Cost-effectiveness in clinical trials: using multiple imputation to deal with incomplete cost data

Background Cost-effectiveness has become an important outcome in many clinical trials and has resulted in the collection of resource use data and the calculation of costs for individual patients. A specific example is a Cancer Research UK phase III trial comparing chemotherapy (CT) against standard palliative care in patients with advanced non-small cell lung cancer. Resource usage from trial entry until death were collected and costs obtained on a subset of 115 trial patients. For some patients, however, the unavailability of medical notes resulted in some cost components, and hence total cost, being missing. The 82 patients with complete data were not representative of all trial patients in terms of effectiveness and thus it was necessary to address the missing data problem. Methods Multiple imputation (MI) was used to impute values for the unobserved individual cost components, allowing total cost to be calculated and cost-effectiveness carried out for all patients in the cost sub-study. The results are compared with those from a complete case analysis. Results After MI, the results indicated that CT had a high probability of being cost-effective for a societal willingness to pay over £20 000 per life-year gained. This was in stark contrast with the complete case analysis, which suggested that CT was not a cost-effective use of resources at any reasonable level of willingness to pay for a life-year. Limitations Our findings are based on a relatively small retrospective study with all events observed. Conclusion In conclusion, cost-effectiveness analysis of the complete cases only may give biased results, and therefore, in situations where there are missing costs, MI is recommended. Clinical Trials 2007; 4: 154—161. http://ctj.sagepub.com

[1]  Simon G Thompson,et al.  How Sensitive Are Cost-Effectiveness Analyses to Choice of Parametric Distributions? , 2005, Medical decision making : an international journal of the Society for Medical Decision Making.

[2]  J Carpenter,et al.  Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians. , 2000, Statistics in medicine.

[3]  R. Souhami,et al.  Mitomycin, ifosfamide, and cisplatin in unresectable non-small-cell lung cancer: effects on survival and quality of life. , 1999, Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

[4]  Joseph L Schafer,et al.  Modeling and imputation of semicontinuous survey variables , 1999 .

[5]  Roderick J. A. Little,et al.  Statistical Analysis with Missing Data , 1988 .

[6]  J. Schafer,et al.  Missing data: our view of the state of the art. , 2002, Psychological methods.

[7]  H. Boshuizen,et al.  Multiple imputation of missing blood pressure covariates in survival analysis. , 1999, Statistics in medicine.

[8]  Roger A. Sugden,et al.  Multiple Imputation for Nonresponse in Surveys , 1988 .

[9]  S. Thompson,et al.  Analysis of cost data in randomized trials: an application of the non-parametric bootstrap. , 2000, Statistics in medicine.

[10]  Daniel F Heitjan,et al.  Bayesian estimation of cost‐effectiveness from censored data , 2004, Statistics in medicine.

[11]  A H Briggs,et al.  Handling uncertainty in economic evaluations of healthcare interventions , 1999, BMJ.

[12]  J KristinN.,et al.  Multiple Imputation for Incomplete Data With Semicontinuous Variables , 2003 .

[13]  Judi Scheffer,et al.  Dealing with Missing Data , 2020, The Big R‐Book.

[14]  Bradley Efron,et al.  Missing Data, Imputation, and the Bootstrap , 1994 .

[15]  S. Bryan,et al.  Patterns, costs and cost-effectiveness of care in a trial of chemotherapy for advanced non-small cell lung cancer. , 2002, Lung cancer.

[16]  D. Rubin,et al.  Statistical Analysis with Missing Data , 1988 .

[17]  N. Zethraeus,et al.  Definition, interpretation and calculation of cost-effectiveness acceptability curves. , 2000, Health economics.

[18]  A. Briggs,et al.  Something old, something new, something borrowed, something blue: a framework for the marriage of health econometrics and cost-effectiveness analysis. , 2002, Health economics.

[19]  Andrew Briggs,et al.  Missing... presumed at random: cost-analysis of incomplete data. , 2003, Health economics.

[20]  Joseph L Schafer,et al.  Analysis of Incomplete Multivariate Data , 1997 .