Using Data Well: NewPublicHealth Q&A with John Lumpkin
At the Health Data Initiative III: The Health Datapalooza this summer, which looked at the role health data plays in transforming health and health care, the Robert Wood Johnson Foundation (RWJF) announced the three winners of an app challenge calling for developers to create applications that would allow consumers to easily access and make use of comparative information about the quality of care provided in various regions of the country. The data behind the apps came from Aligning Forces for Quality, the Foundation’s initiative to improve the quality of health care in 16 targeted communities.
John Lumpkin, MD, MPH, senior vice president and director of the Health Care Group at RWJF, presented the awards to the top three winners of the challenge. Recently, NewPublicHealth spoke with Dr. Lumpkin about the explosion of health care data, and how to find the strongest data and use it well.
NewPublicHealth: Quantities of health data are increasing at a speedy clip. What kind of challenge does that pose in order to make sense of the numbers?
Dr. Lumpkin: In a single day there are over 300 million photos that are being posted on Facebook, and 3.2 billion likes and comments. We have 115,000 people participating in web sites such as Patients Like Me and about 9 million tweets each day. So we have all this data, but how do we access it in ways that make it meaningful? Kerr White, MD, the chairman for the U.S. National Committee on Vital and Health Statistics once said: “With the advent of new technology, data can be collected in any format, aggregated by the computer and arrayed in any desired output…untouched by human thought.” And so our challenge is to present it in a way that we actually have it touched by human thought.
Some good examples of this would include the County Health Rankings and the Aligning Forces for Quality Initiative where we have coalitions of consumers, purchasers and providers who, through public reporting of quality and price information and consumer engagement, we encourage to adopt quality improvement methods and use the data in ways that move that process forward.
NPH: Does everyone mean the same thing when they talk about “evidence-based” data?
Dr. Lumpkin: A key question is whether data is analyzed in such a way that’s it’s valid to make assumptions. Sometimes data can be pulled together in ways that disproportionally heightens minorities’ involvement with a problem. Sometimes expanding the frame in which a problem is analyzed identifies the need for further analysis. When I was the State Health Director in Illinois, I looked at some of the data used to compare two counties, Gallatin and Cook counties. Both counties had the same rates of death due to violence. Cook County has a high rate of minorities; Gallatin County statistically has no minorities. But what they both had in common was the rate of poverty. So while there may be a correlation between race and number of violent deaths, there is a much stronger correlation between violence and poverty level, and this is the one where the evidence is clearer.
So, yes, I think that there are some dangers in how data is used, and that’s a reason why it’s important that statisticians get involved and that there is also some external peer review proof to say, yes, this is a valid assessment.
NPH: What else concerns you about the current accumulation of data?
Dr. Lumpkin: I think that the real risk in this big data stuff is some of the basic tenets of statistics. Correlation is not causation. So you can do a correlation that says people who walk on the right hand side of the street are more likely to have skin cancer. That doesn’t say that if you walk on the right side of the street you’re going to get skin cancer. So there needs to be some process of making sure there’s enough evidence that can stand up to the test.
>>Read more on moving data into action.