Hospital Costs and Severity of Illness in Three Types of Elective Surgery

Background If patients who are more severely ill have greater hospital costs for surgery, then health-care reimbursements need to be adjusted appropriately so that providers caring for more seriously ill patients are not penalized for incurring higher costs. The authors' goal for this study was to determine if severity of illness, as measured by either the American Society of Anesthesiologists Physical Status (ASA PS) or the comorbidity index developed by Charlson, can predict anesthesia costs, operating room costs, total hospital costs, or length of stay for elective surgery. Methods The authors randomly selected 224 inpatients (60% sampling fraction) having either colectomy (n = 30), total knee replacement (n = 100), or laparoscopic cholecystectomy (n = 94) from September 1993 to September 1994. For each surgical procedure, backward-elimination multiple regression was used to build models to predict (1) total hospital costs, (2) operating room costs, (3) anesthesia costs, and (4) length of stay. Explanatory candidate variables included patient age (years), sex, ASA PS, Charlson comorbidity index (which weighs the number and seriousness of coexisting diseases), and type of insurance (Medicare/Medicaid, managed care, or indemnity). These analyses were repeated for the pooled data of all 224 patients. Costs (not patient charges) were obtained from the hospital cost accounting software. Results Mean total hospital costs were $3,778 (95% confidence interval +/- 299) for laparoscopic cholecystectomy, $13,614 (95% CI +/- 3,019) for colectomy, and $18,788 (95% CI +/- 573) for knee replacement. The correlation (r) between ASA PS and Charlson comorbidity scores equaled 0.34 (P < .001). No consistent relation was found between hospital costs and either of the two severity-of-illness indices. The Charlson comorbidity index (but not the ASA PS) predicted hospital costs only for knee replacement (P = .003). The ASA PS, but not the Charlson index, predicted operating room and anesthesia costs only for colectomy (P < .03). Conclusions Severity of illness, as categorized by ASA PS categories 1-3 or by the Charlson comorbidity index, was not a consistent predictor of hospital costs and lengths of stay for three types of elective surgery. Hospital resources for these lower-risk elective procedures may be expended primarily to manage the consequences of the surgical disease, rather than to manage the patient's coexisting diseases.