Data Envelopment Analysis (DEA) was used to estimate the relative efficiencies of nursing homes in Maine for 2 years before the introduction of a prospective payment system and for 2 years afterward. A multiple regression was performed in which the DEA-estimated efficiency was the dependent variable and a number of facility characteristics and study year were the independent variables. Study findings indicate that nursing home efficiency scores fell after the introduction of prospective reimbursement; that is, generally speaking, more inputs were being used per unit of output in the period after the introduction of prospective reimbursement than in the period preceding it. This may be the result of poor management, degraded quality of inputs, improved quality of outputs, or some other unobserved trend during this same period. The study also concludes that higher occupancy rates are associated with lower efficiencies and that patients in restraints produce severe negative effects on efficiency. In addition, it appears that a sudden increase in administrative hours in the final study year played a role in that year's lower average efficiency scores, but also that all input levels generally rose while output levels remained virtually constant.