youngprewett

not attempt to collect the same information, so a recording bias probably exists.As already noted, we are not confident we could identify all files associated with a single case, so some double-counting may have occurred. 199 The methods we used to review the files also had important limitations We relied on single-person review to screen cases. In the eight weeks we were given to review the files, we could not do dual review (which is standard in all our other EPC work) of over 18,000 cases. Therefore, more coding errors may have occurred than in situations where we use dual review. Mitigating this limitation is that we did do formal inter-rater reliability testing and demonstrated excellen Super C Shield t reliability among reviewers. Also, the principal investigator reviewed all cases that were identified as serious.Furthermore, we identified nearly all the serious events identified by Metabolife, plus many more that Metabolife did not identify. So, while we acknowledge that there may still be errors in the data, we do not think they are so numerous or egregious as tothreaten our conclusions.The Metabolife analysis did not undergo as extensive a review process as did the other sections of this report. The Metabolife analysis was reviewed by three experts and two federal agencies, in contrast to the much more extensive review process for the other sections of the report. Furthermore, because of the timeline necessary to produce the final report, the time available to the reviewers of the Metabolife .

http://www.optimalstackfacts.org/before-buying-super-c-shield-read-it-first/