Annual Relapse Rates in Placebo Groups Decreasing Over Time
Developed by Khawai.
The data visualization below demonstrates a mystery related to relapsing remitting multiple sclerosis (RRMS). For some unknown reason, the annualized relapse rates (ARRs) among RRMS patients in the placebo arms of phase 3 trials have decreased substantially over the last two decades. In the 1990s ARRs tended to be above 1.0, but more recently ARRs have been 0.5 and below.
This mystery has more than academic interest. Sample size calculation is a key aspect in the planning of any trial. Planning a phase 3 randomized, placebo-controlled trial in RRMS requires investigators to estimate the expected ARR in the placebo group in advance.
The aim of this data visualization is to assess the timing variance of ARR in placebo groups. In the chart below, we’ve displayed the ARR of 31 trials reported between 1993 and 2014. Each circle corresponds to a trial and the size of each circle is proportional to the sample size. We used a log-linear model, with ARR on the y-axis represented logarithmically and time on the x-axis represented linearly. The red line depicts the regression of log ARR as a function of time (year/month of publication).
The regression model can be written as follows:
log10 (ARR) = 0.104055 + (-0.02449 x Year)
The slope (-0.02449) is significantly different from 0, with a p-value = 0.000001599.
It’s unclear why people in the placebo groups of recent trials show a less active disease course but this fact brings new challenges in the design of upcoming RRMS trials. Lower event rates require larger trials, and this of course will increase costs.
What are some possible reasons for this change, and how would one go about testing for these hypotheses? Please add your ideas to the discussion in our Forum .
Note: mouse-over the circles to display trial information.
Data source: click here to download the database (csv format)