Robustness of crossover trials against subject drop-out – Examples of perpetually connected designs

Abstract

When performing a repeated measures experiment, such as a clinical trial, there is a risk of subject drop-out during the experiment. If one or more subjects leave the study prematurely, a situation could arise where the eventual design is disconnected, implying that very few treatment contrasts for both direct effects and carryover effects are estimable. This paper aims to identify experimental conditions where this problem with the eventual design can be avoided. It is shown that in the class of uniformly balanced repeated measurement designs consisting of two or more Latin squares, there are planned designs with the following useful property. Provided that all subjects have completed the first two periods of study, such a design will not be replaced by a disconnected eventual design due to drop-out, irrespective of the type of drop-out behaviour that may occur. Designs with this property are referred to as perpetually connected. These experimental conditions are identified and examined in the paper and an example of at least one perpetually connected uniformly balanced repeated measurement design is given in each case. The results improve upon previous contributions in the literature that have been confined largely to cases in which drop-out occurs only in the final periods of study.

Keywords

Crossover design subject drop-out missing data clinical trial uniformly balanced repeated measurement design perpetually connected

Get full access to this article

View all access options for this article.

References

Jones

Kenward

. Design and analysis of cross-over trials ▪▪. 3rd ed. Monographs on Statistics and Applied Probability 138. CRC Press. 2014.

Wellek

Blettner

. On the proper use of the crossover design in clinical trials: Part 18 of a series of evaluation of scientific publications 20122nd edn. Deutsches Arzteblatt International, pp. 276–281.

Senn

D’Angelo

Potvin

. Carry-over in cross-over trials in bioequivalence: theoretical concerns and empirical evidence. Pharmaceut Stat 2004; 3: 133–142.

Grattan

Hickman

Darby-Dowman

. A five way crossover human volunteer study to compare the pharmacokinetics of paracetamol following oral administration of two commercially available paracetamol tables and three development tablets containing paracetamol in combination with sodium bicarbonate or calcium carbonate. Eur J Pharmaceut Biopharmaceut 2000; 49: 225–229.

National Research Council. The prevention and treatment of missing data in clinical trials ▪▪. Panel on Handling Missing Data in Clinical Trials. Committee of National Statistics, Division of Behavioral and Social Sciences and Education. Washington, DC: The National Academies Press, 2010.

Rubin

. Inference and Missing Data. Biometrika 1976; 63: 581–592.

Molenberghs

Kenward

. Missing data in clinical studies, Chichester, UK: Wiley, 2007.

Rosenkranz

. Analysis of cross-over studies with missing data. Stat Meth Med Res 2015; 24: 420–433.

Matthews

JNS

Henderson

. Two-period, two-treatment crossover designs subject to non-ignorable missing data. Biostatistics 2013; 14: 626–638.

10.

Higgins

JPT

Green

. Cochrane handbook for systematic reviews of interventions Version 5.1.0, The Cochrane Collaboration, 2011.

11.

Low

Lewis

Prescott

. Assessing robustness of crossover designs to subjects dropping out. Stat Comput 1999; 9: 219–227.

12.

Shih

. Problems in dealing with missing data and informative censoring in clinical trials. Curr Control Trial Cardiovasc Med 2002; 3: 4–4.

13.

Godolphin

. Simple pilot procedures for the avoidance of disconnected experimental designs. Appl Stat 2004; 53: 133–147.

14.

Mills

Chan

et al.

Design, analysis and presentation of crossover trials. Trials 2009; 10: 1–6.

15.

Cochran

Autrey

Cannon

. A double change-over design for dairy cattle feeding experiments. J Dairy Sci 1941; 24: 937–951.

16.

Hedayat

Afsarinejad

. Repeated measurements designs 2. Ann Stat 1978; 6: 619–628.

17.

Cheng

. Balanced repeated measurement designs. Ann Stat 1980; 8: 1272–1283. Corrigendum. Ann Stat 1983; 11: 349.

18.

Hedayat

Yang

. Universal optimality of balanced uniform crossover designs. Ann Stat 2003; 31: 978–983.

19.

Bate

Godolphin

. Choosing cross-over designs when few subjects are available. Comput Stat Data Analys 2008; 52: 1572–1586.

20.

Majumdar

Dean

Lewis

. Uniformly balanced repeated measurements designs in the presence of subject dropout. Stat Sinica 2008; 18: 235–253.

21.

Zhou

Majumdar

. On uniformly balanced crossover designs efficient under subject dropout. J Stat Theory Pract 2012; 6: 178–189.

22.

Williams

. Experimental designs balanced for the estimation of residual effects of treatments. Aus J Scient Res 1949; 2: 149–168.

23.

Newcombe

. Sequentially balanced three-squares cross-over designs. Stat Med 1996; 15: 2143–2147.

24.

Godolphin

. The robustness of resolvable block designs against the loss of whole blocks or replicates. J Stat Plan Inference 2015; 163: 34–42.

25.

Godolphin

. The use of treatment concurrencies to assess robustness of binary block designs against the loss of whole blocks. Aus New Zealand J Stat 2015; 57: 225–239.

26.

CHMP. Guideline on the investigation of bioequivalence. London, 2010.

27.

FDA Guidance. Guideline for industry: statistical approaches to establishing bioequivalence, 2001.

28.

Srivastava

Anderson

. Some basic properties of multidimensional partially balanced designs. Ann Math Stat 1970; 41: 1438–1445.

29.

Godolphin

. On the connectivity problem for m-way designs. J Statistical Theory Pract 2013; 7: 732–744.

30.

Harville

. Matrix algebra from a statisticians’s perspective, New York, NY: Springer-Verlag, 1997.