Part of the reason why, he said, can be boiled down to simple numbers. Prevention programs must include huge numbers of people to be effective, making them expensive, and there is no guarantee they will work. As an example, Cutler pointed to smoking cessation.
“If you stop smoking and then you don’t have a heart attack, you save the money you’d spend on treating the heart attack,” he said. “But the argument has been that, because very few people will manage to stop smoking, you have to intervene with a lot of people, so the number you need to treat is large. And then the second reason is that maybe you don’t die of a heart attack, but you’re going to die of something, and that will still be expensive … so it’s largely a wash.”
Cutler’s study upends those arguments, saying that even relatively modest investments in preventive care can produce significant savings.
“When we looked at the trend in per-capita spending by the elderly, in 2005 is where we began to see the increases slowing,” Cutler said. “And a large part of that is due to cardiovascular health, because what used to disable people were heart disease and strokes, and those have declined immensely … so about half of the decline we saw was related to cardiovascular issues.”
By 2012, Cutler said, the slowdown in health care spending was beginning to add up.
“It was almost $ 2,900 a year, per person — that’s a lot of money,” he said. “Toward the end of Obama’s first term, you may recall that there was a great deal of talk about debt reduction. But when a deal didn’t get made, the issue just faded away. One reason why is because Medicare costs came in less rapidly than we thought … and this study shows one reason why.”
Deciphering what was behind the slowdown in health care spending was not easy.
“There were three main technical challenges in this paper,” Cutler said. “The hardest one, where we spent the most time, was in trying to de-compose the spending by disease. That turns out to be very difficult because if someone goes to the doctor for a cardiovascular condition, but they also have a history of mental illness, how do you separate the spending on those items?”
The solution, Cutler said, came when he and colleagues compared a random sample of billing records of people with similar diagnoses.
“We did an analysis of everyone who has X, we compared them to people who look just like them, but they don’t have X,” Cutler said. “And then we can ask how much more did people with X spend?”
The next challenge, Cutler said, came in teasing apart how much of the spending slowdown was due to fewer cases of cardiovascular disease and how much to each case simply costing less. What researchers found was that there were both fewer first-time illnesses and fewer cases where a patient had a problem, like a heart attack, and later experienced additional problems.
Finally, Cutler said, the team faced the challenge of gauging the impact of medications on spending. To do that, he said, they created a combined measure of how various drugs lower the risk for cardiovascular disease. When the team compared the predicted drop in disease to the actual drop, it found about half of the dip could be attributed to medications.
While the study found that significant cost savings came from improvements in cardiovascular health, Cutler said there is still room for improvement.