Standard Article

Statistics for Toxicology

Methodology

  1. Peter N. Lee MA, CStat1,
  2. David Lovell PhD2

Published Online: 15 DEC 2009

DOI: 10.1002/9780470744307.gat037

General, Applied and Systems Toxicology

General, Applied and Systems Toxicology

How to Cite

Lee, P. N. and Lovell, D. 2009. Statistics for Toxicology. General, Applied and Systems Toxicology. .

Author Information

  1. 1

    P N Lee Statistics and Computing Ltd, Independent Consultant in Statistics, Sutton, Surrey, UK

  2. 2

    University of Surrey, Reader in Medical Statistics, Postgraduate Medical School, Guildford, Surrey, UK

Publication History

  1. Published Online: 15 DEC 2009

Abstract

This chapter discusses statistical issues in the design, analysis and interpretation of toxicological data. Aimed at the statistically unqualified reader, its concern is principles rather than techniques. Although discussion is often in relation to the long-term rodent carcinogenicity study, the principles and methods described apply widely in toxicology. After describing the statistician's role, some general issues including bias, hypothesis testing, p-values and multiple comparisons are referred to. The chapter then discusses the relevance of Good Laboratory Practice (GLP) and quality assurance (QA), before considering various aspects of experimental design and conduct, including numbers of animals, choice of controls, dose levels, duration, stratification and randomization. Various considerations general to statistical analysis are discussed next, including combining, pooling and stratification, trend analysis, multiple control groups, age adjustment, missing data, historical controls and the need to take context of observation into account in pathology data. The chapter advocates the use of simple statistical methods and recommends the appropriate techniques to use in many standard situations. Details of the actual methods are not presented, the reader being pointed to appropriate publications and software.

Keywords:

  • experimental design;
  • statistical analysis;
  • pathology;
  • carcinogenicity;
  • toxicology;
  • probability;
  • stratification;
  • randomization;
  • age-adjustment;
  • dose-related trend;
  • context of observation