Growth in spend on healthcare grew faster than growth in the gross domestic product in every OECD country between 2000 and 2008. That’s the main reason why every country, including Britain, is trying some form of health reform. The growth in spend on healthcare is unsustainable, and, as I’ve joked before, the US may be the first empire to be brought down by health care rather than disease. But why the growth, and can anything be done about it?

The anxiety of governments over the expenditure on healthcare is amplified by their observation that higher expenditure doesn’t lead to better outcomes. Alain Enthoven, the US health economist who devised the idea of the internal market for Britain, long ago observed that increasing expenditure did lead to better outcomes until a point he christened “flat of the curve,” where increased expenditure did not lead to better outcomes. There may even be a point where more expenditure means worse outcomes as doctors launch into unnecessary procedures.

Where is the point where greater expenditure ceases to deliver better outcomes? A plot of adult mortality between 15 and 60 against health care expenditure per head ( expressed in purchasing power parity) for many countries suggests that the point is about $2500 a year. Above that expenditure not only does adult mortality not improve but it worsens, particularly for the US, where expenditure is over $7000 and mortality is roughly 50% higher than the lowest. There are also countries where expenditure is below $2500 per head but outcomes very good – Chile, China, Singapore, and Korea.
Another insight into diminishing returns comes from a US study that showed that between 1985 and 1997 life expectancy for patients with lung cancer improved by one month but costs increased by $20 000.

The usual explanation for rising costs comes from the demand side – ageing populations and rising expectations. But a new report from Volterra Health argues that the supply side is more important. New technology alone has been estimated to account for half of the increased expenditure in the US since 1960, and more hospitals and doctors means increased expenditure and activity. We might like to think that more demand comes first, but followers of Jack Wennberg, the US health policy expert, will be familiar with the idea of “demand induced supply.” He points out that Los Angeles has twice as many hospital beds, doctors, and intensive care beds per head and twice the health spend compared with Minneapolis – and yet worse outcomes. And every new doctor means a lifetime health expenditure of $5m.

Volterra Health, which has market enthusiast Nick Bosanquet as its chair, has an interesting analysis of “old and new models of healthcare.” The “old model” ran uninterrupted from 1880 to 1980 and is still the predominant model – but with “new” bolt ons. With the “old model” “patients visited doctors, presented their symptoms and were then either treated for these symptoms there and then, or let through the gate into the specialist universe of diagnostics and treatment in hospital.” There may also have been public health and school health programmes, but they were sideshows.

Mainly driven by worries about rising costs but also by the increasingly old fashioned inflexibility of healthcare systems, healthcare funders, mainly governments, set about devising “new models.” They comprise, say Volterra Health, prevention, screening and early diagnosis, ambulatory treatment, workforce specialisation, disease management, and new treatments and technologies. None of them has achieved the Holy Grail of unequivocally improving outcomes and reducing costs.

Prevention “has been more success in reducing early mortality than in reducing long term illness,” and unfortunately it has probably increased inequalities. The effectiveness of screening is highly variable among and within disease groups, and, although it may have improved outcomes, it has probably increased total costs. Ambulatory care, although cheaper than hospital care, has stoked demand. Volterra Health believes that workforce specialisation and disease management have improved outcomes and reduced costs, but many might dispute both claims, particularly as the higher the ratio of secondary to primary care within a health system the higher the costs and the poorer the outcomes.
As is almost always the case, the identification and analysis of the problem is better than the proposed answer. Somewhat half heartedly, Volterra Health prescribe “conditions for greater innovation,” particularly new players, and disease level strategies based around the “new model.” Perhaps the best thing about the report is its title, “The medical arms race,” and the most interesting outcome for me was circulating it to a listserve of 300 colleagues, most of them in low and middle income countries. They reacted passionately, saying “Why don’t these groups from high income countries stop preaching to us.” Indeed, they suggested a reversal of the usual pattern and proposed that a group from low income countries tell high income countries how they could improve their health systems. Not a bad idea.

Competing interest: RS is employed by the UnitedHealth Group, which has a UK subsidiary, UHUK, which might benefit from this report if it leads to encouragement for “new players,” but the present government is keen on that anyway. RS runs a philanthropic programme for United and does not work for UHUK. He has shares in United, but it’s hard to see that what happens in the UK will have much influence on the share price for a long time to come if ever. He’ll probably be dead by then.

Richard Smith was the editor of the BMJ until 2004.