Faculty Sheng Luo, PhD was recently profiled on the DCRI website about a recent grant and the importance of biostatistics in research. The article below appeared on the DCRI website on November, 2018.
When the Duke Department of Anesthesiology recently re-applied for a research grant from the National Institutes of Health (NIH), it invited DCRI’s Sheng Luo, PhD, to contribute his biostatistical expertise. The NIH grant would be for a study related to chronic pelvic pain affecting reproductive-aged women.
The trial was a multi-site, randomized, double-blinded trial to compare the efficacy of individualized treatments in alleviating pain and improving outcomes for patients suffering from vestibulodynia (VBD). Investigators hoped the study’s findings would help patients, their partners, and their clinicians make more informed decisions about pain management related to VBD. The study design was longitudinal, with patients followed over a period of time and returning for follow-up visits. “Because data will be collected at each visit,” Luo said, “it was especially important to add a statistical section in the proposal on how we analyze that kind of longitudinal data and what model the study will use. We want to plan for many variables and outcomes of interest.”
A major challenge in all clinical studies is patient enrollment and retention, and especially the probability of patient dropout, according to Luo. “We want to be ready to handle that with the right statistical analysis plan,” he said. “A study could see as much as a 10 to 20 percent dropout rate, greatly reducing the chances of statistically and clinically significant results. So we use a whole set of criteria to assess how to increase sample size from the beginning to compensate for dropout.” In general, he said, when preparing a grant application, “it’s ideal to have a statistician join the conversation as early as possible to tackle a wide range of challenges and issues. When properly designed, the study’s data should have sufficient power to show that therapies do or do not work. This is what we looked at with the VBD study.”
DCRI biostatisticians also help predict whether investigators can meet study targets on time, according to Luo, and if patients will be randomized and blinded properly. “The NIH is very concerned about these issues, which can create huge red flags. When studies are properly designed with enough patients and power to produce usable results, time and money will have been well-spent.” Working with data related to living organisms, DCRI biostatisticians both crunch numbers and help design studies where enough data and the right kind of information are collected. Along the way, they analyze, evaluate, and interpret results while accounting for biases, and missing data, and other relevant issues.
Luo said that including biostatisticians as part of a study team during the grant application process can help boost the chances for a winning proposal. Thanks in part to Luo’s contributions, this fall the Department of Anesthesiology received the five-year NIH grant for its VBD study. Luo also credited Kevin Anstrom, the DCRI’s Director of Biostatistics, for his assistance in applying for the grant. “Whether a clinical trial is for government or industry,” Luo said, “biostatistics should always play a major role. We’re glad to be brought into the process as early as possible to provide investigators with feedback on the protocol. The sooner the better, really.”