not be the tool we need to drive any required acute service reconfiguration in the next few years.