Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Risk Prediction With Electronic Health Records

Risk Prediction With Electronic Health Records Opinion EDITORIAL The Importance of Model Validation and Clinical Context Benjamin A. Goldstein, PhD; Ann Marie Navar, MD, PhD; Michael J. Pencina, PhD Historically, risk models have been derived using data from algorithm for identification of atrial fibrillation (AF) cases large epidemiologic cohorts or clinical trials. Although these adopted by Kolek et al incorporated billing codes, echocardio- data sources are often high quality, their external generaliz- graphic analyses, and natural language processing, a sophis- ability may be limited for at least 2 reasons. First, the popula- ticated approach that is unlikely to be easily replicated across tions included in the cohort all EHRs and for all conditions to be studied. or trial are often narrowly de- Even using their advanced EHR-derived phenotype, the Related article page 1007 fined and not representative authors noted high degrees of miscalibration among high- of all adults. Recent efforts to and low-risk individuals, which underpredicted risk in the low- combine data from multiple cohorts have led to risk predic- est-risk groups and overpredicted risk in the highest-risk tion models with broader external generalizability. The pooled groups. The authors concluded that the performance of the cohort equations used in the 2013 American College of Cardi- http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JAMA Cardiology American Medical Association

Risk Prediction With Electronic Health Records

Download PDF
2 pages

Loading next page...
 
/lp/american-medical-association/risk-prediction-with-electronic-health-records-aDniyq04CM

References (14)

Publisher
American Medical Association
Copyright
Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
ISSN
2380-6583
eISSN
2380-6591
DOI
10.1001/jamacardio.2016.3826
Publisher site
See Article on Publisher Site

Abstract

Opinion EDITORIAL The Importance of Model Validation and Clinical Context Benjamin A. Goldstein, PhD; Ann Marie Navar, MD, PhD; Michael J. Pencina, PhD Historically, risk models have been derived using data from algorithm for identification of atrial fibrillation (AF) cases large epidemiologic cohorts or clinical trials. Although these adopted by Kolek et al incorporated billing codes, echocardio- data sources are often high quality, their external generaliz- graphic analyses, and natural language processing, a sophis- ability may be limited for at least 2 reasons. First, the popula- ticated approach that is unlikely to be easily replicated across tions included in the cohort all EHRs and for all conditions to be studied. or trial are often narrowly de- Even using their advanced EHR-derived phenotype, the Related article page 1007 fined and not representative authors noted high degrees of miscalibration among high- of all adults. Recent efforts to and low-risk individuals, which underpredicted risk in the low- combine data from multiple cohorts have led to risk predic- est-risk groups and overpredicted risk in the highest-risk tion models with broader external generalizability. The pooled groups. The authors concluded that the performance of the cohort equations used in the 2013 American College of Cardi-

Journal

JAMA Cardiology – American Medical Association

Published: Dec 12, 2016

There are no references for this article.