In this International Year of Statistics, I'd like to describe the major role of statistics in public health advances. In our modern society, it is sometimes difficult to recall the huge advances in health and medicine in the 20th century. To name a few: penicillin was discovered in 1928, risk factors for heart attacks and stroke were established in the 1950s, and vaccines were created throughout the latter half of the century to prevent diseases that once killed thousands of children annually.
A few weeks ago, SAS was fortunate to receive a visit from Christy Chuang-Stein, Vice President and Head of Statistical Research and Consulting Center at Pfizer, and a candidate for president of the American Statistical Association. One of Christy's slides mentioned that the Centers for Disease Control and Prevention (CDC) published a list of the "Top 10 Great Public Health Achievements in the 20th Century." The CDC articles are fascinating but lengthy, so let me give you the executive summary and simultaneously emphasize the role of statistics in a few of these achievements:
- Routine immunization of children: During the 20th century, researchers developed vaccines that prevent smallpox, measles, polio, and other diseases. The safety and efficacy of these life-saving vaccines were tested by using statistically designed clinical trials and statistical quality control during manufacturing.
Many of these studies were conducted out of the public's eye, but in 1954 there was a massive public trial to test Jonas Salk’s polio vaccine. More than 1.8 million children participated in a randomized, double-blind trial. This is a statistical design in which subjects are randomly assigned to either the control group or the vaccine group. Neither the doctors nor the parents knew which child received the vaccine instead of a placebo. This famous experiment was a success. Today, pharmaceutical companies run similar statistical studies as they develop drugs for the treatment and management of a wide range of maladies.
- Motor-vehicle safety: In 1925, about 18 Americans died for every million miles traveled. By the 1990s, that average mortality rate had dropped to 1.7 deaths per million miles. Engineering (both in vehicles and on the roads) had a large part to do with that decrease, but statistics played a role in identifying key risk factors that contributed to vehicular deaths, including statistics about the use of seat belts, infant restraint systems, booster seats, and statistics about the value of graduated licensing for teenage drivers.
- Declines in deaths from heart disease and stroke: Although heart disease and stroke are the first and third leading causes of death in the US, respectively, death rates due to heart disease have decreased 56 percent since the 1950s, and death rates from stroke have decreased by 70 percent. As I described in a previous article about Jerome Cornfield, carefully designed statistical studies such as the Framingham Heart Study established the major risk factors: high cholesterol, high blood pressure, smoking, and dietary factors such as cholesterol, fat, and salt.
- Safer foods: Several times a year, the news tells us about recalls of food that are linked to a foodborne disease such as Salmonella or E. coli. Tomatoes, spinach, lettuce, ground beef—these and other food products have recently been in the news as sources of outbreaks that are geographically diverse, yet are eventually traced to a common cause, such as poor sanitation at a single processing facility. Kaiser Fung's book, Numbers Rule Your World, presents a fascinating look at how statistical methods (coordinated through the CDC) are used to detect, investigate, trace, and control outbreaks of foodborne diseases.
- Tobacco as a health hazard: The CDC article notes that smoking is known to be the "leading preventable cause of death and disability" in the US. But the health risk of smoking was unknown before epidemiologists and statisticians began analyzing data in the 1940s and 1950s. By 1964 the evidence from thousands of studies convinced the US Surgeon General to conclude that "cigarette smoking is causally related to lung cancer" and to other diseases. The statistics associated with establishing causality led to a lengthy debate between statistician Jerome Cornfield (and colleagues) at the National Institutes of Health and Sir Ronald Fisher, a heavy smoker who was a brilliant statistician but a bully toward those who disagreed with him. In the end, science prevailed over intimidation, and the weight of statistical evidence has led to laws that have decreased the prevalence of smoking in the US, as shown in the following graph:
The other achievements in the top 10 list were workplace safety, control of infectious diseases, healthier mothers and babies, family planning, and fluoridation of drinking water.
After her talk, I asked Chuang-Stein whether the 21st century has produced any advances that compare with those on the CDC list for the 20th century. She replied that the following two achievements are likely to make the list for the 21st century:
- Personalized medicine: In personalized medicine, an individual's genetic profile and his or her unique biochemistry are used to customize treatment. For example, which medicines are likely to provide the best results with the fewest side effects?
- Disease modification: If a person is diagnosed early enough, it might be possible to inhibit the disease so that it never debilitates the person. For example, current research on Alzheimer's disease aims to inhibit the progression of the disease. Disease-modifying treatments are available for several diseases that develop slowly, including multiple sclerosis and rheumatoid arthritis
Certainly medicine, chemistry, engineering, and other fields also contributed to these public health successes. Statistics is a collaborative science, and it does not make its contributions in isolation. Cornfield called it "the bedfellow of the sciences." Nevertheless, statistics made these public health advances possible. And that is truly something to celebrate!