Practical approaches to minimize problems with missing quality of life data

Stat Med. 1998;17(5-7):725-37. doi: 10.1002/(sici)1097-0258(19980315/15)17:5/7<725::aid-sim817>3.0.co;2-1.

Abstract

Missing information on quality of life (QOL) is a significant problem in many cancer trials particularly for patients with advanced disease, where clinical deterioration may be a reason for not responding to quality of life assessments. Examples from four clinical trials are presented where non-respondents to quality of life assessments have poorer health than respondents. In this context, auxiliary outcome variables, such as health status, may be useful proxies in assessing the impact of missing QOL data on estimated treatment effects. This approach is illustrated in a trial of palliative treatment in advanced cancer. A method for imputation of missing QOL data based on auxiliary outcome variables is also illustrated. However, the most effective method of minimizing the problem of missing data is in designing the trial with preventative strategies in place. Since some missing data due to deteriorating health may still occur, the design should include the collection of auxiliary QOL information. Preventative strategies are illustrated with an ongoing trial in advanced breast cancer.

Publication types

  • Research Support, Non-U.S. Gov't
  • Review

MeSH terms

  • Breast Neoplasms / drug therapy
  • Clinical Trials as Topic / methods*
  • Female
  • Humans
  • Linear Models
  • Models, Statistical*
  • Myocardial Infarction / drug therapy
  • Quality of Life*
  • Research Design*