Personal Decisions are the Leading Cause of Death

In the November-December 2008 issue of Operations Research, Dr. Ralph L. Keeney of Duke University, in the paper “Personal Decisions are the Leading Cause of Death”, argues that personal decisions cause a large fraction of the deaths in the United States, with the vast majority of those decisions resulting in the death of the decision maker. Some of the key decisions people make involve diet, alcohol and drug use, exercise and driving habits. From the abstract:


This paper analyzes the relationships between personal decisions and premature deaths in the United States. The analysis indicates that over one million of the 2.4 million deaths in 2000 can be attributed to personal decisions and could have been avoided if readily available alternative choices were made. Separate analyses indicate 46% of deaths due to heart disease and 66% of cancer deaths are attributable to personal decisions, about 55% of all deaths for ages 15–64 are attributable to personal decisions, and over 94% of the deaths attributable to personal decisions result in the death of the individual making the decisions. Relative to the current 45%, retrospective appraisal suggests that roughly 5% of deaths in 1900 and 20%–25% of deaths in 1950 could be attributed to personal decisions. These results suggest that more effort directed toward improving personal choices regarding life risks may be an effective and economical way to save lives.


You can read the full paper here: pdf Personal Decisions are the Leading Cause of Death .

The editors of Operations Research, led by Area Editor Stefanos Zenios, asked a number of leading researchers to provide commentary on this paper.

Dr. Jonathan Caulkins of Carnegie Mellon University provides the first commentary. He agrees with Dr. Keeney’s basic thesis:


There is little doubt about Keeney’s basic points that very many premature deaths in the U.S. (~1million/year) could have been avoided if the decedents had lived their lives differently and that living differently would not have required unusual insight or great resources. In that sense, it was entirely within the power of these people to have prolonged their lives.


He offers, however, “three respects [in which] there is more at play than decision making, at least in the narrow sense of identifying the best alternative for a decision maker through reflection and analysis.” Those three are:

  1. alternatives are selected without people making specific, conscious decisions,
  2. picking the best option seems to be hard, and
  3. perhaps only individuals make decisions, but they do so in a social context.

You can read Dr. Caulkin’s full commentary pdf here .

The second commentary is provided by Drs. Anupam B. Jena, Tomas J. Philipson, and Eric Sun of the University of Chicago. In their commentary, they point out that the issue might not be one of “poor” decision making:


… another way of interpreting Professor Keeney’s estimates is that the costs of living a healthy lifestyle are so large and so prevalent that nearly one million individuals annually would be willing to die prematurely to have higher utility in the years of life that they are alive. For the same reason that people often prefer to save time by driving (less safe) highways rather than local roads, individuals may rationally choose to live unhealthy because the benefi ts outweigh the costs.

The interesting question raised by Professor Keeney’s estimates, then, is the approximate value of interventions aimed at promoting healthy lifestyle choices. The answer depends on one’s view of the world. At one extreme, if people are fully informed about the risks of “poor” lifestyle choices, then these interventions will have little value. At the other extreme, if people are completely uninformed about the benefi ts of healthy behaviors, then these interventions could be of great value. Indeed, at a firrst pass, the value to society would be the value of life generated by saving one million premature deaths times the additional years of life lived. Of course, new compensating behaviors may arise that might mitigate the e ects of such interventions, but in general, the value would be quite large. Professor Keeney’s estimates suggest that interventions aimed at promoting healthy lifestyles may well have a large social value to society. Further research, then, is needed to understand why people make the choices that they do, and how this might impact policies aimed at promoting healthier lifestyles.


You can read the full commentary pdf here .

The third commentary comes from Drs. Victoria L. Brescoll and Kelly D. Brownell of Yale University. Brescoll and Brownell agree with the data and classification, but worry about the prescriptive aspects of the paper:


As we see it, the major weakness in this paper does not lay in his calculations or classification scheme for estimating the proportion of premature deaths due to personal decisions but rather in the prescriptions that he draws from these data.

Specifically, the proposition that “more effort directed at improving personal choices” is an effective and economical way to save lives is the public health equivalent of the bridge to nowhere. It is an old and tired argument and, unfortunately, has not shown to be true in a number of public health domains. In fact, historically, campaigns focused on changing individual’s behavior at the expense of changing the environmental factors that contribute to these behaviors have not been successful. As Keeney shows in his analyses, deaths due to personal choices have only increased in the last century, a time when health information available to consumers has grown dramatically and government officials have implored people to stop making decisions that harm their health.


You can read their full commentary pdf here .

The final commentary comes from Dr. Milton C. Weinstein of the Harvard School of Public Health. Dr. Weinstein believes that Dr. Keeney’s work may “overestimate the number of choice-related deaths”. For instance:


…some of the behaviors that Keeney considers to be personal choices may be less than fully voluntary. For example, recent research suggests a genetic link to obesity, and that people who carry such genes may have higher thresholds for appetite satisfaction than others. Likewise, addictive behaviors such as cigarette smoking cannot be easily changed. Even if the initial decision to start smoking is made before it becomes addictive, the decision maker may be a child at the time.


Dr. Weinstein’s full commentary is here: pdf Weinstein commentary .

Dr. Keeney was given the opportunity to provide a rejoinder to these comments. Here are some excerpts of that rejoinder:


I agree with many of the comments and feel that some of the apparent differences are perhaps due to shortcomings in my presentation of the intent of the paper and its suggestions that I believe follow from the data and analysis.

[Purpose of Article.] As Jena et al. (2009) note,
“personal decisions are involved in nearly all deaths.” I limited the range of personal decisions to those with readily available alternative choices that were not expensive or time consuming (e.g. use your car seat belt, don’t smoke) and to where alternatives, if they could be followed, would potentially appeal to many individuals (e.g. never riding in a vehicle would greatly diminish the chance of dying in a vehicle accident, but this alternative would not appeal to most individuals).

[Meaning of Good and Poor Decisions] In simple terms, a good decision is one made that is consistent with the information that the decision maker has about the possible consequences of that decision and his or her values for those possible consequences. A poor decision is one made inconsistently with such information and values. As I state in the article, “an individual can coherently decide that he would rather continue to smoke because the enjoyment he gains from smoking outweighs his concern for the risks to his health and longevity.” Such a decision would be a good decision.

[Improving Decision Making] Brescoll and Brownell (2009) e ffectively say that improving personal decision making is ineffective and uneconomical, and support this with “campaigns focused on changing individual’s
behavior have not been successful.” These campaigns, I suspect, tell people what to do and why it is good for them. Many people are not overly receptive to being told what to do. My suggestion is different, as it does not tell people what to do. Instead, it empowers them to make their own decisions consciously and it suggests a common-sense process to think clearly and consistently with their information and values.

[Quality of the Data] The main sources of data used in my article are statistics of medical causes of death and analyses by various medical, health, or safety professionals about the actual contributing causes of death. It is my judgments, explained in the article, that specify which actual causes of death could have been averted by choosing different readily available alternatives. Based on the analysis, and a few small sensitivity analyses, I concluded that approximately one million fatalities (of the 2.4 million deaths) could have been averted in 2000 if individuals had chosen different healthier and safer alternatives. As I stated in the article, I gave precise estimates of the fatalities averted so readers could understand and appraise my calculations. Even if many of the fatalities that I concluded could have been averted by making different personal decisions were not counted using different judgments, it would seem that many hundreds of thousands of lives would not have been lost using any reasonable set of different assumptions.


Dr. Keeney’s full pdf rejoinder is here .

Now it is your turn. What thoughts do you have on Dr. Keeney’s theses? Why do so many of us make decisions that shorten our lives?