Sumatriptan and lost productivity time: a time series analysis of diary data

Two previously conducted clinical studies assessed lost nonworkplace activity time and lost workplace productivity time due to migraine symptoms in subjects using sumatriptan for 6 months to treat their migraines after a 12- to 18-week period of using their usual therapy without sumatriptan. Although statistically significant differences in lost nonworkplace activity time and lost workplace productivity time between the usual therapy and sumatriptan treatment periods were detected using the Wilcoxon signed-rank test, this test could not determine whether differences were attributable to inherent trends in the data. This current study employed time series analysis, which detects and controls for preexisting trends in data, to further explore the possibility that the observed reductions in lost time in the two clinical studies were related to management of the subjects with sumatriptan. The intercepts and slopes of the computed linear models suggest that the initiation of sumatriptan therapy produced savings of 0.8 hours of nonworkplace activity time and 0.5 hours of workplace productivity time per patient per week. These savings were sustained throughout the sumatriptan treatment period. Preexisting trends in the data were not detected in the models. Thus the productivity gains are not associated with either time effects or the statistical phenomenon of regression to the mean, but variables that are extreme in initial measurements will tend to be closer to the center of the distribution in subsequent measurements. This strengthens the hypothesis that management of migraine with sumatriptan is associated with reductions in lost productivity time.