The primary goal of the Data Monitoring Committee (DMC) in a drug development paradigm is to evaluate the ongoing safety of a study or a program (i.e., group of studies) and determine if there are any glaring safety concerns for the patients. If safety concerns become apparent, the DMC has the responsibility of recommending the study be amended or stopped completely. It is the independent statistician’s (or reporting statistician’s) responsibility to provide the DMC with all the information necessary to make that decision. If the reporting statistician fails to provide the DMC with the necessary information, the safety of many future subjects could be in jeopardy. To fulfill this role, what are the needed characteristics that make a successful reporting statistician?
First and foremost, the reporting statistician must be a good statistician with broad knowledge of research designs and statistical techniques. Statisticians in the pharmaceutical industry, whether working for a pharmaceutical company or CRO, generally work on a limited number of studies for a long period of time. By doing so, they become experts in an indication’s characteristics and the analysis methods needed for these studies. The reporting statistician on the other hand works on multiple studies of diverse indications simultaneously, so while never becoming an expert on one indication, the reporting statistician has a greater breadth of familiarity with many more medical conditions than most other statisticians. Their statistical knowledge must be applicable for the wide range of statistical tasks they may be asked to perform. While being a generalist in terms of statistics, the reporting statistician must be a specialist in the analysis of safety data. The analysis of safety data often does not conform to standard analysis methods taught in graduate school and the reporting statistician must show creativity and insight when summarizing safety data.
While most statisticians do not go to medical school, the reporting statistician needs to have medical foresight into the many indications he or she will be dealing with. The safety signals that will indicate a problem in a diabetes medication will not be the same signals as those for a cancer treatment, or the labs that may be of concern in a trial with adult patients may not be the same as with pediatric patients. The reporting statistician must rely on careful reading of the study literature, the support of the medical experts on the DMC, and his/her own common sense to identify the important safety signals in a particular study.
An important quality of a good reporting statistician is good communication skills. The closed session DMC report must provide the DMC members with the information needed to make their safety evaluation of the study. The report must be comprehensive yet succinct, and it must be scientific yet understandable to non-statisticians. Presenting masses of numbers and tables is insufficient, the DMC report must tell the safety story of the study. The report must engage the DMC readers. Therefore, the reporting statistician must be not only a writer but also a graphic designer to present the information in novel and enticing ways. It must be kept in mind that the reporting statistician doesn’t have a lot of time to write the closed session report. Often study data is received four weeks before a meeting and must be reported to DMC members one week prior to the meeting. In those three weeks the reporting statistician is reviewing the data, performing a multitude of analyses, isolating safety trends, and characterizing subjects with safety issues, and writing this masterpiece of a report. The reporting statisticians’ work doesn’t end there. The reporting statistician must also present the findings of the analyses at the DMC meeting. During the one to two hour meeting, the reporting statistician needs to be informative, entertaining, engrossing, and ready to answer diverse questions. This is often one of the harder skills for reporting statistician to master, since statisticians are by nature introverted creatures.
Another important distinction between the reporting statistician and statisticians working for a pharmaceutical company concerns the quality of data being analyzed. The DMC involves looking at interim data, and by nature interim data is not clean. But data does not come with a label ‘clean’ or ‘dirty. It is up to the statistician to determine if an unusual finding reflects a true safety concern or is due to data cleanliness. In a sense, the reporting statistician must be a data whisperer, trying to interpret the integral meaning of data. The analysis of safety data is often guided by one aberrant finding and if this finding is due to incomplete or incorrect data, the reporting statistician’s navigation through the river of discovery may be misguided.
ACI is one of the nation’s leaders in organizing and conducting DMCs. The reporting statisticians at ACI are surrounded by a team of talented SAS programmers and project managers who are committed to providing our clients with the highest quality analyses and services. The highly renowned doctors who constitute our DMCs consistently praise ACI for providing the quality reports that allow them to confidently make their safety evaluations. But the success of the DMC centers around the reporting statistician, the person who obtains and presents the information that drives the DMC. The reporting statisticians are ACI proudly possess the qualities needed for successful DMCs.
Written by Brian Cohen, Director of Statistics