COST-EFFECTIVENESS ANALYSIS OF CARDIOPULMONARY RESUSCITATION IN CRITICALLY ILL PATIENTS WITH CANCER

Objective:  Cardiopulmonary resuscitation (CPR) is thought to be a very low-yielding intervention in critically ill patients with cancer. This paper illustrates the incremental cost-effectiveness and cost-utility of CPR administration in critically ill patients with solid tumors, lymphoma/myeloma, leukemia, and hematopoietic stem cell transplantation (HSCT) at a comprehensive cancer center.

Methods:  A decision analysis model was developed and populated with retrospective data for each of four cancer groups with patients who underwent CPR from 1993 to 2000 in a medical intensive care unit (ICU).   All costs from the healthcare institution perspective post CPR were captured and adjusted to 2007 dollars.   For branches without any patients, the average costs and outcomes from the other cancers were utilized with ensuing sensitivity analysis.  Outcomes of importance included both life years saved (LYS) and quality-adjusted life years (QALYs), the latter calculated using utility values in the literature.  Tornado diagrams were created to show the impact of probability of survival >24 hours following the initial resuscitation, probability of survival to ICU discharge, probability of survival to hospital discharge, and probabilities of discharge to another facility or home.  Subsequent one-way sensitivity analyses were conducted on the most important variables.

Results: During the study period, 1,651 patients experienced cardiac arrest in the ICU of whom, 404 (24.5%) underwent CPR; 232 (57%) were male and the average age was 53.  The costs per LYS were:  solid tumors - $95,882/LYS; lymphoma/myeloma - $41,037/LYS; leukemia - $1,671,684/LYS; HSCT recipient - $1,840,700/LYS.  The costs per QALYs were: solid tumors - $153,408/QALY; lymphoma/myeloma - $56,593/QALY; leukemia - $21,633,563/QALY; HSCT recipient - $24,542,667/QALY.  Tornado diagrams revealed that the probability of surviving to ICU discharge had the highest impact on the cost per LYS in three of the models, except leukemia, where probability of survival to hospital discharge had the highest impact.  Subsequent one-way sensitivity analysis exhibited a wide range of values with patients with solid tumors and lymphoma/myeloma groups exhibiting the least variation.

Conclusions: Compared to other interventions in the cost-effectiveness range of $50,000 to $100,000 per QALY, the application of CPR is comparable in patients with lymphomas/myelomas.  However, the cost per QALY exceeds this range in the other three cancer groups.  The results of this study provide a means to guide decision making regarding end-of-life intervention alternatives.