A generous gift from the Minerva Foundation supported NIH cost sharing expenses and made possible the funding of the new Biostatistics Workshop program. Five workshop grants were awarded in the fall to support meetings exploring cutting edge research areas; three were held in 2016.
- Statistical and Epidemiologic Issues in Evaluating Safety of Antiretroviral Exposures in HIV-exposed Uninfected Children
Organizer: Senior Lecturer Paige Williams
This workshop was held on February 26th, 2016 at the Marriott Copley Plaza. The event brought together an interdisciplinary group of 40 faculty, students, post-doctoral fellows, and clinical experts in HIV research, to address multiple statistical and epidemiologic issues that arise in evaluating safety of ARV exposures in children born to mothers with HIV infection. Presenters included Katharine Correia, Dr. Kelesitse Phiri, Kathryn Rough, and Ellen Caniglia from the Harvard Chan School; as well as Jeanne Sibiude from the Université Paris Descartes, Fatima Kakkar from the University of Montreal, Hermione Lyall from the Imperial College London, and Rebecca Zash from the Beth Israel Deaconess Medical Center. The workshop came at the end of this year¹s annual Conference on Retroviruses and Opportunistic Infections (CROI).
- Causal Inference with Highly Dependent Data in Communicable Diseases Research
The organizing committee included Professor Victor DeGruttola, Professor Eric Tchetgen Tchetgen, Dr. Laura Balzer, Associate Professor Judith Lok, Assistant Professor Jukka-Pekka Onnela and Assistant Professor Rui Wang.
On May 25, 2016, the Department of Biostatistics hosted a workshop “Causal Inference with Highly Dependent Data in Communicable Diseases Research”. Topics included novel study designs, causal parameters, identifiability results (linking the causal parameter to a statistical parameter), and analyses to handle complex interactions and dependence between individuals within a cluster (e.g. household, clinic, village or social networks). There were many lively discussions and an interactive poster session. The workshop was attended by 60 students and faculty from 18 different universities and institutes. Overall, the workshop was a great success in that the attendees debated pressing topics in causal inference, epidemiology and network science, as well as forged cross-discipline collaborations.
- Reproducibility in Personalized Medicine Research
Organizer: Assistant Professor Lorenzo Trippa
More than 75 faculty, students and researchers attended this workshop, held on September 29, 2016 at DFCI. The goal of the workshop was to discuss methods to identify potential causes of lack of reproducibility and statistical approaches to enhance reproducibility in precision medicine. During the past decade, one of the most notable transformations in science has
been the availability of large and diverse sets of data. This trend has been accompanied by the increased use of machine learning and statistical techniques to classify patients and optimize treatments in precision medicine. In this context several concomitant factors can cause poor reproducibility levels, including unmeasured and heterogenous covariates’ distributions across studies, new technologies and ascertainment mechanisms. The focus of the workshop was
on statistical techniques and applications to understand and prevent the most important and common causes of lapses of reproducibility.
