Clinical Trials

Clinical Trials: A Practical Guide ■❚❙❘Figure 2. Repeated measures with no clear pattern or trend for a subject that could be better summarizedby a mean.140= mean130SBP (mm Hg)1201101000 1 2 3 4 5 6 7 8 9 10Time (weeks)SBP = systolic blood pressure.maximum CD4 T-cell count was determined for four age quartiles. The medianmaximum CD4 T-cell count by increasing age quartile was 500 × 10 6 cells/L,448 × 10 6 cells/L, 430 × 10 6 cells/L, and 419 × 10 6 cells/L (P < 0.0001), illustratingthat older patients had a lower maximum CD4 count and indicating a poorerresponse to HAART. This result was consistent with the other results of the studysuch as the maximum CD4 T-cell gain (251 × 10 6 cells/L, 246 × 10 6 cells/L,212 × 10 6 cells/L, and 213 × 10 6 cells/L, for increasing age quartiles [P = 0.0003])and was still significant after adjustments for other variables influencing CD4T-cell response.Time to reach the maximumA statistic related to the maximum value is the time to reach the maximum.This is a measure of how quickly a subject responds to the drug. Although theprevious two statistics avoid the problem of correlation structure seen withrepeated measurement data, by reducing multiple measurements to a single indexcorrelation they are susceptible to missing observations. The missing valuescould be higher than the actual measurements that were recorded and could comeat earlier time points. If the time point at which the unknown theoreticalmaximum value would be achieved is not selected in the design stage, inaccurateestimates will be obtained for both the maximum value and the time to reach it(see Chapter 13).323