This article examines the relationship between 1996 health plan enrollment
and both HEDIS-based plan performance ratings and individual HEDIS measures
. Data were obtained from a large firm that collected, aggregated, and diss
eminated plan performance ratings to its employees. Plan market share regre
ssions are estimated controlling for out-of-pocket price and model type in
addition to the plan ratings and HEDIS measures. The results suggest that e
mployees did not respond strongly to the provided ratings. There are severa
l potential explanations for the lack of response, including difficulty und
erstanding the ratings and never having seen them. In addition, employees m
ay base their plan choices on information that is obtained from their own p
ast experience, friends, family, and colleagues. The pattern of results sug
gests that such information is important. Counterintuitive signs most likel
y reflect an inverse correlation between some HEDIS ratings (or measures) a
nd attributes employees observe informally.