Abstract:
|
The integration of biostatisticians into diverse research teams is becoming increasingly important for addressing multidisciplinary problems with increasingly complex data sources. The current case study presents one such situation in which the collaboration between two biostatisticians and a team of surveillance-focused public health researchers led to the use of an as-yet unconventional method. In analyses of a rate over time, statisticians may intuitively consider time-series or growth models the gold standard. However, in the case of proportional outcomes—presently, the case fatality rate in the study of opioid overdoses—a myriad of novel data conditions may arise, thus necessitating use of alternative methods. The discussions between all members of the research team led to the decision to use beta regression to model the proportional outcome (e.g., the case fatality rate) over time. It was only through these extensive discussions and close collaboration that this appropriate methodology was selected. Ultimately, this case study presents a compelling rationale for the increasing integration of collaborative biostatisticians on research teams.
|