Laboratory data analyses responders analysis

The term "responders analysis" was first introduced in Chapter 9 with regard to clinical laboratory data. A responders analysis approach can be used in the context of efficacy data, as well. Consider a double-blind, placebo-controlled, therapeutic confirmatory trial of an investigational antihypertensive ("test drug"). Based on earlier experience, a period of 12 weeks is considered sufficient to observe a clinically meaningful treatment effect that can be sustained for many months. In this study, a participant whose SBP is reduced by at least 1 0 mmHg after 1 2 weeks... 

In a way, these norms correspond to the ideal organisation of a laboratory. Certain rules are difficult to respect, particularly those of recording the flow of information acquired by instruments, which would require enormous amounts of computer memory. As a result, no laboratory presently has the capacity to respond 100% to the norms of archiving all raw data acquired by an instrument. Nevertheless, GPL represents a charter of laboratory operations, which has the advantage that it defines problems, provides answers and gives responsibilities to every person implicated in each stage of analysis. 

ABRF-96SEQ offered a unique opportunity for participating laboratories ability to interpret the same data. There were 95 participants in the study, of which 98% perform sequence analysis, 54% use mass spectrometty routinely, and 64% provide an internal sequence analysis service. The total number of sequence reviewers were 157 or 1.65 per response and the total years of experience were 1061.4 or 6.8 years per reviewer. Concerning everyday sequence calling, 37% report only one person reviews the data while only 17% use more than 1 person 72% say that a simple sequence is called by one person and that complex data is evaluated by more than one reviewer. There were 165 protein sequencers in the responding laboratories and are distributed as follows 133 Applied Biosystems/Perkin Elmer, 22 Hewlett-Packard, 8 Porton/Beckman, and 2 MiUigen. Respondents reported the most difficult amino acids to identify were Cys (64.5%), Trp (55%), and Ser (22.5%). 

---