|
Abstract:
|
New medical device approvals are to be based on the results of in vitro and in vivo studies through research. However, a design of an in vitro experiment and the subsequent statistical analysis may be complicated. For such cases, applying simple summary statistics to characterize variability is insufficient; more sophisticated techniques are required. A short comparison of applicable approaches (e.g., quality control summary, GLM, mixed model) will be presented with special emphasis paid to the underlying assumptions, correct interpretation of results, and evaluation of methods. Also, examples will be discussed in the context of hemodialyzer reuse studies using data collected by device manufactures. Factors of interest will be type of dialyzer, lot of production, blood flow, and number of reuse.
|
