Transl Clin Pharmacol.  2017 Jun;25(2):93-100. 10.12793/tcp.2017.25.2.93.

Bioequivalence data analysis for the case of separate hospitalization

• Bae KS
• Kang SH

• KMID: 2386766
• DOI: http://doi.org/10.12793/tcp.2017.25.2.93

Abstract

A bioequivalence study is usually conducted with the same-day drug administration. However, hospitalization is occasionally separated for logistical, operational, or other reasons. Recently, there was a case of separate hospitalization because of difficulties in subject recruitment. This article suggests a better way of bioequivalence data analysis for the case of separate hospitalization. The key features are (1) considering the hospitalization date as a random effect than a fixed effect and 2) using "PROC MIXED" instead of "PROC GLM" to include incomplete subject data.

Keyword

Bioequivalence; Separate hospitalization; Mixed effects model

MeSH Terms

Hospitalization*
Statistics as Topic*
Therapeutic Equivalency*

Figure

• 

  Figure 1 Subject disposition.

• 

  Figure 2 SAS script for data loading. ADM, hospitalization (drug administration) group code (1, 2, or 3); SEQ, treatment sequence group (RT, reference then test treatment; TR, test then reference treatment); PRD, period (1 or 2); TRT, treatment (T, test treatment; R, reference treatment); SUBJ, subject ID; CMAX, maximum concentration (Cmax) value in original scale; LNCMAX, Cmax value in natural log scale.

• 

  Figure 3 Results of the independent two-group t-test.

• 

  Figure 4 Result of conventional 2 × 2 model (Model 2). (a) ANOVA result, (b) 90% confidence interval, SEQ, treatment sequence group; SUBJ, subject ID; PRD, period; TRT, treatment; PE, point estimate; LL, lower limit; UL, upper limit; WD, width of confidence interval.

• 

  Figure 5 Result of full Model (3A) with drug administration (ADM) as a fixed factor and period (PRD) nested within ADM. (a) ANOVA result, (b) 90% confidence interval, ANOVA, analysis of variance; SEQ, treatment sequence group; SUBJ, subject ID; TRT, treatment; PE, point estimate; LL, lower limit; UL, upper limit; WD, width of confidence interval.

• 

  Figure 6 Result of reduced Model (3B) with drug administration (ADM) as a fixed factor and period (PRD) nested within ADM. (a) ANOVA result, (b) 90% confidence interval, ANOVA, analysis of variance; SEQ, treatment sequence group; SUBJ, subject ID; TRT, treatment; PE, point estimate; LL, lower limit; UL, upper limit; WD, width of confidence interval.

Reference

1. Park SH. Design of Experiments. 2nd ed. Seoul: Min-Young Sa;2003. p. 58–60. p. 107–109. p. 146–148.

2. EMA. Questions & Answers: positions on specific questions addressed to the Pharmacokinetics Working Party (PKWP). 2015. p. 32.

Smith MK. Inappropriately Designating a Factor as Fixed or Random. Accessed 1 May 2017. https://www.ma.utexas.edu/users/mks/statmistakes/fixedvsrandom.html.

4. SAS. SAS/STAT 9.3 User's Guide. SAS Institute;2011. p. 217–218.

5. SAS. SAS/STAT 14.1 User's Guide. SAS Institute;2015. p. 123.

6. Elliott AC, Woodward WA. SAS Essentials: Mastering SAS for Data Analytics. 2nd ed. Wiley;2015.

7. Galwey NW. Introduction to Mixed Modeling: Beyond Regression and Analysis of Variance. 2nd ed. Wiley;2014.

8. Haney SA, Bowman D, Chakravarty A, Davies A, Shamu C. An Introduction to High Content Screening: Imaging Technology, Assay Development, and Data Analysis in Biology and Drug Discovery. Wiley;2015.

9. Torbeck LD. Pharmaceutical and Medical Device Validation by Experimental Design. CRC Press;2017.

10. Aris VM. 8. Using microarrays to measure cellular changes induced by biomaterials. Characterization of biomaterials. Woodhead Publishing;2012.

11. Gibson D. Methods in Comparative Plant Population Ecology. 2nd ed. Oxford University Press;2014.

12. Diletti E, Hauschke D, Steinijans VW. Sample size determination for bioequivalence assessment by means of confidence intervals. Int J Clin Pharmacol Ther Toxicol. 1992; 30(Suppl 1):S51–S58. PMID: 1601532.