Atherly A; Academy for Health Services Research and Health Policy. Meeting.
Abstr Acad Health Serv Res Health Policy Meet. 2000; 17: UNKNOWN.
Presented by: Adam Atherly, Ph.D., Assistant Professor, Emory University, Department of Health Policy and Management, 1518 Clifton Rd. NE, Atlanta, GA 30322. Tel: 404-727-1175. E-mail: aatherl@sph.emory.edu.
Research Objective: To examine the effect of biased selection into Medicare supplemental insurance plans that include a prescription drug benefit on premiums for that benefit. Coverage for prescription drugs among Medicare beneficiaries is a very mixed bag. The vast majority of those with supplemental insurance from a third party (employer or the government) and those in managed care, have some degree of prescription drug coverage. But, for those without third party insurance and who either cannot join an HMO, or choose not to, voluntary coverage is highly unlikely. This then provides a conundrum: why do employer sponsored plans and HMOs overwhelmingly include some prescription drug benefit, while individually purchased plans do not? One possibility is that risk adjustment by insurers in the individual market increases the premium for high-risk (sick) individuals to the point they decide not to purchase drug coverage. Although this has been speculated to be true in the past, there are no published studies that quantify the effect of biased selection on the premiums for the prescription drug benefit on supplemental insurance plans.Method of Analysis: Supplemental insurance plan choice was modeled using a multinomial logit. Premiums were modeled using regression analysis. Premiums were regressed against individual and community characteristics. Using the estimated coefficients, premiums were predicted for supplemental insurance plans with and without a prescription drug benefit both for the groups who actually purchased the policies and for those who did not. Unobserved biased selection was corrected for using a Heckman Two-Step. Data was drawn from the 1995 Medicare Current Beneficiary Survey (MCBS).Principal Findings: Descriptive statistics suggest that those who buy individual supplemental insurance policies with a prescription drug benefit are slightly younger than those who do not include a drug benefit in their supplemental insurance (75.1 years versus 76.7), and have better self rated health and fewer ADL conditions. In the plan choice model, the purchase of a plan without a prescription drug benefit is positively associated with age, worse self-rated health, cancer and arthritis. Purchase of a plan with a prescription drug benefit was not found to be significantly associated with age or chronic illness. Predicted annual premiums for the drug group for a policy with a prescription drug benefit ($1,415) was less than predicted premiums for the non-drug group for a similar policy ($1,599); the result was statistically significant at the 0.01 level.Conclusions: The risk group formed in the individual supplemental insurance market with prescription drug benefits is, overall, healthier than the risk group formed by those without prescription drug benefits. The average predicted increase in premium associated with a drug benefit for the group that actually purchased a drug policy ($280) is significantly lower than the predicted increase for those who opted not to purchase drug coverage ($472).Implications for Policy, Delivery or Practice: In the individual supplemental insurance market, insurers are, in most states, free to experience rate policies. As a result, relatively unhealthy Medicare beneficiaries forgo prescription drug coverage due to high, though perhaps actuarially fair, premiums. This analysis suggests several possible market based methods for extending a prescription drug benefit through the existing Medicare structure, including the use of subsidies or the formation of joint risk pools.
Publication Types:
Keywords:
- Choice Behavior
- Health Maintenance Organizations
- Managed Care Programs
- Medicare
- Prescriptions, Drug
- Risk Adjustment
- economics
- hsrmtgs
Other ID:
UI: 102272459
From Meeting Abstracts