To compute cost-effectiveness/cost-utility (CE/CU) ratios, from the treatment clinic and societal perspectives, for high-dose palliative radiotherapy treatment (RT) for advanced non-small-cell lung cancer (NSCLC) against best supportive care (BSC) as comparator, and thereby demonstrate a method for computing CE/CU ratios when randomized clinical trial (RCT) data cannot be generated.
Unit cost estimates based on an earlier reported 1989-90 analysis of treatment costs at the Vancouver Island Cancer Centre, Victoria, British Columbia, Canada, are updated to 1997-1998 and then used to compute the incremental cost of an average dose of high-dose palliative RT. The incremental number of life days and quality-adjusted life days (QALDs) attributable to treatment are from earlier reported regression analyses of the survival and quality-of-life data from patients who enrolled prospectively in a lung cancer management cost-effectiveness study at the clinic over a 2-year period from 1990 to 1992.
The baseline CE and CU ratios are $9245 Cdn per life year (LY) and $12,836 per quality-adjusted life year (QALY), respectively, from the clinic perspective; and $12,253/LY and $17,012/QALY, respectively, from the societal perspective. Multivariate sensitivity analysis for the CE ratio produces a range of $5513-28,270/LY from the clinic perspective, and $7307-37,465/LY from the societal perspective. Similar calculations for the CU ratio produce a range of $7205-37, 134/QALY from the clinic perspective, and $9550-49,213/QALY from the societal perspective.
The cost effectiveness and cost utility of high-dose palliative RT for advanced NSCLC compares favorably with the cost effectiveness of other forms of treatment for NSCLC, of treatments of other forms of cancer, and of many other commonly used medical interventions; and lies within the US $50, 000/QALY benchmark often cited for cost-effective care.
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.