hsg logo white

HD InsightsCompiling the Evidence: Study Statistician

Most of the Study Statistician’s work in a clinical trial occurs before the trial begins or after it ends. Before the trial, we assist investigators in the detailed planning of the study, including enrollment criteria, choice of treatment(s), specification of primary and secondary outcome measures and follow-up period. We carry out what is quaintly known as a “power calculation” to assess whether the study is able to answer the clinical question addressed.

In a placebo-controlled study, we need to ensure that if the active treatment provides a clinically meaningful benefit over placebo, the study will have high probability of demonstrating this. Note the word “if” in the previous sentence. Ideally, before the study starts, we also consider what level of information we would need from an interim analysis to declare the data conclusive enough to halt the study because of greater than expected efficacy or because of “futility,” trends in data that indicate that the trial, if completed, would be very unlikely to show benefit of the treatment.

Responsibilities During the Study

While the study is in progress, the primary Study Statistician, along with other study personnel will remain blinded as to treatment assignments. This period is used to develop and test the programs that will be used in the statistical analysis. The Study Statistician can also assist the study team in monitoring key measures such as recruitment, retention and compliance and checking the accumulating data for inconsistencies or outliers. If external factors (such as a pandemic) require major protocol revisions, the Study Statistician can assist in assessing and hopefully ameliorating — the consequences of these for the study.

Most double-blind randomized studies will have ongoing monitoring by a Data and Safety Monitoring Board (DSMB), which may have access to individual treatment assignments. Reports to the DSMB should be prepared by a statistical liaison, who will need to be unblinded, not the Study Statistician. The statistical liaison will perform any interim analyses prespecified in the protocol or requested by the DSMB.

statistician

Wrapping up the Stats

After the study concludes, but before the randomization code is revealed, we assist investigators in identifying and correcting data errors to finalize the database. After data lock and unblinding of treatment assignments, we run our statistical analyses, providing the data output that will be included in the clinical study report, publications, and presentations. We work with the investigators on the preparation of these documents. We also assist with the archiving of data and programs so that other researchers may benefit from the hard work done by the study team.  

Prerequisites for the Role Most clinical trial statisticians have a master’s or doctoral degree in statistics, biostatistics, or a closely related field. Good communication skills are essential, as in any collaborative endeavor. Prior knowledge of the disease area or mentoring by an expert is always helpful.