The British Doctor’s Study: The Link Between Smoking and Lung Cancer

In science, you often hear the phrase ‘correlation does not equal causation’. It is a ubiquitous line in science that often requires a lot of explaining. One of the best ways to demonstrate correlation vs causation is a website by Tyler Vigen called Spurious Correlations.

For example (see below), as the US government spends on science, space and technology there is an increase in the number of suicides by strangulation, suffocation and hanging. Does this mean science, space and technology are causing people to kill themselves? Surely this is a thing we should know about? Why is nobody angry?

This is a correlation. As one goes up so does the other – but this doesn’t mean the two are related to each other, it could just be a coincidence. In the example from Vigen, it is purely a coincidence that there is an increase in the number of hangings as US spending increases – there is no causation as the two are clearly unrelated.

Another clear coincidence is that an increase in ice cream sales correlates with an increase in the sale of meat. This does not mean that people are getting wild with their ice cream meat combinations (chocolate fudge pork anyone?) there is a much simpler explanation. The weather.

As it gets hot outside you are more likely to buy ice cream and likely to have a meaty BBQ. The increase in both ice cream and meat sales correlates with the weather, not with each other.

The conclusion between the weather and ice cream is an obvious one to make. Establishing the causal link between smoking and lung cancer and ending a habit millions have had for millennia was going to be a lot more difficult.

---

In the 1950s, smoking was a normal part of daily life. According to the charity Action on Smoking and Health (ASH), the percentage of male adults who smoked daily in 1948 was as high as 82%. The smoking rate of women was also relatively high at 41%.

The idea that smoking caused any particular harm was a little-known fact at the time and except for a small number of individuals, many did not see an issue in smoking, no causal link had ever been established. Throughout the US in particular, smoking was advertised as a treatment for numerous medical conditions with doctors advocating the behaviour.

For example, the cigarette company Lucky Strike ran an ad campaign that stated 20,679 physicians approved the cigarettes for protection again throat irritation (ironically). Lucky Stike sent a bunch of doctors free packets of cigarettes and asked if they found it less irritating on their throat than other brands. 20,679 of those said yes and an advertising slogan was born. This wasn’t just Lucky Strike however, all cigarette manufacturers got in on the deal of misinforming customers to varying degrees of success.

At the time, a small correlation between smoking and an increase in cancer, particularly tongue cancer, was observed but the consensus at the time was that smoking wasn’t a medical issue. There was a large increase in the number of cases of lung cancer, however. There was a causal link, there was no correlational link either. Initial suspicions believed this was caused by the rapid increase in polluting industrial activities but this conclusion was a simple one to make, nearly everyone smoked, but not everyone developed lung cancer, so it had to be something else, right?

With an increasing suspicion that smoking increases the prevalence of lung cancer, what was needed was a large population of smokers and non-smokers that could be contacted over several years and also supply relatively accurate self-reported data on smoking habit and health.

What was needed was doctors.

---

In 1951, Sir Richard Doll and Sir Austin Bradford Hill contacted all General Medication Council UK registered doctors and asked them to complete a questionnaire about their smoking habits. Of the 60,000 doctors who were contacted, more than 70% responded. (This did only end up being limited to males over 35 years of age, as the rate of cancer in this age group was low, as was the number of female doctors).

They collected 34,000 responses from doctors born between pre-1900 and 1930 and continued to observe patients for the next 50 years. After their first 10 years of follow-up, there were 4597 deaths from peptic ulcers, bronchitis, tuberculosis, coronary heart disease, and of course, lung cancer.

This was the first study to link the number of cigarettes smoked to the risk of death from lung cancer. For doctors who didn’t smoke the death rate of the disease was 0.07 per 1000, however, for those who smoked more than 35 cigarettes per day, the number was 45 times higher at 3.15 per 1000 dying of lung cancer.

Finally, a causal link between smoking and the prevalence of lung cancer.

The findings didn’t end there. Those who stopped smoking also saw a reduction in their risk and also demonstrated a link between other diseases (namely heart disease, heart attacks, and chronic liver disease) and those who smoked a higher number of cigarettes.

The decades of follow-up questionnaires continued to add more and more credence to the problems of smoking and allowed people to make a truly informed decision that would increase their health. Almost 50% of smokers’ cause of death was due to a smoking-related disease and the years shaved off of life for those who continue to smoke. For example, those who smoke until they are 50 years old lost 4 years of their life, and smoking up to 60 deducts 7 years of life. An informed decision that shows the earlier you stop smoking the larger the increase in life expectancy.

Over the following decades since the study began, there has been a marked decrease in the number of smokers. At its peak during the 1970s, 35% of all deaths in men were likely caused by smoking-related activities. The peak for women came much later in 2000 with 16% of all deaths being caused by smoking. The work done by Sir Richard Doll and Sir Austin Bradford Hill in this study has contributed to a global increase in smoking cessation and led to millions of individuals saving themselves from a bucket kicked far too early.

There is still work to be done in reducing the number of smokers from more disadvantaged backgrounds. 1 in 4 unemployed people in the UK is a smoker almost twice as high as those who are employed. Only 7% of those with a degree regularly smoke whilst those with no formal qualifications in over 4 times higher at 29%. Even those who are divorced at 7% more likely to smoke than those who are still married or in a civil partnership.

14% of the UK population smoke cigarettes and whilst this is falling year on year smoking is still the leading cause of preventable death in the UK with around 95,000 deaths linked to smoking. Treatment for smoking-related diseases has an estimated cost of £2.5 billion to the NHS and £1.4 billion to social care. Large-scale efforts are now being used to discourage more and more people from taking up smoking in the first place. Plain packaging has been one that has had a stark effect in reducing the number of smokers and also making it less attractive for children. In the years before plain packaging, an estimated 200,000 children tried smoking every year. The displaying of smoking on TV programmes in the UK is also heavily regulated, smoking indoors has been banned since 2006/2007 and advertising cigarettes or tobacco on TV and radio is also banned.

---

As discussed demonstrating a causal link is a difficult one that can take years. If you make an observation that as one thing increases so does the other, this could just be a coincidence or be a completely different factor – this is a correlation. Until we focus our efforts on finding more of these links there will continue to be a lot of ‘coincidences’. Fixing this needs public pressure, private government and money, and scientific willpower.