CDC looks to localize its health data even more

Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.

The Centers for Disease Control and Prevention for several years has maintained a partnership with a couple of foundations. It brings geographically oriented health data to counties, zip codes and other census designated spots. Now the initiative, called Places, adds data for even smaller zones that had been unavailable. With more about the program, the director of the CDC’s National Center for Chronic Disease Prevention and Health Promotion. Dr. Karen Hacker, joined Federal Drive with Tom Temin.

Interview transcription:

Tom Temin: Dr. Hacker, good to have you on.

Dr. Karen Hacker: Thank you so much.

Tom Temin: Tell us more about Places, what does it purport to do and what kinds of information does it deliver to? And to who does it deliver it to in these locales?

Dr. Karen Hacker: So the data for Places comes from a very large surveillance system known as a Behavioral Risk Factor Surveillance System. This is a survey that has been done by phone for many, many years. And it’s really how we look at chronic diseases in our country, and also health behaviors — have you gone to the doctor in the last year, things like that. But that isn’t very helpful if you’re at the local level because you’re looking at the state, you may be looking at the nation, but you don’t really know what’s going on in your own community. So now, we have used a methodology that allows us to take that data and do something called small area analysis so that we can actually look at the data, as you said, at this very granular level. And for those of us who are in government, both at the federal level, but I’d also say as a former county employee, having that information is really unique. It’s something we did not have access to before.

Tom Temin: And how is this information generated? Who makes the phone calls to whom and where does the data get gathered together?

Dr. Karen Hacker: The data for the Behavioral Risk Factor Surveillance System is done on a state by state basis. And in some cases, the state has a contractor that they are working with, maybe even a university. In other cases, the state may be doing it themselves. And that data, there’s a variety of modules that are used for that, there’s some consistent ones across all states. And then there’s even some that may might want to look at a particular issue beyond the generic sort of grouping that they have.

Tom Temin: So the data they gather is information about individuals then that is rolled up to the national level at CDC?

Dr. Karen Hacker: Absolutely, it is all anonymous data. But questions like when were you at the doctor last? Have you ever been diagnosed with diabetes? Do you have high blood pressure? Do you take medication for high blood pressure? And on and on. And again, those are data that are very hard to access because most of that information is in the healthcare system. And we in public health don’t have access to that, as you know.

Tom Temin: Right, the HIPAA laws and so forth. But you can get individuals to tell it then, back out their names and specific addresses and that kind of thing.

Dr. Karen Hacker: Correct. And I believe that when they take the data, you may even have ever gotten one of those phone calls, it’s completely anonymous from the very beginning.

Tom Temin: And are the phone calls directed at the aged or certain populations, or are they random, almost like a census survey?

Dr. Karen Hacker: Yes, it’s totally random. It is one of the things we used to call as random digit dial surveys. And the idea being that you were using a much more scientific method of sampling when you do that randomization.

Tom Temin: And just a question of process. Is it harder to get the calls completed nowadays, since nobody answers their phone, for the most part, thinking it’s somebody selling something?

Dr. Karen Hacker: It’s interesting that you say that, because both that but also the movement to cell phones and going away from landlines. So it is a little bit more difficult. And we’re really thinking about data monetization, and how best to validate that data on an ongoing basis. But I think it also depends a little bit on who is trying to get the data, and whether or not the individual is willing to pick up the phone and spend the time.

Tom Temin: That would be quite a project to change the data gathering basis. I guess there’s other methodologies, but it would be quite an effort, I imagine.

Dr. Karen Hacker: Well, and also we’ve had this data for a long time, so we can compare it to prior years and look at trends, which is very important for us. The challenges, though, as people use new technology is how do you get access to this information and what’s the best way to collect it?

Tom Temin: And how do the locales use the information? Is it local health officials, local hospital administrators, all of the above?

Dr. Karen Hacker: I think at this point, it’s pretty new. The 500 cities, which was at that level previously, in many cases certainly was used by health department directors. I think in some cases, it might have been used by other local officials as they thought about where they wanted to put their services or where they might expand services. And the other thing is, you can take this data and you can use other datasets and combine them together. So if you wanted to look at, for example, environmental issues, you could combine that with now chronic disease and you can sort of see them together. I think that gives you a new window into how these are connected to one another.

Tom Temin: Now you are enhancing the data. Tell us what is new here, what’s going on with the Places program?

Dr. Karen Hacker: What’s new is the level of data that’s now available across the entire country. So previously, we did have data available at the 500 largest cities. But as a health department director, for example, my own had had a county, that meant the rest of the county I didn’t have information on, and now you can get data literally at the neighborhood level.

Tom Temin: And how is that possible? That’s a lot of phone calls.

Dr. Karen Hacker: No, we took the data from the Behavioral Risk Factor Surveillance System, and then we applied an analytic strategy called small area analysis. And it’s just the area is getting smaller and smaller, and so that’s why it’s available now. And as you know, where you live really matters.

Tom Temin: Sure. But I guess I’m wondering how you have enough data density, for example, to project what the health trends are, say, in a rural county, where there might be one person per square mile versus Brooklyn or something where you have 10,000 people per square mile, or maybe it’s 100,000, a million?

Dr. Karen Hacker: I don’t have all of the information on these the statistical methods. If you’re really interested, there are papers written on this. But you’re absolutely correct. The challenge is how do you use this data and the methodology to make it valid, even at a very small level. So one of the things that you’ll notice, if you look at this, that there’s something called confidence intervals, which speak to how confident you are in the results. And yes, they are pretty wide when you look at very small areas.

Tom Temin: But even for a local official or a local hospital that’s serving in a low density population area, at the least the data could give them ideas on trends and things to look out for, I imagine.

Dr. Karen Hacker: Oh, absolutely. And again, going from nothing to something is an enormous difference in thinking about how you apply those resources.

Tom Temin: And does this methodology or does this data gathering extend to say, opioid addiction and types of behaviors like that are really public health issues, but not diseases per se, like heart disease, or liver disease or something?

Dr. Karen Hacker: Well, currently, the major modules in the Behavioral Risk Factor Surveillance System are really focused on chronic disease, and opioids is not necessarily a part of that. There are some modules available, but not every state decides to use them. So there are 27 variables that are included in places at this point, it’s going to be the ones that you would probably expect, questions on smoking, obesity, diabetes, hypertension, heart disease, those kinds of things are going to be what you’re going to see, and physical activity as well.

Tom Temin: And how do people get at the data locally or nationally?

Dr. Karen Hacker: Go to cdc.gov.

Tom Temin: I know you’d say that.

Dr. Karen Hacker: Yes. And you can start to drill into it. It’s interactive, so you have an ability to sort of magnify smaller areas or to highlight certain areas. I haven’t personally used it myself. It’s pretty user friendly. And again, you can look from one place to another and see the differences and variation, even across what might seem like a relatively small jurisdiction.

Tom Temin: And a final question. Is the methodology that’s been in place for so many years, is it agile enough, fast enough to possibly be applied in a pandemic situation that we now feel could happen at any time?

Dr. Karen Hacker: Right. Well actually, one of the challenges is that these surveys are really done only on an annual basis. And so one could argue that data is a little old. However, most of the types of things we’re talking about do not change that rapidly. So it is in use, for example, the CDC actually has a COVID tracker, and this data is also being used there. And that COVID data it’s a lot more real time then the data that we’re talking about.

Tom Temin: Dr. Karen Hacker is director of the National Center for Chronic Disease Prevention and Health Promotion at the CDC. Thanks so much for joining me.

Dr. Karen Hacker: You’re welcome. It’s a pleasure.

Related Stories

Comments

Sign up for breaking news alerts