Creating a Legacy of Healthy Decisions

Laura Roser’s Q&A with Emily Oster, Professor, Brown University


LR: Tell us about your work and where your passion lies related to your research.

EO: My work focuses on health economics and, in particular, on understanding when people make choices that do not seem in their best interest with respect to their health. Much of my earlier work focused on these issues in the developing world. I studied, for example, why there was not much change in sexual behavior in Africa in response to the HIV epidemic. My more recent work takes up similar question in the USA. At the moment, I am engaged in a series of projects which study dietary choices and what types of information or other changes can prompt improvements in diet.

The thing that drives me in much of my work is the goal of understanding what data is telling us about how people behave, about what policies are the best, or simply about how the world works. There is no better moment in my research when I realize I have learned something new from data – something no one else knows – and I’ll be able to share it.

This goal of understanding data also drives the second arm of my research , which focuses on improving the use of statistical tools to learn from data. Data can be wonderful, but used incorrectly it can also pull us far astray. Developing better, easy to use, tools to improve our understanding of statistical relationships can help us draw the right conclusions – and not the wrong ones – from the data.

LR: What influenced your research toward economics and health?

EO: Prior to becoming an economist I had designs on being a doctor. I think this is still evident in much of my work. I like the tools and the central ideas in economics, but topic-wise my heart is really in the health space. One of the things that is really special about my job is that I can combine these two.

LR: How should our readers think about health related to the legacy they are leaving?

EO: In the United States, in particular, there is enormous inequality in health outcomes across socioeconomic groups. Several years ago I did a project on infant mortality in the US and Europe. It is well known that the USA has extremely high infant mortality relative to Europe. By some metrics it is twice or three times as high as the most successful European countries. Our paper delved into the reasons for this and one of the most striking findings was that well-off women in the US (those with a college education or more) had infant mortality rates virtually indistinguishable from their counterparts in Europe. The difference was that less well educated mothers in the US has dramatically higher infant mortality than similar cohorts in Europe. This inequality extends to other parts of the life course. Poorer individuals in the US are more likely to die at all ages, are more likely to be overweight and obese, have more heart disease, strokes and cancer. They are also more likely to have conditions like opioid dependence. Knowing how to use a legacy to impact these problems is difficult. We do not have many good evidence-based solutions for these issues. It may be self-serving, but funding research into solutions is an approach which could help in the even slightly longer run. We have little or no idea, for example, how to produce sustained improvements in diet, which would have positive impacts on all kinds of health outcomes. Learning more about what works would be a first step.

Related: Why Write Your Legacy?

LR: What kind of legacy would you like to leave behind for your family, community, and world?

EO: I hope my best legacy will be my children. But from my work I hope to ultimately deepen our understanding of what works to change behavior and what doesn’t.

LR: What is one principle from your research that our readers could use for practical decision making for future health care decisions?

EO: Take this example. Let’s say there is a disease that affects 1 in 10,000 people. There is a test for the disease. Everyone who has the disease receives a positive test result. Most people who do not have the disease get a negative
test result. 2% of healthy people will get a positive test result. These are “false positives.” Still, this is a great test.

Now imagine you get a positive test result. What is the chance you have the disease? Most people would say quite high – maybe close to 100%, 98%, maybe 99%. In fact, this is also what many doctors answer when faced with this question. The answer is actually about 0.5% – that is, one half of one percent, or 1 in 200.

Once you know this, it is easy to see why it’s true. Since the disease is rare, most people do not have it. 2% of those who do not have it – so 200 of every 10,000 – will be a false positive. But within that 10,000 people there is only ONE
true positive. So even if your test result is positive, it is still very likely you are healthy.

This rule is important because of the tendency most people have to want to ACT on health information, especially bad news. Over-reacting to a test of this type – and this is a common feature for many kinds of tests – can create problems where there are none. Taking a step back, repeating the test if necessary, trying another diagnostic test to be sure – these are all crucially important things in making sure you end up making the right decisions.