Principal investigator: Eduarda Miller de Figueiredo
Authors: Jason Abaluck and Jonathan Gruber
Location of the Intervention: United States
Sample Size: 477.393 individuals
Sector: Health Insurance
Primary Variable of Interest:
Type of Intervention: Health insurance
Methodology: Logit
Summary
O Medicare It is a program that provides universal health insurance coverage for people over 65 and for those in the disability insurance program. In 2003, through a law modernizing this program, benefits for prescription drugs were added – the Part dThis article sought to investigate the choices of older adults based on available plans after modernization. Using the conditional logit model and discrete choice models, the authors found that individual choices are consistent with maximizing behavior.
- Policy Problem
The Modernization Law of Medicare[1]The reform, which took place in 2003, represented a more significant expansion of U.S. public insurance programs by adding the benefit of prescription drugs. Part d to the program MedicareFor this expansion, several private insurance companies were used to provide this new product for public insurance.
O Medicare It is a program that provides universal health insurance coverage for people over 65 and for those in the disability insurance program. Until its expansion, the program covered most medical needs but excluded coverage for prescription drugs. Thus, beneficiaries of the program... Medicare They spent an average of $2.500 each on prescription drugs in 2003, which is more than double what the average American spent on all healthcare in 1965 (the year the National Health Surveillance Agency was created). Medicare).
This article seeks to investigate the choices of older adults regarding the newly created program. Part d in 2006.
- Implementation and Evaluation Context
In 2003, the U.S. government and Congress agreed on a prescription drug benefits package at a projected cost to the federal government of $40 billion per year for the first ten years. The innovation of Part d This is due to the fact that it is provided by private insurance companies under contract with the government.
According to Duggan, Healy, and Morton (2008), standard economic theory would suggest that the beneficial characteristic of the plan is precisely the permission for individuals to choose from a wide variety of plans that meet their needs, rather than restricting them to a limited set of choices made by the government. However, Iyengar and Kamenica (2006) emphasize that not only the decision to participate in a market, but also the very nature of the choice is affected by the size of the set of options.
Heiss, McFadden and Winter (2006) when studying the Part dThey assessed whether the intentions to enroll in the plan were "rational," which led to the discovery that for most potential enrollees, the decision to enroll or not appears to be made rationally.
- Policy/Program Details
Beneficiaries could choose between three types of private insurance plans that cover their medication expenses, namely:
- Medicare Prescription Drug Plans (PDP): offers only the benefits of prescription drugs.
- Medicare Advantage (MA): provide all the benefits of Medicareincluding prescription medications.
- Standard Plan of Part dBeneficiaries could retain their current employer plan, provided the coverage was credible or at least as generous as the standard plan. Part d, for which they would receive a subsidy from the government.
According to Part dBeneficiaries are entitled to basic coverage for prescribed medications according to the following structure:
- None of the first $250 in medication costs each year;
- 75% of the costs for the next $2.250 in medication expenses (up to $2.500);
- 0% of the costs for the next $3.600 in medication expenses (up to a total of $5.100, the "donut hole");
- 95% of costs exceeding $5.100 are medication expenses.
Enrollment in Part D plans was voluntary for qualified citizens of MedicareHowever, low-income seniors who received coverage for prescribed medications through state programs... Medicaid Those who were automatically enrolled – the “doubly eligible” – could have copayments of just $1 for generics and $3 for brand-name drugs below the poverty line, and $2 for generics and $5 for brand-name drugs above the poverty line.
There was enormous interest from insurance companies in participating in the program; more than 3 plans were being offered to potential enrollees. Part dIn June 2006, there were 10,4 million people enrolled in the PDP plan, 5,5 million in the MA plan, and approximately 6 million "doubly eligible".
- Assessment Method
The authors used a sample of prescription drug records from the Wolters Kluwer (WK) Company for 1,53 million older adults who:
- They had a claim from Part d with any type of coverage;
- They are not insured by the employer, doubly eligible, or eligible for subsidies;
- Having claims for only one region of the country;
- Avoid complaints regarding missing payment information;
- They are part of the sample of pharmacies with consistent reports;
- It contains data from 2005 and 2006.
Information regarding the availability of plans was also used. Part d Based on four files provided by the CMS: plan information, beneficiary cost, form, and geographic location.
After thoroughly cleaning the sample, focusing only on PDP plans and excluding individuals with fewer than 500 observations in their status, the final sample consists of 477.393 individuals.
To analyze the plan using the conditional logit model, the authors utilize several discrete choice models, since these allow for controlling additional plan characteristics, provide a more precise understanding of how preferences combine with the characteristics of the choice set, and allow for quantifying the welfare consequences of the choices.
- Main results
Model 1, which includes only the premium, direct costs incurred, variation in direct costs, and quality variables, shows that a $100 increase in premiums leads to a 32% reduction in the probability of a given plan being chosen. Model 2, which adds additional covariates to control for deductibles, "donut hole" coverage, average cost sharing, formulary coverage, and plan quality, shows that the coefficient for premiums increased, suggesting that it was initially biased downwards due to the bias of the omitted variable. That is, the coefficient in the variance term falls even further when a control for the number of the 100 most popular medications included in the plan's formulary was added.
According to the authors, one explanation for the result of model 2 is that, although individuals prefer plans that cover more medications, they lack sufficient foresight to choose plans that cover medications they may need in the future but are not currently taking.
In models 3 and 4, binary variables for brand and brand status were added. The premium coefficient decreased when the authors included fictitious brands, but the premium effects remain large: a $100 increase in annual premiums leads to a 50% reduction in the probability of a plan being chosen. The coefficients for plan features are very large across all specifications. Model 4, which has the lowest plan features, suggests that individuals are willing to pay over $300 for full coverage of the "donut hole," $50 for generic coverage of the "donut hole," and $12 for each of the top 100 medications listed on the formulary.
In the end, the authors conclude that the distribution of health plan coverage would be quite different if there were no inconsistencies in choices. It was estimated that the share with some coverage of the "donut hole" would fall by 40% if these inconsistencies were corrected.
- Lessons in Public Policy
Although individual choices are consistent with maximizing behavior, such as preferring lower out-of-pocket expenses and higher quality, they are inconsistent with the standard model in three respects: individuals underestimate out-of-pocket expenses relative to premiums, overestimate plan features, and do not fully appreciate the risk-reducing aspects of plans for themselves.
References
Duggan, M.; Healy, P.; Morton, F. S. (2008). Providing prescription drug coverage to the elderly: America's experiment with Medicare Part D. Journal of Economic perspectives, 22
Heiss, F.; McFadden, D.; Winter, J. (2009). Regulation of private health insurance markets: lessons from enrollment, plan type choice, and adverse selection in Medicare Part D (No. w15392). National Bureau of Economic Research.
Iyengar, SS; Kamenica, E. (2006). Choice overload and simplicity seeking. University of Chicago Graduate School of Business Working Paper, 87, 1-27.
[1] Medicare Modernization Act of 2003.
