A recent BMJ study showed that government cuts in England have caused extra deaths compared with trends before 2010. But how reliable are studies that link austerity with increased deaths?
There is strong evidence that budget cuts affect health. In Greece, the recent economic crisis led to large cuts in government spending. Spending on health was capped at six per cent of rapidly falling gross domestic product (GDP). A single example of the effect of this policy is on drug users and needles. In 2009, cuts were made in the number of syringes and condoms distributed by the government. In just three years, the number of new HIV infections increased from 15 to 484 a year.
In Eastern Europe, the collapse of communism led to mass privatisations of key industries. This was strongly linked to increased deaths in men of working age. This was partly due to an increase in alcohol consumption, some of which was not produced for human use.
We see that there is good evidence linking cuts to deaths, but these studies have limitations. The first is that we don’t see the alternative scenario. What would have happened if the cuts had not been made; we see increases in deaths, but could they have been avoided? A partial solution is to look at countries that avoided austerity.
One example is Iceland, which rejected an International Monetary Fund rescue package in a referendum. This led to a collapse in their currency, but little or no ill effect on health. So we can only guess how a similar policy would have affected Greece. And would a much larger country in the Eurozone have had a similar experience?
It is also well known that association does not imply causation. An association only needs two outcomes to be correlated. Causation means there is a direct link from one outcome to another, which is difficult to prove.
There are many examples of outcomes that are spuriously correlated. However, it is plausible that austerity caused extra deaths. This suggests that governments should at the very least consider whether their policies are causing unnecessary deaths.
Although some studies have attempted to predict future mortality and life expectancy, the findings may be implausible or depend heavily on certain assumptions. Even sophisticated statistical models can give quite different answers, and can be affected by quirks in the data.
There are also issues with the data these assumptions are based on. Current mortality data in wealthy countries is known to be generally accurate. This was not the case in the past, and is not the case now in poorer countries, for example in sub-Saharan Africa.
To say what caused deaths is much harder than counting them, particularly for older patients who may be suffering from several illnesses.
Much of the data and analysis is at national level. This may hide important differences between cities or regions. It may also miss unnecessary illnesses or deaths in certain social groups.
Finally, there will not be any data on the long-term effects of recent austerity for several decades.
These issues are either unavoidable or difficult to resolve. This means scientists, including in academia and at bodies like the World Health Organisation, need to study the effects of economic policies in as much detail as they can. As statisticians, we have much more data than in the past and can analyse it more easily. We need to give governments and the media as much evidence as possible. This will will help politicians and voters to make informed choices during the next economic crisis.