Pay Attention To The Data: Q&A With Austin Frakt

posted by Alicia Caramenico

on February 20, 2019

In March, health care economist Dr. Austin Frakt will speak at AHIP’s National Health Policy Conference in Washington, D.C. We caught up with the director of the Partnered Evidence-Based Policy Resource Center at the V.A. Boston Healthcare System ahead of the event to learn about the importance of evaluating health system change.

Your session focuses on Improving Health Care Through Evidence-Based Guidance. Why evidence from data-driven research is so important?

Dr. Austin Frakt: We want to get away from anecdotes and things that aren’t generalizable. If we want to know truth about the effectiveness of something, we can’t necessarily rely on an exceedingly small sample. Just talking to peer experts may not be accurate.

How can we move to an evidence-based health system?

Frakt: There’re a few things you need. A very important, necessary thing is financing—we spend a minuscule fraction of the total health budget on generating evidence and measuring what we’re doing. So, I would advocate for a larger budget for research. I would advocate for when programs are implemented that are new, they are always accompanied by funding and plans for rigorous evaluations.

We also need data—so you have to collect data and make it available. You have to use good methods, and you have to pay attention to the evidence. You have to actually do something with the data. If it shows something is not working well, you need to be ready to stop doing that thing. If it shows something is working very well, you might be ready to expand it. If it shows something is working but could be better in certain ways, that suggests other actions.

We do all those things badly. We don’t fund enough research. We don’t do enough research. We don’t pay attention to research when we have it. And we let things that aren’t working continue and we don’t expand things that do work for a variety of reasons.

What can insurance providers do better when it comes to evidence?

Frakt: On the generation side, the biggest thing insurance providers can do is help make their data available for research. There’s a lot of issues with that—one has to pay attention to privacy and security. But insurers that participate in some way in making their data available for research are doing the right thing.

Then there’s heeding evidence, in terms of clinical activities. Making benefit packages align with evidence on the value of different treatments is a concrete step. This is a variation of value-based insurance design—where something is demonstrated to be of very high value, in terms of the health effects it delivers or its cost, then the cost sharing of that is made much lower because we want to increase affordability for those things.

You wrote a piece last week on hospital consolidation. What do we know about how mergers affect the price of hospital care? What about the effect on quality?

Frakt: We know a lot—mergers increase price. There’s also studies that show quality does not improve when hospitals consolidate, and some studies show that quality can get worse. Increased consolidation is not helpful on price or quality. To put it another way, competition is very good in terms of driving prices lower and improving quality.

You’ve also written extensively about the high cost of pharmaceutical innovation. Can you talk about how to balance prices and access to drugs?

Frakt: People and organizations that are looking to invest will invest more in areas where they’re going to make more money, either through higher prices or broader market. Everybody does that—why do we invest for retirement in stocks or bonds? We’re all looking for a high return—that’s normal and we shouldn’t be ashamed of that.

So, you’ll get more money flowing into pharmaceutical research and development if the payout is bigger. Anything that reduces the payout is in that sense a threat to future innovation. Now we can’t afford an infinite amount of innovation. There’s a tradeoff. The big question is how you trade off the revenue pharmaceutical manufacturers receive and the rate of innovation. And that’s a big struggle and it comes down to values. How much future innovation do you want? I don’t know how to answer that question. I want as much as I can afford, I suppose.

The organization I think is doing good things in this area is the Institute for Clinical and Economic Review. They have a very good process, an open process for assessing clinical and economic value of new technologies and drugs. They have stakeholder engagement—it’s all public. They take feedback and refine what they do. And they have no authority to make any changes. They’re not a government agency, they just provide information to the market as best they can.

How do you envision drug prices evolving in an ideal situation?

Frakt: It’s very complicated. We have different drug pricing problems for different kinds of drugs. Small-molecule drugs are different than biologics, biosimilars are different than generics, the generic market is different than the brand market. Different classes of drugs are different. Vaccines you want everybody to get. Antibiotics, in an ideal world, nobody needs them. We kind of hope there’s no market for antibiotics but on the other hand we want and desperately need new antibiotics.

So, the way you would want reward innovation in vaccines or oncology would be different than for antibiotics. And that’s why we have a problem with not enough antibiotics innovation—we’re not paying attention to those nuances.

What are some of the big health industry issues your blog The Incidental Economist will tackle this year?

Frakt: I will be paying attention to the value of health care versus the value of other things that determine our health (social determinants of health). The big question of what affects health and by how much is a hard question and we don’t fully know the answer.

It would be good for health insurance providers, hospital systems, and ACOs to try to understand where our limited resources should be spent to maximize health. It might not be on the next drug; it might be on nutrition or housing or some other thing that we think isn’t health care but is actually very important to a person’s health. Health plan data could be very informative. A lot of what researchers can do is limited by the data they have. If there are insurance providers experimenting with benefit design that gets at social determinants in different ways, evaluating those would be very valuable.

" "