Translational Medicine-Big Data

Translational Medicine-Big Data: Does it really lead to better decisions?
InnovationWell 2015

Barry Hardy (Douglas Connect), Frank Hollinger (Sphaera Pharma) and Michael Liebman (Strategic Medicine

Session Chair: 

Michael Liebman (Strategic Medicine)

Friday, 13 February, 2015 - 13:45

Charles Commons, Johns Hopkins University
10 East 33rd Street


Baltimore, MD, USA 21218

Translational medicine, while contributing to the identification of new therapeutics and diagnostics and to improvement of patient care, has remained a uni-directional process, from laboratory to clinic. This primary focus is based on good science but has been more opportunistic in producing results than positioned to specifically address unmet clinical needs. To bridge this gap, it is critical that translation move from “bench–to-bedside” to become “bedside-to-bench-to-bedside”.

It is necessary that the implementation of big data approaches recognize the complexity involved in unmet clinical needs, i.e. the real world patient and physician.

To accomplish this, it is necessary to understand the full complexity of the disease process, the interaction of the patient with the healthcare system and existing needs to:

- understand the limitations in clinical diagnosis and its implications
- understand the quality of clinical guidelines and physician compliance
- understand the impact of patient (lack of)adherence and its relation to their perception of risk
- enhance disease stratification, not just patient stratification
- incorporate the reality of co-morbidities and concurrent medication
- incorporate clinical history, lifestyle and environmental exposure

In spite of the incorporation of Big Data approaches, the solutions do not come from simple data mining and statistical analysis. It is critical to develop functional models that support qualitative reasoning and sensitivity analysis. The ability to reason about the quality and completeness of data, the impact of missing data and the resolution of conflict from multiple data sources is critical to support optimal clinical decision support. The development of models to provide the foundation for such data analysis provides the cornerstone by which better decision-making can occur.