Fischbeck’s model predicts the probability of dying
For many people, death is high up on the list of biggest fears in life. With a Carnegie Mellon-based website, DeathRiskRankings.com, people can now find statistical evidence that will tell them their odds of dying within a certain period of time. Professor Paul Fischbeck, who is part of both the social and decision science and engineering and public policy departments, originally came up with the idea following his research into risks of all types, from engineering risks to investment risks. The website began as an undergraduate research project within the information systems department as early as the fall of 2008.
The project began primarily due to the ongoing congressional health care debate, a major issue facing the government since mid-2009. In arguments concerning different healthcare issues, the two major metrics that are often used are life expectancy and infant mortality. DeathRiskRankings.com concerns itself with the other variables in between, such as age, race, and whether the participant is from America or Europe. It even takes into account different types of cancer, and the differing mortality rates between countries and genders. Using these criteria, one can calculate the chances of dying in the next year, the next five years, and so on. After infancy, annual risk of dying increases exponentially, according to a Carnegie Mellon press release.
DeathRiskRankings.com followed another one of Fischbeck’s risk assessment projects, TrafficStat. This website calculated the risk of dying from automobile accidents, depending on factors like the brand of car and where a person was driving. Despite its modest publicity, appearing on The Today Show and CNN, the site did not see the attention that Fischbeck had expected. “The ultimate risk is life and death,” Fischbeck said from his office in Porter Hall. “There’s always been an interest in quantifying how risky various things are." The website was released in August 2009, and during the first few weeks, the website received about a million hits an hour. Now, according to Fischbeck, “it’s gone way down in traffic … and I get e-mails almost daily from people wanting to use it for policy discussion.”
The variables themselves were based on death certificate data held by the Centers for Disease Control in the United States and data held by various organizations around Europe. Due to several differences in the variables and the time increments in which the data was collected, the Carnegie Mellon group had to make simplifications and compromises.
The Internet is no stranger to sites predicting dates of death: Death Clock has satiated people’s fear of dying for many years. Fischbeck is quick to point out differences. “What we do is we allow this direct comparison between all 50 states and about 20 European countries. And we have 66 causes of death. No other site does that.” Side-by-side analysis between, for example, a white American 55-year-old man and a white French 40-year-old man can yield some comparative results.
The overall message of the website was to establish the numbers. “The risks are often overestimated,” Fischbeck asserted. “People think that the risks of dying in the next year are much higher than they really are.” Statistics on the Kentucky Post website support his claim: 19-year-olds, for example, only have a one- or two-in-a-thousand chance of dying this year depending on their race, which is a small probability. With these newfound, reassuring numbers in mind, though, Fischbeck remains aware of the fact that people might take these numbers for granted and engage in more risky behavior. However, Fischbeck found economic trends in health care around the world. Americans spend much more money on health care every year for people over 80 years old than Europeans, even though the chance of dying is lower for these Americans than Europeans.
Next up on Fischbeck’s plans is the inclusion of several other countries into the DeathRiskRankings database. Canada (which could not provide data at the time of the website’s creation), Japan, and South Korea were definite interests. “[Asian countries] have much better life expectancies,” Fischbeck said, which would make for useful comparisons. Other countries, like Qatar, have already given data to be used in the future.
Other plans include creating other risk-assessment websites. One of these websites will include the odds of going to the hospital in the next year, and it will give the reasons for going, how long the stay will be, and what the cost will be. Fischbeck cited better decision-making as what he hopes visitors get out of his website. “People are now more informed, and I hope are making better decisions because of it,” he said.