preview

Clinical Synthesis Essay

Good Essays

We consider data from a clinical trial, with two treatments and two measurement times as analyzed by \cite{matt}. There are 422 subjects in this trial, but not all completed it i.e there is dropout. The data are real but not public. The trial was randomized and observations were taken pre-randomization and at two times post-randomization: in this work we consider only the post-randomization observations. For confidentiality reasons, \cite{matt} scaled all observed responses by the mean and standard deviation of the responses at time 1 and we will refer simply to treatment and response without describing what they are. There are no covariates other than treatment type and time. Without treatment by time interaction the parameter vector is $(\beta_1,\beta_2,\beta_3,\beta_4)$. All 422 subjects provided a response at time 1, but 24.4\% dropped at time 2. Of 212 subjects receiving Treatment A, only 126 provided a response at time 2 and the other 86 dropped out which is equivalent to 40.57\% missing. We have no information …show more content…

each average is over those patients who have not yet dropped out. All three groups have a decreasing mean response, perhaps at a slower rate towards the third time point. The overall reduction in mean response within each treatment group is very roughly from between 90 and 95 to around 70 and 75. This appears close to the criterion for clinical improvement, which was stated in advance of the trial, to be a reduction of 20\% in the mean PANSS scores. The decrease in group 2 was smaller overall. However, at each timepoint these observed means are, necessarily, calculated only from those subjects who have not yet dropped out. Figure \ref{fig:fig4} in the right side, shows the behavior of complete cases: large differences between groups are evident, with the highest decrease in the groups 1 and 3 and the lowest in the group

Get Access