Access to Alcohol Treatment and Selective Enrollment in Health Care Plans (8/01)

Background: Behavioral health care benefits are less comprehensive than medical care benefits in most private insurance plans. In response, federal and state parity laws require private insurance plans to offer equal coverage for behavioral health care and medical care. However, most of these parity laws exclude substance abuse care because of uncertainty about the costs and because reliable new data is lacking. There is continuing resistance to adding this coverage because of uncertainty about the costs and because reliable new data is lacking. One concern by employers is that offering relatively generous substance abuse benefits will attract an employee population with high treatment costs, high medical costs, and low productivity.

There is a presumption in health services literature that people who know they are likely to need specific services are more likely to enroll in insurance plans that afford easier access to those services. This belief provides a rationale for health plans to restrict access to treatment services for fear of attracting potentially high-cost enrollees, a phenomenon known as adverse selection. The validity of this presumption is rarely challenged in health economics literature for most diseases. However, in the case of alcohol dependence, which is usually accompanied by a failure to recognize that treatment is needed (denial), one can question if individuals with alcohol dependence are more likely to choose particular enrollment options because they make access to alcoholism treatment easier. If proven false, it can weaken arguments against making alcoholism treatment more accessible. If proven true, it would raise questions about how such foresight coexists with denial.

Study goals: This study investigates the relationship between access to alcoholism treatment and health plan enrollment decisions.

Methods: This study uses longitudinal data on firms contracted with a large managed behavioral health organization to test for adverse selection of alcohol-dependent enrollees into plans with generous treatment benefits. Because conclusions about adverse selection are confounded with demand responses to cost-sharing (moral hazard), two different approaches will be taken:

1. A series of "natural experiments" will be identified. These arise when employers implement substantial changes in substance abuse treatment benefit offered to employees. The degree of adverse selection will then be measured by assessing whether the use of alcohol treatment services by new plan enrollees differs systematically from that of old plan enrollees who joined their employers prior to the design change.

2. The relationship between the generosity of alcohol treatment benefits and the length of time that treatment users remain enrolled in their health plan will be examined. Disenrollment rates of users and nonusers of alcohol treatment services, and high- and low-cost users by level of benefit generosity will be compared. Compared to natural experiments, this approach provides increased statistical power with which to detect adverse selection.

Current status: Ongoing.