The authors present a methodology that measures improvement in customer satisfaction scores when those scores are already high and the production process is slow and thus does not generate a large amount of useful data in any given time period. The authors used these techniques with data from a midsized rehabilitation institute affiliated with a regional, nonprofit medical center. Thus, this article functions as a case study, the findings of which may be applicable to a large number of other healthcare providers that share both the mission and challenges faced by this facility. The methodology focused on 2 factors: use of the unique characteristics of panel data to overcome the paucity of observations and a dynamic benchmarking approach to track process variability over time. By focusing on these factors, the authors identify some additional areas for process improvement despite the institute's past operational success.

DEA has not been applied to rehabilitation, especially in the case where managers have limited data available.