Identifying heart failure using EMR-based algorithms

Geoffrey H. Tison, Alanna Chamberlain, Mark J. Pletcher, Shannon M Dunlay, Susan A. Weston, Jill M. Killian, Jeffrey E. Olgin, Veronique Lee Roger

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Background: Heart failure (HF) is a major clinical and public health problem, the management of which will benefit from large-scale pragmatic research that leverages electronic medical records (EMR). Requisite to using EMRs for HF research is the development of reliable algorithms to identify HF patients. We aimed to develop and validate computable phenotype algorithms to identify patients with HF using standardized data elements defined by the Patient Centered Outcomes Research Network (PCORnet) Common Data Model (CDM). Methods: We built HF computable phenotypes utilizing the data domains of HF diagnosis codes, prescribed HF-related medications and N-terminal B-type natriuretic peptide (NT-proBNP). Algorithms were validated in a cohort (n = 76,254) drawn from Olmsted County, MN between 2010–2012 a sample of whose records were manually reviewed to confirm HF according to Framingham criteria. Results: The different algorithms we tested provided different tradeoffs between sensitivity and positive predictive value (PPV). The highest sensitivity (78.7%) algorithm utilized one HF diagnosis code and had the lowest PPV (68.5%). The addition of more algorithm components, such as additional HF diagnosis codes, HF medications or elevated NT-proBNP, improved the PPV while reducing sensitivity. When added to a diagnostic code, the addition of NT-proBNP (>450 pg/mL) had a similar impact compared to additional HF medication criteria, increasing PPV by ∼3–4% and decreasing sensitivity by ∼7–10%. Conclusions: Algorithms derived from PCORnet CDM elements can be used to identify patients with HF without manual adjudication with reasonable sensitivity and PPV. Algorithm choice should be driven by the goal of the research.

Original languageEnglish (US)
Pages (from-to)1-7
Number of pages7
JournalInternational Journal of Medical Informatics
Volume120
DOIs
StatePublished - Dec 1 2018

Fingerprint

Electronic Health Records
Heart Failure
Patient Outcome Assessment
Research
Phenotype
Brain Natriuretic Peptide
Public Health

Keywords

  • Cohort studies
  • Electronic health records
  • Heart failure
  • Learning health system
  • Outcomes assessment

ASJC Scopus subject areas

  • Health Informatics

Cite this

Identifying heart failure using EMR-based algorithms. / Tison, Geoffrey H.; Chamberlain, Alanna; Pletcher, Mark J.; Dunlay, Shannon M; Weston, Susan A.; Killian, Jill M.; Olgin, Jeffrey E.; Roger, Veronique Lee.

In: International Journal of Medical Informatics, Vol. 120, 01.12.2018, p. 1-7.

Research output: Contribution to journalArticle

Tison, Geoffrey H. ; Chamberlain, Alanna ; Pletcher, Mark J. ; Dunlay, Shannon M ; Weston, Susan A. ; Killian, Jill M. ; Olgin, Jeffrey E. ; Roger, Veronique Lee. / Identifying heart failure using EMR-based algorithms. In: International Journal of Medical Informatics. 2018 ; Vol. 120. pp. 1-7.
@article{c881d5dfcdb04f5794420e4cac3e602d,
title = "Identifying heart failure using EMR-based algorithms",
abstract = "Background: Heart failure (HF) is a major clinical and public health problem, the management of which will benefit from large-scale pragmatic research that leverages electronic medical records (EMR). Requisite to using EMRs for HF research is the development of reliable algorithms to identify HF patients. We aimed to develop and validate computable phenotype algorithms to identify patients with HF using standardized data elements defined by the Patient Centered Outcomes Research Network (PCORnet) Common Data Model (CDM). Methods: We built HF computable phenotypes utilizing the data domains of HF diagnosis codes, prescribed HF-related medications and N-terminal B-type natriuretic peptide (NT-proBNP). Algorithms were validated in a cohort (n = 76,254) drawn from Olmsted County, MN between 2010–2012 a sample of whose records were manually reviewed to confirm HF according to Framingham criteria. Results: The different algorithms we tested provided different tradeoffs between sensitivity and positive predictive value (PPV). The highest sensitivity (78.7{\%}) algorithm utilized one HF diagnosis code and had the lowest PPV (68.5{\%}). The addition of more algorithm components, such as additional HF diagnosis codes, HF medications or elevated NT-proBNP, improved the PPV while reducing sensitivity. When added to a diagnostic code, the addition of NT-proBNP (>450 pg/mL) had a similar impact compared to additional HF medication criteria, increasing PPV by ∼3–4{\%} and decreasing sensitivity by ∼7–10{\%}. Conclusions: Algorithms derived from PCORnet CDM elements can be used to identify patients with HF without manual adjudication with reasonable sensitivity and PPV. Algorithm choice should be driven by the goal of the research.",
keywords = "Cohort studies, Electronic health records, Heart failure, Learning health system, Outcomes assessment",
author = "Tison, {Geoffrey H.} and Alanna Chamberlain and Pletcher, {Mark J.} and Dunlay, {Shannon M} and Weston, {Susan A.} and Killian, {Jill M.} and Olgin, {Jeffrey E.} and Roger, {Veronique Lee}",
year = "2018",
month = "12",
day = "1",
doi = "10.1016/j.ijmedinf.2018.09.016",
language = "English (US)",
volume = "120",
pages = "1--7",
journal = "International Journal of Medical Informatics",
issn = "1386-5056",
publisher = "Elsevier Ireland Ltd",

}

TY - JOUR

T1 - Identifying heart failure using EMR-based algorithms

AU - Tison, Geoffrey H.

AU - Chamberlain, Alanna

AU - Pletcher, Mark J.

AU - Dunlay, Shannon M

AU - Weston, Susan A.

AU - Killian, Jill M.

AU - Olgin, Jeffrey E.

AU - Roger, Veronique Lee

PY - 2018/12/1

Y1 - 2018/12/1

N2 - Background: Heart failure (HF) is a major clinical and public health problem, the management of which will benefit from large-scale pragmatic research that leverages electronic medical records (EMR). Requisite to using EMRs for HF research is the development of reliable algorithms to identify HF patients. We aimed to develop and validate computable phenotype algorithms to identify patients with HF using standardized data elements defined by the Patient Centered Outcomes Research Network (PCORnet) Common Data Model (CDM). Methods: We built HF computable phenotypes utilizing the data domains of HF diagnosis codes, prescribed HF-related medications and N-terminal B-type natriuretic peptide (NT-proBNP). Algorithms were validated in a cohort (n = 76,254) drawn from Olmsted County, MN between 2010–2012 a sample of whose records were manually reviewed to confirm HF according to Framingham criteria. Results: The different algorithms we tested provided different tradeoffs between sensitivity and positive predictive value (PPV). The highest sensitivity (78.7%) algorithm utilized one HF diagnosis code and had the lowest PPV (68.5%). The addition of more algorithm components, such as additional HF diagnosis codes, HF medications or elevated NT-proBNP, improved the PPV while reducing sensitivity. When added to a diagnostic code, the addition of NT-proBNP (>450 pg/mL) had a similar impact compared to additional HF medication criteria, increasing PPV by ∼3–4% and decreasing sensitivity by ∼7–10%. Conclusions: Algorithms derived from PCORnet CDM elements can be used to identify patients with HF without manual adjudication with reasonable sensitivity and PPV. Algorithm choice should be driven by the goal of the research.

AB - Background: Heart failure (HF) is a major clinical and public health problem, the management of which will benefit from large-scale pragmatic research that leverages electronic medical records (EMR). Requisite to using EMRs for HF research is the development of reliable algorithms to identify HF patients. We aimed to develop and validate computable phenotype algorithms to identify patients with HF using standardized data elements defined by the Patient Centered Outcomes Research Network (PCORnet) Common Data Model (CDM). Methods: We built HF computable phenotypes utilizing the data domains of HF diagnosis codes, prescribed HF-related medications and N-terminal B-type natriuretic peptide (NT-proBNP). Algorithms were validated in a cohort (n = 76,254) drawn from Olmsted County, MN between 2010–2012 a sample of whose records were manually reviewed to confirm HF according to Framingham criteria. Results: The different algorithms we tested provided different tradeoffs between sensitivity and positive predictive value (PPV). The highest sensitivity (78.7%) algorithm utilized one HF diagnosis code and had the lowest PPV (68.5%). The addition of more algorithm components, such as additional HF diagnosis codes, HF medications or elevated NT-proBNP, improved the PPV while reducing sensitivity. When added to a diagnostic code, the addition of NT-proBNP (>450 pg/mL) had a similar impact compared to additional HF medication criteria, increasing PPV by ∼3–4% and decreasing sensitivity by ∼7–10%. Conclusions: Algorithms derived from PCORnet CDM elements can be used to identify patients with HF without manual adjudication with reasonable sensitivity and PPV. Algorithm choice should be driven by the goal of the research.

KW - Cohort studies

KW - Electronic health records

KW - Heart failure

KW - Learning health system

KW - Outcomes assessment

UR - http://www.scopus.com/inward/record.url?scp=85053826591&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053826591&partnerID=8YFLogxK

U2 - 10.1016/j.ijmedinf.2018.09.016

DO - 10.1016/j.ijmedinf.2018.09.016

M3 - Article

C2 - 30409334

AN - SCOPUS:85053826591

VL - 120

SP - 1

EP - 7

JO - International Journal of Medical Informatics

JF - International Journal of Medical Informatics

SN - 1386-5056

ER -