A nearly unbiased test for individual bioequivalence problems using probability criteria |
| |
| Institution: | 1. Children''s Health and Exercise Research Centre, University of Exeter, Exeter, UK;2. Population Health Sciences and Medical Schools, University of Bristol, Bristol, UK;3. Congenital Heart Unit, Bristol Royal Hospital for Children and Heart Institute, Bristol, UK;4. Department of Paediatrics, “Carol Davila” University of Medicine and Pharmacy, Bucharest, Romania;5. Athlete Health and Performance Research Centre and the Sports Medicine Department, Aspetar Qatar Orthopaedic and Sports Medicine Hospital, Doha, Qatar;6. Institute of Sport Exercise and Health, University College London, London, UK;7. St. Bartholomew''s Hospital, Barts Health National Health System Foundation Trust, London, UK;8. Research Institute for Sport and Exercise Sciences, Liverpool John Moores University, Liverpool, UK;9. Cardiovascular Institute, Hospital Clinic de Barcelona and Barcelona Football Club Medical Services, Barcelona, Spain;10. Manchester United Football Club, Football Medicine & Science Department, AON Training Complex, Manchester, UK;11. National Institute for Health Research Cardiovascular Biomedical Research Centre, Bristol Heart Institute, UK;1. Departement of Statistics, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia;2. Badan Pusat Statistik-Statistics Indonesia, Jakarta, Indonesia |
| |
| Abstract: | Statistical bioequivalence has recently attracted lots of attention. This is perhaps due to the importance of setting a reasonable criterion on the part of a regulatory agency such as the FDA in the US in regulating the manufacturing of drugs (especially generic drugs). Pharmaceutical companies are obviously interested in the criterion since a huge profit is involved. Various criteria and various types of bioequivalence have been proposed. At present, the FDA recommends testing for average bioequivalence. The FDA, however, is considering replacing average bioequivalence by individual bioequivalence. We focus on the criterion of individual bioequivalence proposed earlier by Anderson and Hauck (J. Pharmacokinetics and Biopharmaceutics 18 (1990) 259) and Wellek (Medizinische Informatik und Statistik, vol. 71, Springer, Berlin, 1989, pp. 95–99; Biometrical J. 35 (1993) 47). For their criterion, they proposed TIER (test of individual equivalence ratios). Other tests were also proposed by Phillips (J. Biopharmaceutical Statist. 3 (1993) 185), and Liu and Chow (J. Biopharmaceutical Statist. 7 (1997) 49). In this paper, we propose an alternative test, called nearly unbiased test, which is shown numerically to have power substantially larger than existing tests. We also show that our test works for various models including 2×3 and 2×4 crossover designs. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
