Missing Data in Clinical Studies Molenberghs G. and Kenward M. G.

Abstract

This book is part of a series under the general heading ‘Statistics in Practice’. The series provides detailed coverage of statistical concepts, methods and worked case studies in specific aspects of statistical theory and practice.Missing data are a nightmare most researchers face from time to time. As research designs have become more complex and often multicentred, the problem of missing data has become that much more common. Maintaining control over the problem of missing data is easier in studies where information is gathered prospectively, and much more difficult when information is obtained from patient records. As a rule of thumb when more than 20% of the intended information is missing the results are not very meaningful. Then the researcher faces many dilemmas. There is the ethical commitment to participants who may have undergone painful procedures to provide samples. There are also issues of financial accountability to the sponsors of the research. Also there is the moral responsibility, when coming up with a truthful scientific answer to the research question gets compromised.