Implicit rationing criteria in non-small-cell lung cancer treatment.

Data collected from lung cancer patients attending the Victoria Clinic of the British Columbia Cancer Agency are used to investigate how resources are rationed in the treatment of non-small-cell lung cancer (NSCLC). An ordered logit model is estimated to analyse empirically the relationship between treatment selection and: tumour stage, size and differentiation; the Feinstein index; Karnofsky performance status (KPS); and the patient's age, gender and marital and smoking status. Implicit rationing is found to occur with respect to all of these factors except the Feinstein index, gender and marital status. With respect to age, KPS and smoker status the main empirical results are: (a) an increase in age from 50 to 85 reduces the expected treatment expenditure by 50-70%, depending on the patient's KPS and smoker status; (b) patients with a KPS less than 80 and of 80, receive 30-46% and 75-85%, respectively, of the expected treatment expenditure for patients with a KPS of 90 or 100, depending on age and smoker status; (c) the expected treatment expenditure for active smokers is about 71-86% of the expenditure for non- or former smokers depending on age and KPS.