Do-not-admit versus inpatient surgery in an HMO: determinants of choice and the implications for medical care costs.

We develop a model for investigating the implications of policies that have encouraged a shift from inpatient to do-not-admit (DNA) surgery. We use discriminant function analysis on date for two surgical procedures from the Kaiser Permanente Medical Care Program of Portland, Oregon. Case attributes found to be significantly associated with the choice of surgery mode are surgeons' rate of inpatient surgery, number of chronic conditions per patient, time in surgery, number of procedures performed, and type of anesthesia used. Our estimates of cost savings provide support on economic grounds for the use of DNA surgery, for the types of surgery investigated. Our results also suggest that simple evaluation methods, based on the mean length of stay and on extrapolation of proportion of DNA cases from the base year to the current year, may overestimate the cost savings derived from the shift to DNA surgery.