Seniors Aren’t Learning to Choose Better Prescription Insurance Plans
The Medicare Part D program, which gives seniors access to privately run prescription drug plans, was supposed to harness the power of the market to keep costs down. But a series of studies by Yale SOM economist Jason Abaluck and his co-authors have found that seniors often aren’t picking the plans that offer the best value.
By Roberta Kwok
The Medicare Part D program, launched by the U.S. in 2006, gave seniors the opportunity to obtain prescription drug coverage by choosing from a wide array of subsidized private insurance plans. Because private insurers were running the plans instead of the government, the reasoning went, the drive for profit would motivate those companies to increase efficiency. In a competitive market, those savings would be passed down to consumers.
However, there was a hitch: seniors often didn’t select insurance plans that offered the best value. As a result, many people paid more for drug-related expenses than necessary, reducing the incentive for companies to provide high-quality coverage at lower costs.
“If people aren’t choosing plans that are cheaper for them, then it’s unclear whether you will get this market discipline,” says Jason Abaluck, an assistant professor of economics at the Yale School of Management who studied this trend.
Now, a new study by Abaluck’s team published in American Economic Review suggests that the problem has persisted: Beneficiaries aren’t learning to make better choices. And changes in features such as premiums mean that patients are missing out on even more savings than they did at the program’s start.
Matters are “definitely not getting better over time,” Abaluck says. “If anything, they seem to be getting worse.”
In an earlier study, Abaluck’s team found that seniors often pay more attention to premiums than out-of-pocket expenses when selecting Medicare Part D prescription insurance plans. Beneficiaries could have spent about 30% less if they had chosen more suitable plans.
But that study included data only from 2006 and did not explore whether the situation improved. Perhaps seniors would learn to choose more wisely, or those who joined the program later would have more resources to help them sift through the options. In addition, insurance plans might change so that beneficiaries did not lose as much money due to poor choices.
To investigate that question, Abaluck and Jonathan Gruber, a professor of economics at the Massachusetts Institute of Technology, examined a larger data set from the Centers for Medicare & Medicaid Services. The team analyzed claims from 1.76 million beneficiaries from 2006 to 2009 and simulated how their Medicare Part D expenses would have differed if they had selected other plans.
As in the previous study, the researchers found that seniors were paying more than necessary. In 2006, beneficiaries could have saved an average of 19-33% of their Medicare Part D expenses if they had selected a plan that best matched their medical circumstances. Only 11% of patients chose the best plan in 2006; this number fell to 8-9% in 2007-08 and 2% in 2009.
The results were similar when the team analyzed overall “welfare,” a measure of the plan’s value that accounted for factors such as quality ratings and level of risk protection. The average annual amount of lost welfare increased from $161 to $251 over the course of the study.
Abaluck and Gruber then created a model to explore which factors might contribute to this trend. Inertia played a large role; about 90% of people simply stuck with the same plan they chose the previous year. But even seniors who changed their plans didn’t pick the most suitable options. The problem “isn’t just that they’re not bothering to switch plans,” Abaluck says. When they do switch, “they’re not finding the best ones.”
In addition, people who had been enrolled in the Part D program longer did not choose any better than those with less experience. Nor did seniors who entered the market in 2009—when more resources to help customers were available—fare better than those who entered in 2006. And because premiums for suboptimal plans that patients chose tended to increase more than those for optimal plans, the patients’ lost savings also rose.
Part of the reason for these poor decisions is that people are putting too much weight on factors that won’t make much difference to them, Abaluck says. For instance, in 2006, seniors were willing to pay $350 more for “donut hole” coverage—which covers a gap in drug costs—even if this feature did not turn out to cut their out-of-pocket expenses.
The results raise the concern that patients’ lack of discernment is reducing the pressure on insurance companies to offer better value. If seniors are not paying close attention to their choices, private insurers might be able to get away with charging higher premiums. Even if companies still strove to become more efficient, the savings might simply allow them to reap more profits. “The benefits wouldn’t get passed on to consumers,” Abaluck says.
The issue is too complex for Abaluck to say definitively whether seniors would have been better off with a government-administered plan. But the results do suggest that the current model could be improved.
For instance, people could use a calculator on the Medicare Part D website that estimates how drug expenses will differ on various plans. Abaluck and his team also found in another study that reducing the number of available plans in each area could help. This is not because seniors choose more carefully when they have fewer choices, but rather because the benefit managers assembling the plans are less likely to include options that are unsuitable for a lot of people if the number is constrained. When patients then pick from this limited set of plans, “there’s just less scope for going wrong,” he says. (Abaluck notes that if the number of plans is determined in some other way, small choice sets may not always include better plans.)
Abaluck acknowledges that making the right decision isn’t easy. “It’s just an immensely complicated problem that is almost impossible to solve unless you have a sophisticated computer algorithm,” he says.