Author + information
- Tal Lorberbaum, MA and
- Nicholas P. Tatonetti, PhD∗ ()
- ↵∗Department of Biomedical Informatics, Columbia University, 622 West 168th Street, PH20, New York, New York 10032
We thank Dr. Lazzerini and colleagues for sharing their clinical observations supporting the drug−drug interaction (DDI) findings in our recent paper (1). In isolation, their observations could be dismissed as idiosyncratic by skeptics, but data science—the application of rigorous analytical methods to large datasets—provides the context needed to explain these effects across large patient populations. The integration of adverse event report mining, electronic health record corroboration, electrophysiology experiments, and now these case studies paints a clear picture of a DDI that could not have been identified using traditional surveillance approaches (2).
We have further capitalized on the promise of health data science by participating in Observational Health Data Sciences and Informatics (OHDSI), an international network of >140 researchers in >20 countries performing large-scale analyses on distributed databases that in total comprise >600 million patients (3). Each participating site converts their patient data to a common data model that enables analyses created at one location to be replicated across the entire network while maintaining patient privacy. This open and international collaboration has developed open source software tools for data exploration and evidence generation, conducted multivariate analyses to predict drug side effects, and improved definition of the heterogeneity in treatment pathways (4,5). The latter study was performed on an underlying population of >250 million patients from 11 data sources in 4 countries and took 20 days from conception to results from 7 sources, highlighting the efficiency with which these analyses can be performed at scale.
In the case of ceftriaxone and lansoprazole, we used a combination of public and private clinical data to identify the interaction. Through collaborations like OHDSI, these data are available to all researchers with interests in large-scale retrospective studies. We hope that the current study motivates others to participate in these new data science initiatives.
Please note: Dr. Tatonetti is a compensated advisor to Advera Health, Inc. Mr. Lorberbaum has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Lorberbaum T.,
- Sampson K.J.,
- Chang J.B.,
- et al.
- ↵Observational Health Data Sciences and Informatics. Available at: https://www.ohdsi.org/. Accessed March 3, 2017.
- Hripcsak G.,
- Ryan P.B.,
- Duke J.D.,
- et al.