What follows are quotes of interest (to me) from sections 1-3 of Quality information and consumer health plan choices, by Nancy Dean Beaulieu (JHE, 2002).
Section 1, Introduction:
[T]here have been few published studies of the effects of health plan performance data on consumers’ health plan choices. The existing literature on this topic does not provide convincing evidence that health plan performance data affect consumers’ health plan choices […]. These findings are troublesome for practitioners and policymakers implementing managed competition. Embedded in the theory of managed competition is an assumption that health plan quality can be measured and that all else equal consumers will enroll in plans with higher measured quality. If these assumptions hold and if consumers have a choice of health plan, then market forces will generate incentives for health plans to improve the quality of care. […]
The analyses [in this paper] are based on data from a natural experiment in which a large employer began providing health plan performance data to its employees to facilitate their health plan selections. […]
Analysis of plan switching behavior suggests that the provision of quality information had a small, but significant effect on consumer plan choices. Employees were more likely to switch from plans with lower reported quality. Cross-sectional analyses of plan choice indicate that reported quality played a role in plan selection even after controlling for other health plan characteristics frequently associated with plan choice.
Section 2, Methodology:
[I hypothesize that] older consumers are more likely to have strong ties to specific physicians and therefore may be less responsive to quality information. Consumers that should be most responsive to the quality information are those that currently have little information about the relative quality of the different health plans and who are least likely to have established strong relationships with specific physicians. New employees meet these criteria.
Consumers with greater expected needs for health care services may be more or less responsive to quality information. They could be more responsive because they stand to benefit more from an improvement in the health care services delivered by a higher quality plan; on the other hand, they could be less responsive because of stronger ties to specific physicians or concerns about continued coverage for pre-existing conditions.
Section 3, Literature Reveiw
Scanlon et al. (1997) present a useful framework for understanding how these variables affect health plan choice. They identify a set of primary variables representing attributes of health plans that directly impact health plan choice. These primary variables include measures of price, quality, choice of provider, benefit design, coverage, and convenience. The set of secondary variables represents characteristics of the consumer that might amplify or attenuate the effects of some primary variables and, hence, indirectly affect health plan choice. Secondary variables include demographic and health status variables. One might think of the primary variables as arguments to the consumer’s utility function and the secondary variables proxying for the relative importance of the primary variables. […]
[In Chernew and Scanlon (1998)], the authors fail to find evidence confirming their hypothesis that newly-hired employees will be most responsive to the report card data. […]
In [Scanlon et al. (2002)] specifications that aggregate multiple health plan performance measures into an overall performance rating the authors find evidence that employees avoid plans with below-average ratings.
Chernew, M., Scanlon, D.P., 1998. Health plan report cards and insurance choice. Inquiry 35, 9–22.
Scanlon, D.P., Chernew, M., Lave, J., et al., 1997. Consumer health plan choice: current knowledge and future directions. Annual Review of Public Health 18, 507–528.
Scanlon, D.P., Chernew, M.,McLaughlin, C., Solon, G., 2001. The Impact of Health Plan Report Cards on Managed Care Enrollment. Unpublished manuscript.