Literature DB >> 33712852

Electronic phenotyping of health outcomes of interest using a linked claims-electronic health record database: Findings from a machine learning pilot project.

Teresa B Gibson1, Michael D Nguyen2, Timothy Burrell1, Frank Yoon1, Jenna Wong3, Sai Dharmarajan4, Rita Ouellet-Hellstrom5, Wei Hua2, Yong Ma6, Elande Baro7, Sarah Bloemers1, Cory Pack1, Adee Kennedy3, Sengwee Toh3, Robert Ball8.   

Abstract

OBJECTIVE: Claims-based algorithms are used in the Food and Drug Administration Sentinel Active Risk Identification and Analysis System to identify occurrences of health outcomes of interest (HOIs) for medical product safety assessment. This project aimed to apply machine learning classification techniques to demonstrate the feasibility of developing a claims-based algorithm to predict an HOI in structured electronic health record (EHR) data.
MATERIALS AND METHODS: We used the 2015-2019 IBM MarketScan Explorys Claims-EMR Data Set, linking administrative claims and EHR data at the patient level. We focused on a single HOI, rhabdomyolysis, defined by EHR laboratory test results. Using claims-based predictors, we applied machine learning techniques to predict the HOI: logistic regression, LASSO (least absolute shrinkage and selection operator), random forests, support vector machines, artificial neural nets, and an ensemble method (Super Learner).
RESULTS: The study cohort included 32 956 patients and 39 499 encounters. Model performance (positive predictive value [PPV], sensitivity, specificity, area under the receiver-operating characteristic curve) varied considerably across techniques. The area under the receiver-operating characteristic curve exceeded 0.80 in most model variations. DISCUSSION: For the main Food and Drug Administration use case of assessing risk of rhabdomyolysis after drug use, a model with a high PPV is typically preferred. The Super Learner ensemble model without adjustment for class imbalance achieved a PPV of 75.6%, substantially better than a previously used human expert-developed model (PPV = 44.0%).
CONCLUSIONS: It is feasible to use machine learning methods to predict an EHR-derived HOI with claims-based predictors. Modeling strategies can be adapted for intended uses, including surveillance, identification of cases for chart review, and outcomes research.
© The Author(s) 2021. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For permissions, please email: journals.permissions@oup.com.

Entities:  

Keywords:  administrative claims; electronic health records; electronic phenotyping; healthcare; rhabdomyolysis; supervised machine learning

Mesh:

Year:  2021        PMID: 33712852      PMCID: PMC8279790          DOI: 10.1093/jamia/ocab036

Source DB:  PubMed          Journal:  J Am Med Inform Assoc        ISSN: 1067-5027            Impact factor:   4.497


  28 in total

1.  The FDA Sentinel Initiative - An Evolving National Resource.

Authors:  Richard Platt; Jeffrey S Brown; Melissa Robb; Mark McClellan; Robert Ball; Michael D Nguyen; Rachel E Sherman
Journal:  N Engl J Med       Date:  2018-11-29       Impact factor: 91.245

Review 2.  Rhabdomyolysis: pathogenesis, diagnosis, and treatment.

Authors:  Patrick A Torres; John A Helmstetter; Adam M Kaye; Alan David Kaye
Journal:  Ochsner J       Date:  2015

3.  Association of aspartate aminotransferase in statin-induced rhabdomyolysis.

Authors:  Xu Cong Ruan; Lian Leng Low; Yu Heng Kwan
Journal:  J Prim Health Care       Date:  2017-12

4.  Diagnostic markers of acute myocardial infarction.

Authors:  Sabesan Mythili; Narasimhan Malathi
Journal:  Biomed Rep       Date:  2015-07-29

5.  Advances in Electronic Phenotyping: From Rule-Based Definitions to Machine Learning Models.

Authors:  Juan M Banda; Martin Seneviratne; Tina Hernandez-Boussard; Nigam H Shah
Journal:  Annu Rev Biomed Data Sci       Date:  2018-05-23

Review 6.  Rhabdomyolysis: a review of the literature.

Authors:  F Y Khan
Journal:  Neth J Med       Date:  2009-10       Impact factor: 1.422

7.  Adapting electronic health records-derived phenotypes to claims data: Lessons learned in using limited clinical data for phenotyping.

Authors:  Anna Ostropolets; Christian Reich; Patrick Ryan; Ning Shang; George Hripcsak; Chunhua Weng
Journal:  J Biomed Inform       Date:  2019-12-19       Impact factor: 6.317

8.  Rhabdomyolysis: Patterns, Circumstances, and Outcomes of Patients Presenting to the Emergency Department.

Authors:  Emily G Knafl; James A Hughes; Goce Dimeski; Rob Eley
Journal:  Ochsner J       Date:  2018

Review 9.  Beyond muscle destruction: a systematic review of rhabdomyolysis for clinical practice.

Authors:  Luis O Chavez; Monica Leon; Sharon Einav; Joseph Varon
Journal:  Crit Care       Date:  2016-06-15       Impact factor: 9.097

10.  Rhabdomyolysis with different etiologies in childhood.

Authors:  Demet Alaygut; Meral Torun Bayram; Belde Kasap; Alper Soylu; Mehmet Türkmen; Salih Kavukcu
Journal:  World J Clin Pediatr       Date:  2017-11-08
View more
  3 in total

1.  Pharmacovigilance and Pharmacoepidemiology as a Guarantee of Patient Safety: The Role of the Clinical Pharmacologist.

Authors:  Giada Crescioli; Roberto Bonaiuti; Renato Corradetti; Guido Mannaioni; Alfredo Vannacci; Niccolò Lombardi
Journal:  J Clin Med       Date:  2022-06-20       Impact factor: 4.964

2.  "Artificial Intelligence" for Pharmacovigilance: Ready for Prime Time?

Authors:  Robert Ball; Gerald Dal Pan
Journal:  Drug Saf       Date:  2022-05-17       Impact factor: 5.228

Review 3.  Broadening the reach of the FDA Sentinel system: A roadmap for integrating electronic health record data in a causal analysis framework.

Authors:  Rishi J Desai; Michael E Matheny; Kevin Johnson; Keith Marsolo; Lesley H Curtis; Jennifer C Nelson; Patrick J Heagerty; Judith Maro; Jeffery Brown; Sengwee Toh; Michael Nguyen; Robert Ball; Gerald Dal Pan; Shirley V Wang; Joshua J Gagne; Sebastian Schneeweiss
Journal:  NPJ Digit Med       Date:  2021-12-20
  3 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.