A major demonstration and evaluation project was undertaken to study the consequences of using incentive payments to change admission, discharge, and outcome patterns for Medicaid patients in nursing homes. Thirty-six proprietary, Medicaid-certified, skilled nursing homes in San Diego County with a combined Medicaid inpatient census of approximately 3,600 residents participated. Baseline data were collected from November 1980 through April 1981. Homes were then assigned to either a control group or a treatment group. During the next two years, an incentive reimbursement system was employed in paying for care in nursing homes in the treatment group. The results of the outcome-incentive component of the study revealed no statistically significant differences in patient outcomes, as measured by goal achievement, between treatment and control facilities. There was no evidence that goal-related reimbursement incentives improved patient outcomes. The results do suggest that goals related to clearly defined conditions are the ones most likely to improve patient outcomes. Incentive reimbursement may still prove effective if it is limited to certain clearly delineated conditions. If this is true, incentive reimbursement could be used to supplement rather than replace other approaches to quality assurance in nursing homes. (NB)