Factors associated with changes in satisfaction with care.

Objective: Satisfaction with care is an important outcome for evaluating the effectiveness of medical care. Many factors can influence satisfaction, including disease state, healthcare utilization, and health-status changes. However, few studies have investigated the association between these factors and changes in satisfaction.

Design: This study examined the influence of personal characteristics, type of health plan, disease states, and healthcare utilization on changes in satisfaction with care in a prospective cohort over a 12-month period through two surveys, baseline and follow-up.

Participants: Enrollees in one of three different commercial health plans: point-of-service product, an unrestricted fee-for-service product, and a preferred-provider organization product.

Measurements and main results: Two multivariate logistic regression models were constructed. The first model evaluated factors that predicted increased satisfaction with care between the two surveys. Compared with respondents who reported no change in health status, both those with improved health status (odds ratio [OR], 1.29, 95% confidence interval [CI95], 1.03-1.61) and those with declines in health (OR, 1.29, CI95, 1.03-1.61) were significantly more likely to report an increase in satisfaction with care. Those with a history of hospitalization were also more likely to report an increase in satisfaction with care (OR, 1.27, CI95, 1.01-1.59). The second multivariate logit model evaluated factors that predicted decreases in satisfaction with care from the baseline survey. Those with reported declines in health status were more likely to report decreases in satisfaction with medical care (OR, 1.43, CI95, 1.13-1.79). Neither age, gender, race, type of health plan, disease state, nor doctor's office visits were related to observed changes in satisfaction with medical care.

Conclusion: Changes in satisfaction with care appear to be related to changes in health status. However, the relation between these two attributes is not intuitively apparent.