This study discusses the use of robust
regression analysis using the least trimmed of squares method to evaluate the relative
efficiency of decision making units in organizations in the case of the existence of outliers and the effect of nonhomogeneus outside factor.
efficiency unit evaluated in this study is the efficiency of hospital physicians where the service cost is the input and the number of high severity discharges and low severity discharges of patients as the outputs. The results show that the estimation of the unit efficiency using LTS method is more stable.