Statin use may be widening health inequalities in England

“Mass prescription of statins ‘will widen social inequalities’," The Independent reports. 

The headline is based on a new study looking at deaths from coronary heart disease in England from the years 2000 to 2007.

The good news is that overall deaths from heart disease were estimated to be down by a third (34.2%) during the time period.

The bad news, at least for those concerned about health inequalities, is that the use of statins (a cholesterol-lowering drug), benefitted the richest 20% more than the poorest 20% of society.

This is unlikely to be due to any biological factor and may instead arise due to a combination of socioeconomic and cultural reasons, such as people with chaotic lifestyles associated with poverty being less likely to stick to a treatment plan.

The study also found that population-based approaches – such as encouraging people to stop smoking, eat a healthy diet and take regular exercise – have had a much bigger impact than medical approaches, like statins.

This led the study authors to suggest that there needs to be a greater emphasis on population-based approaches in the future, if we are not to see health inequalities widen further.

This study usefully informs debate in the public health sector about the best and fairest way of continuing this reduction in the future. 

Where did the story come from?

The study was carried out by researchers from the University of Liverpool, University of Chester, University College London, Public Health Wales and the University of British Columbia (Canada). It was funded by the National Institutes for Health Research School of Public Health Research and Liverpool PCT FSF scheme.

The study was published in the peer-reviewed medical journal BMJ Open. As the name suggests, this journal is open-access, meaning anyone can read that full article online for free.

Different UK newspapers emphasised different angles of the story (which seemed to be linked to their political editorial line), but they all covered the facts of the study accurately. 

What kind of research was this?

This was a modelling study trying to work out what proportion of a fall in coronary heart disease deaths in England was due to preventative medications, like statins, and what proportion was due to population-wide changes like diet and exercise. They were also interested in exploring the relative effects on different socioeconomic groups.

The UK, the study authors inform us, has experienced a remarkable 60% reduction in coronary heart disease mortality since the 1970s, largely due to reductions in things like smoking. However, coronary heart disease remains the leading cause of premature death.

This study wanted to find out whether the decline was mainly due to medicines, like statins, or population-wide approaches like stopping smoking, good diet and exercise. They also knew that many of the risk factors of coronary heart disease show a social gradient, with the poorest worst affected. The team were interested in whether medicines or lifestyle changes made these social inequality differences bigger or smaller.

Modelling studies like this use existing data to estimate the relative impact of different variables (e.g. statin use) on an outcome (e.g. death). The advantage of models are that you can play around with the parameters to see what the most important influences are, and this can help target resources to give the most value for money in the future. However, all models rely on a range of assumptions and are only as good as the quality of their inputs and their design.

As the old software engineer saying goes “GIGO”: garbage in, garbage out.

It’s important to assess whether the model has realistic assumptions and if its data is relevant and of good quality.

What did the research involve?

The study team pulled together data from randomised controlled trials, meta-analyses, national surveys and official statistics to input into a statistical model. They then ran a series of statistical tests to estimate whether the relative contribution preventative medicines, reduction in blood pressure and cholesterol levels had contributed to the decrease in coronary heart disease deaths. The data came from adults over 25 living in England, gathered between 2000 and 2007. 

The main outcome of interest was number of deaths prevented or postponed (DPPs) in 2007, stratified by socioeconomic status.

For the number crunching, they used a model called the “IMPACTSEC model”.

This is a statistical technique that takes results from previous studies to make an estimate about the relative contributions, specific treatment and risk factors make to reductions in death rates. 

Or, in laypersons’ terms: it takes results from previous studies to make an estimate about how likely a particular intervention is in preventing or postponing deaths.

The first part of the IMPACTSEC model calculates the net benefit of statins and antihypertensive treatment in 2007. The second part of the IMPACTSEC model estimates the number of DPPs related to changes in systolic blood pressure and cholesterol levels in the population. They realised that there was overlap between pharmacological and non-pharmacological contributions to risk factors, and adjusted for this in their model.

What were the basic results?

Populations approach vs. medications

In 2007, the model estimated that there were approximately 38,000 fewer coronary heart disease deaths than if death rates had continued at 2000 levels. A large proportion of these, approximately 20,400 DPPs, were attributable to reductions in blood pressure and cholesterol in the English population (population-based approached). A much smaller number, approximately 1,800 DPPs, came from medications such as statins.

The remaining DPPs were attributed to other factors.

Impact by socioeconomic group

Reductions in population blood pressure prevented almost twice as many deaths in the most deprived fifth of society compared with the most affluent.

Reductions in cholesterol resulted in approximately 7,400 DPPs, of which 5,300 DPPs were attributable to statin use and approximately 2,100 DPPs to population-wide changes.

Statins prevented almost 50% more deaths in the most affluent fifth of society compared with the most deprived. Conversely, population-wide changes in cholesterol prevented threefold more deaths in the most deprived fifth of society compared with the most affluent.

How did the researchers interpret the results?

The study team welcomed the reductions in death rates from coronary heart disease over the last 30 years, but raised concerns that the improvements were not spread equally amongst society. They questioned whether health inequalities might get worse if future efforts focussed on policies to increase the use of statins, rather than on population-based approaches.

They concluded: “Our results strengthen the case for greater emphasis on preventive approaches, particularly population-based policies to reduce SBP [systolic blood pressure] and cholesterol”.

Conclusion

This modelling study estimated that population-based approaches to reduce heart disease death rates in England have helped the poorest in society the most, while the effect of statins has benefited the most affluent. This led the study authors to suggest that there needs to be a greater emphasis on population-based approaches in the future, if we are not to see health inequalities widen.

Report author Martin O'Flaherty said in the Telegraph that: “The success of clinical cardiology in providing cost-effective treatments that are based on scientific evidence needs to be celebrated. However, population-wide measures might offer substantially bigger health gains, relieve pressure on an already stressed health system and reduce health inequalities. Measures like controlling tobacco, increasing physical activity, improving the contents of processed food products, restricting the marketing of junk food, taxation of sugary drinks, and subsidies to make healthier foods more affordable require renewed attention not just from academics, but crucially from people and policymakers”.

It is not totally clear how reliable and robust the model used in the study was, or the conclusions that stemmed from it. It is possible that different results and conclusions could have been reached if the inputs had been from different data sources, or the model configured differently.

That said, the researchers took all reasonable measures to mitigate this, and their conclusions remained stable throughout, so we can consider it relatively reliable. The reliability of the conclusions would be increased if they were supported by other studies using a variety of data sources.

The study is useful in informing debate in the public health world about the best and fairest way of reducing heart disease deaths in England, which is always a question of targeting and prioritising finite resources.