# Has statistics made us healthier? The role of statistics in public health

## Features

• Author: Joanna Carpenter
• Date: 17 Jul 2013
• Copyright: Image appears courtesy of iStock Photo. Figure and table copyright of Office for National Statistics

Public health – defined as the science and art of prolonging life, preventing disease and preventing disability through the organized efforts of society by the UK Faculty of Public Health – has without a doubt had great success.

Of the ten great public health achievements in the 20th century (see below), many were made possible by knowledge obtained by statistical analysis. For each achievement, the CDC has published thorough and careful essays on the US context.

The ten great public health achievements identified by the CDC are:

1. Routine immunization of children
2. Motor-vehicle safety
3. Workplace safety
4. Control of infectious diseases
5. Declines in deaths from heart disease and stroke
6. Safer and healthier foods
7. Healthier mothers and babies
8. Family planning
9. Fluoridation of drinking water
10. Recognition of tobacco as a health hazard

Prolonging life

Life has undoubtedly been prolonged. Between 1901 and 2010, life expectancy at birth – the average lifespan – increased from 45 (for a man) and 49 (for a woman) to around 76 (for a man) and 81 (for a woman).

But the picture may be more complicated than these numbers would suggest. Stephen Evans, Professor of Pharmacoepidemiology at the London School of Hygiene and Tropical Medicine, points out that life expectancy doesn’t necessarily tell you the age at which most people died. Life expectancy is an average of mortality rates at all ages,’ he continues.

I think that the contribution of statistics ought to be to make sure that people are thinking clearly about [the data], and not in a naive way just talking about those numbers,’ says Evans.

Most of [the increase] from the 17th century to the middle of the 20th century is determined by infant mortality rates, because a child dying in the first year of life … alters the [average] life expectancy dramatically,’ Evans explains.

Under the mortality rates observed in England and Wales in 1900, of 100,000 baby boys 16,868 were expected to die in their first year of life and 47,569 by their 50th birthday. By 1950, these figures had fallen to 3,361 boys in their first year of life and 13,509 boys and men by their 50th birthday, out of 100,000 births. By 2010, of 100,000 births, only 465 boys died in their first year of life (3% of the rate in 1900), and 4,556 boys and men by their 50th birthday (10% of the rate in 1900).

Figure 1: Number of Survivors by Age, from the Period Life Tables, England and Wales, Males, selected years. Source: Office for National Statistics http://www.ons.gov.uk/ons/dcp171776_292196.pdf

Preventing infectious disease – vaccines and randomized controlled trials

Efforts to improve public health have also had notable success at preventing many infectious diseases. For instance, in 1901 there were 4,793 deaths due to whooping cough in England and Wales. In the 1940s, before vaccination became routine, there were over 100,000 cases of whooping cough each year in England and Wales. In the first 10 months of 2012 there were 13 deaths, the most deaths in a single year since 1982. All of these deaths were in unvaccinated babies aged less than three months.

Central to this achievement is also a key contribution of statistics to public health: the randomized controlled trial (RCT). This was first used in 1948 to test streptomycin for treating pulmonary tuberculosis and designed by British statistician Austin Bradford Hill.

An RCT’s distinguishing feature is that an individual taking part in the trial to test a treatment is assigned at random to the treatment, with only the treatment differing between participants. This ensures that, for a large enough number of participants, differences in overall results are down to the treatment and not chance.

Although it is argued the streptomycin trial was the first RCT, similar methodology had earlier been used to test the efficacy of immunization against whooping cough, and when the UK rolled out the vaccination (by 1957) numbers of cases fell dramatically. Yet in the 1970s, a scare about side effects meant vaccination rates dropped for a while. The inevitable result was that cases rose to up to 60,000 a year.

...statistics has a role to play here, in helping the public to assess the value of vaccination by distinguishing between real and unreal problems, and the balance between benefits and harms.

Evans says that statistics has a role to play here, in helping the public to assess the value of vaccination by distinguishing between real and unreal problems, and the balance between benefits and harms.  I hope that in the future we are going to get better at educating people about this balance,’ he said.

The aim of the Understanding Uncertainty website run by David Spiegelhalter, Winton Professor for the Public Understanding of Risk at Cambridge University, is to help the public make sense of chance, risk, luck and uncertainty. His lecture on how statistics has contributed to public health will shortly be appearing on StatisticsViews.com.

Preventing disease – identifying lifestyle risk factors

Another major contribution to preventing disease is highlighted by the Framingham Heart Study, which began in 1948 – clearly a key year – and still continues. Researchers have gathered observational data’ from physical examinations and interviews about lifestyle from a large group of men and women in Framingham, Massachusetts, who initially had no cardiovascular disease (CVD). Statistical analysis has enabled the researchers to identify high blood pressure, high blood cholesterol, smoking, obesity, diabetes and physical inactivity as risk factors for CVD.

Preventing disability

The major benefit [of lifestyle factors such as obesity] is not just in making you live longer, it’s in making you live better and healthier,’ says Selena Gray, Professor of Public Health at the University of the West of England.

The positive effect of physical activity on older people is really remarkable, but it’s quite challenging to know how you capture that in a way that people understand the added benefits,’ she adds.

Gray says we need better measures of health in populations: How you measure quality of life, people’s functional ability, and their mental health and wellbeing is increasingly important.’

You can have lots of chronic diseases and still be very well. You can have high blood pressure and be on statins and it doesn’t have any impact on your life but you can have a little bit of arthritis in your knee and it has an absolutely devastating effect on your life and your ability to do things,’ Gray explains.

Two fairly blunt measures of wellbeing are published by the Office of National Statistics in addition to life expectancy numbers. Figures on healthy life expectancy (life spent in very good or good general health) and also disability-free life expectancy (life spent free from a limiting persistent illness and disability) show that, on average, people can expect to spend a portion of their lives with a disability: see Table 1.

 2008-2010 figures published in 2012 Life expentancy Healthy life expectancy Years of ill health Disability - free life expectancy Years with disability Man aged 65 17.8 10.1 7.7 10.4 7.4 Woman aged 65 20.4 11.6 8.8 11.2 9.2 Boy at birth 78.1 63.5 14.6 63.9 14.2 Girl at birth 82.1 65.7 16.4 65.0 17.1

Table 1: Office of National Statistics data on life expectancy. Source: http://www.ons.gov.uk/ons/rel/disability-and-health-measurement/health-expectancies-at-birth-and-age-65-in-the-united-kingdom/2008-10/stb-he-2008-2010.html

Evans explains, As we have prolonged life, in many instances what we are doing is keeping alive people who would have previously died – of strokes, of cancer and of heart disease – [but that does] not make people healthy. It means that some are living with some consequence of [a] disease… that in the past would have killed them.’

Official statistics show life expectancy in the least deprived areas of England was 81.4 years compared to 73.3 years in the most deprived. In the least deprived areas, disability-free life expectancy at 69.4 years was 85.3% of life expectancy. In the most deprived areas, disability-free life expectancy at 54.6 years was only 74.4% of life expectancy.

Clearly, there are stark differences between areas, perhaps due to access to health services, lifestyle or both.

Public health involves the organized efforts of society’, and statistical analysis does and should play a role in this, too.

Statistics can help illustrate people’s access to health services and use of emergency and preventive services,’ says Gray, [by] descriptive data of how people use services and whether they use them equitably... There is lots of data from the States on how uninsured people don’t use preventive services around their health.’

In the UK, Evans says, We have prevented diseases through vaccination but in terms of preventing other diseases through lifestyle changes we have been less successful.’

Gray agrees:  Statistics has a role in helping people understand risks and helping them understand how behaviour change affects those risks.’

Knowledge alone is poor at changing behaviour, though. People who smoke know that that increases your risk of lung cancer, but it doesn’t stop them doing it,’ comments Gray.

'Statistics can help illustrate people’s access to health services and use of emergency and preventive services...[by] descriptive data of how people use services and whether they use them equitably...There is lots of data from the States on how uninsured people don’t use preventive services around their health.’

Cultural norms

She points out, Behaviour is framed by our cultural norms. If everyone cycles to work and you have lovely cycle paths, you’re more likely to cycle to work.’

She says we may need action from government to change those norms: The smoke-free legislation has been phenomenally successful at reducing smoking rates and reducing exposure of children to second-hand smoke and in changing cultural norms.’

Evans points out that this is also not sufficient, given a rise in numbers of young women smoking: Physicians are seeing [chronic obstructive pulmonary disease] occurring in young people and in young women in a way that they didn’t before.’ He explains that this disease is an early sign of damage to the lungs.

The 'Nudge' Approach

In 2012 the Cabinet Office published a White Paper Test, Learn, Adapt that argued RCTs should be used much more extensively in public policy. The Behavioural Insight Team there has begun to use RCTs to identify how small changes in public policy can influence behaviour – a so-called nudge’ approach.

This isn’t new for businesses, Gray points out: When you walk around a supermarket you’re probably being nudged to do things that aren’t so good for you.’

Businesses – such as those selling not-so-good-for-you chocolate, cigarettes and cars – invest huge amounts in advertising to adapt our cultural norms to favour their products. The US car industry spent $2.84billion on online – just online – advertising in 2010, and this is projected to rise to$7.44billion by 2016.

Despite the huge advances made in public health, aided by statistics, there is still more that it can contribute in the future, with a shift to include not just providing knowledge of health risk factors, but ways to encourage individuals to live more healthily.

But in the end, statistical analysis is a tool that can be used to make us healthier, but also to make us buy and use products that make us less healthy. It has played its part in great public health achievements in the 20th century. Let’s hope the public health statisticians can outwit their less benign business counterparts.

References

View all

View all