Abstract:
|
When a new diagnostic device with continuous output is developed, we can evaluate its measurement accuracy by comparison with an established device by a so-called method comparison study. The method comparison study may involve the comparison of the new device and comparator via paired results from patient samples. To establish a new device’s measurement range, method comparison study is required in addition to linearity study, Limit of Blank, Limit of Detection, and Limit of Quantitation studies etc. If the comparator does not cover the entire meaning range of the new device, the new device’s measuring range must be truncated and be consistent to that of the comparator’s. However, it will be useful to assess the new device performance for the data range which is not covered by the comparator using some novel statistical method. In this talk, we will examine missingness under various scenarios and try to impute the missing comparator method under various conditions. We will simulate data to assess the condition when is acceptable to extrapolate the relationship where the comparator’s low-end value is missing.
|