Stability of finite difference deconvolution II: simulation studies

✍ Scribed by Jianguo Li; David J. Cutler

• Publisher: John Wiley and Sons
• Year: 1998
• Tongue: English
• Weight: 142 KB
• Volume: 19
• Category: Article
• ISSN: 0142-2782
• DOI: 10.1002/(sici)1099-081x(199812)19:9<595::aid-bdd144>3.0.co;2-4

✦ Synopsis

Theoretical analysis of the stability of finite difference deconvolution (FDD) indicates that if the cumulative amount function is used to characterize the drug input the method is stable for any sampling schedule for an intravenous unit impulse response function. The analysis also indicates that the method is stable for an oral unit impulse response only for well designed sampling schedules. This article confirms these results through numerical simulation experiments. It is shown that the assumption that the unit impulse response is error-free has an influence on the performance of FDD which is generally of no practical significance, except possibly for the first few points estimated. It is also shown that there is no significant interaction between the statistical error due to data noise and the deterministic algorithm error. The major source of error in practice is likely to be the data noise in the input response function. The simulations confirm that, with the estimated cumulative amount function as the quantity estimated and, with a well designed sampling schedule for the case of an oral unit impulse response, FDD is in practice an accurate and stable method with acceptable precision under a typical error disturbance.