Small molecule drugs are the cornerstone of modern medical practice. However, their use is plagued by the onset of unexpected side effects, often seen only in late-stage clinical trials or after release to the market. As a result, there have bee a number of high profile drug withdrawals and a dearth of new drug development. Characterizing the combinatorial effects of drug treatment is of particular concern. It is very difficult to empirically study these interactions before drugs enter the market because of the small samples of co- prescribed drugs in most late stage clinical drug (Phase III) studies. Some interactions can be predicted based on knowledge of shared pathways of metabolism, but many are idiosyncratic and difficult to predict. Thus, we must create surveillance methods to detect unexpected drug effects and interactions that leverage the power of large-scale clinical databases such as the electronic health records. Mining of electronic health record data for the purpose of identifying adverse drug effects is an increasingly important research challenge. For example, in response to a congressional mandate the Food and Drug Administration (FDA) established the mini-sentinel initiative in 2009 -- a pilot study that links claims and administratve data from over 31 institutions for the purpose of monitoring drug safety surveillance. In addition, public-private partnerships (e.g. the Observational Medical Outcomes Partnership) have sprouted to establish data management and analysis standards for safety surveillance. However, the potential of the EHR for drug surveillance is paralleled by an equal number of challenges. Many of these challenges are in the quality (or rather lack thereof) of data when used for secondary analyses. Data stored in the EHR are often dirty, noisy, and missing. In addition to issues regarding data capture, these data also suffer from bias which confounds analysis and makes data mining results difficult to interpret. These issues become especially acute in the context of combination therapies where the exposed patient cohorts are often small and suffer from unknown (i.e. unstudied) biases. In this proposal we present a drug safety surveillance strategy which integrates state-of-the-art signal detection algorithms with chemical systems biology data for the purpose of identifying unexpected effects of combination therapies. We present an integrative methodology which combines quantitative signal detection and chemical systems biology to mine drug effects from a large clinical database. This will require innovations in observational statistical data mining, network analysis, and integrative chemical systems biology. The result will be a set of tools for discovering drug effects and linking them to molecular interaction networks. These resources will aid federal regulators to better monitor the safety of drugs at the population level, pharmacologists who wish to understand the effects of drugs at the physiological level, and drug development researchers to explore new treatments of human disease.

Public Health Relevance

Small molecule drugs are the cornerstone of medical-practice yet their use and development is limited by occurrence unexpected effects and interactions. Adverse events that occur when multiple drugs are taken at once are of particular concern since they are not studied pre-clinically and can have significant health consequences. In this proposal we present a integrating methodology that combines quantitative signal detection and chemical systems biology to study the effects of drugs. We focus on identifying and validating novel adverse events due to combination therapy. The resources and knowledge we produce will enable the development of predictive models of drug effects and potentially the design of safer therapies.

Agency
National Institute of Health (NIH)
Institute
National Institute of General Medical Sciences (NIGMS)
Type
Research Project (R01)
Project #
1R01GM107145-01A1
Application #
8696226
Study Section
Biodata Management and Analysis Study Section (BDMA)
Program Officer
Long, Rochelle M
Project Start
2014-08-01
Project End
2019-04-30
Budget Start
2014-08-01
Budget End
2015-04-30
Support Year
1
Fiscal Year
2014
Total Cost
$597,545
Indirect Cost
$220,544
Name
Columbia University (N.Y.)
Department
Internal Medicine/Medicine
Type
Schools of Medicine
DUNS #
621889815
City
New York
State
NY
Country
United States
Zip Code
10032
Polubriaginof, Fernanda C G; Vanguri, Rami; Quinnies, Kayla et al. (2018) Disease Heritability Inferred from Familial Relationships Reported in Medical Records. Cell 173:1692-1704.e11
Tatonetti, Nicholas P (2018) The Next Generation of Drug Safety Science: Coupling Detection, Corroboration, and Validation to Discover Novel Drug Effects and Drug-Drug Interactions. Clin Pharmacol Ther 103:177-179
Ta, Casey N; Dumontier, Michel; Hripcsak, George et al. (2018) Columbia Open Health Data, clinical concept prevalence and co-occurrence from electronic health records. Sci Data 5:180273
Boland, Mary Regina; Parhi, Pradipta; Gentine, Pierre et al. (2017) Climate Classification is an Important Factor in Assessing Quality-of-Care Across Hospitals. Sci Rep 7:4948
Nissim, Nir; Shahar, Yuval; Elovici, Yuval et al. (2017) Inter-labeler and intra-labeler variability of condition severity classification models using active and passive learning methods. Artif Intell Med 81:12-32
Boland, Mary Regina; Karczewski, Konrad J; Tatonetti, Nicholas P (2017) Ten Simple Rules to Enable Multi-site Collaborations through Data Sharing. PLoS Comput Biol 13:e1005278
Karczewski, Konrad J; Tatonetti, Nicholas P; Manrai, Arjun K et al. (2017) METHODS TO ENSURE THE REPRODUCIBILITY OF BIOMEDICAL RESEARCH. Pac Symp Biocomput 22:117-119
Xu, Katherine; Rosenstiel, Paul; Paragas, Neal et al. (2017) Unique Transcriptional Programs Identify Subtypes of AKI. J Am Soc Nephrol 28:1729-1740
Deng, Changchun; Lipstein, Mark R; Scotto, Luigi et al. (2017) Silencing c-Myc translation as a therapeutic strategy through targeting PI3K? and CK1? in hematological malignancies. Blood 129:88-99
Shameer, Khader; Johnson, Kipp W; Yahi, Alexandre et al. (2017) PREDICTIVE MODELING OF HOSPITAL READMISSION RATES USING ELECTRONIC MEDICAL RECORD-WIDE MACHINE LEARNING: A CASE-STUDY USING MOUNT SINAI HEART FAILURE COHORT. Pac Symp Biocomput 22:276-287

Showing the most recent 10 out of 46 publications