An approach to acquiring, normalizing, and managing EHR data from a clinical data repository for studying pressure ulcer outcomes

William Padula, Leon Blackshaw, C. Tod Brindle, Samuel L. Volchenboum

Research output: Contribution to journalArticle

Abstract

Changes in the methods that individual facilities follow to collect and store data related to hospital-acquired pressure ulcer (HAPU) occurrences are essential for improving patient outcomes and advancing our understanding the science behind this clinically relevant issue. Using an established electronic health record system at a large, urban, tertiary-care academic medical center, we investigated the process required for taking raw data of HAPU outcomes and submitting these data to a normalization process. We extracted data from 1.5 million patient shifts and filtered observations to those with a Braden score and linked tables in the electronic health record, including (1) Braden scale scores, (2) laboratory outcomes data, (3) surgical time, (4) provider orders, (5) medications, and (6) discharge diagnoses. Braden scores are important measures specific to HAPUs since these scores clarify the daily risk of a hospitalized patient for developing a pressure ulcer. The other more common measures that may be associated with HAPU outcomes are important to organize in a single data frame with Braden scores according to each patient. Primary keys were assigned to each table, and the data were processed through 3 normalization steps and 1 denormalization step. These processes created 8 tables that can be stored efficiently in a clinical database of HAPU outcomes. As hospitals focus on organizing data for review of HAPUs and other types of hospital-acquired conditions, the normalization process we describe in this article offers directions for collaboration between providers and informatics teams using a common language and structure.

Original languageEnglish (US)
Pages (from-to)39-45
Number of pages7
JournalJournal of Wound, Ostomy and Continence Nursing
Volume43
Issue number1
DOIs
StatePublished - 2016

Fingerprint

Pressure Ulcer
Electronic Health Records
Iatrogenic Disease
Informatics
Tertiary Healthcare
Operative Time
Language
Databases

Keywords

  • Data mining
  • Data warehouse
  • Informatics
  • Pressure ulcer

ASJC Scopus subject areas

  • Advanced and Specialized Nursing
  • Medical–Surgical

Cite this

An approach to acquiring, normalizing, and managing EHR data from a clinical data repository for studying pressure ulcer outcomes. / Padula, William; Blackshaw, Leon; Brindle, C. Tod; Volchenboum, Samuel L.

In: Journal of Wound, Ostomy and Continence Nursing, Vol. 43, No. 1, 2016, p. 39-45.

Research output: Contribution to journalArticle

Padula, William ; Blackshaw, Leon ; Brindle, C. Tod ; Volchenboum, Samuel L. / An approach to acquiring, normalizing, and managing EHR data from a clinical data repository for studying pressure ulcer outcomes. In: Journal of Wound, Ostomy and Continence Nursing. 2016 ; Vol. 43, No. 1. pp. 39-45.
@article{86f9e776cf3343ae9b939faa53519130,
title = "An approach to acquiring, normalizing, and managing EHR data from a clinical data repository for studying pressure ulcer outcomes",
abstract = "Changes in the methods that individual facilities follow to collect and store data related to hospital-acquired pressure ulcer (HAPU) occurrences are essential for improving patient outcomes and advancing our understanding the science behind this clinically relevant issue. Using an established electronic health record system at a large, urban, tertiary-care academic medical center, we investigated the process required for taking raw data of HAPU outcomes and submitting these data to a normalization process. We extracted data from 1.5 million patient shifts and filtered observations to those with a Braden score and linked tables in the electronic health record, including (1) Braden scale scores, (2) laboratory outcomes data, (3) surgical time, (4) provider orders, (5) medications, and (6) discharge diagnoses. Braden scores are important measures specific to HAPUs since these scores clarify the daily risk of a hospitalized patient for developing a pressure ulcer. The other more common measures that may be associated with HAPU outcomes are important to organize in a single data frame with Braden scores according to each patient. Primary keys were assigned to each table, and the data were processed through 3 normalization steps and 1 denormalization step. These processes created 8 tables that can be stored efficiently in a clinical database of HAPU outcomes. As hospitals focus on organizing data for review of HAPUs and other types of hospital-acquired conditions, the normalization process we describe in this article offers directions for collaboration between providers and informatics teams using a common language and structure.",
keywords = "Data mining, Data warehouse, Informatics, Pressure ulcer",
author = "William Padula and Leon Blackshaw and Brindle, {C. Tod} and Volchenboum, {Samuel L.}",
year = "2016",
doi = "10.1097/WON.0000000000000185",
language = "English (US)",
volume = "43",
pages = "39--45",
journal = "Journal of Wound, Ostomy and Continence Nursing",
issn = "1071-5754",
publisher = "Lippincott Williams and Wilkins",
number = "1",

}

TY - JOUR

T1 - An approach to acquiring, normalizing, and managing EHR data from a clinical data repository for studying pressure ulcer outcomes

AU - Padula, William

AU - Blackshaw, Leon

AU - Brindle, C. Tod

AU - Volchenboum, Samuel L.

PY - 2016

Y1 - 2016

N2 - Changes in the methods that individual facilities follow to collect and store data related to hospital-acquired pressure ulcer (HAPU) occurrences are essential for improving patient outcomes and advancing our understanding the science behind this clinically relevant issue. Using an established electronic health record system at a large, urban, tertiary-care academic medical center, we investigated the process required for taking raw data of HAPU outcomes and submitting these data to a normalization process. We extracted data from 1.5 million patient shifts and filtered observations to those with a Braden score and linked tables in the electronic health record, including (1) Braden scale scores, (2) laboratory outcomes data, (3) surgical time, (4) provider orders, (5) medications, and (6) discharge diagnoses. Braden scores are important measures specific to HAPUs since these scores clarify the daily risk of a hospitalized patient for developing a pressure ulcer. The other more common measures that may be associated with HAPU outcomes are important to organize in a single data frame with Braden scores according to each patient. Primary keys were assigned to each table, and the data were processed through 3 normalization steps and 1 denormalization step. These processes created 8 tables that can be stored efficiently in a clinical database of HAPU outcomes. As hospitals focus on organizing data for review of HAPUs and other types of hospital-acquired conditions, the normalization process we describe in this article offers directions for collaboration between providers and informatics teams using a common language and structure.

AB - Changes in the methods that individual facilities follow to collect and store data related to hospital-acquired pressure ulcer (HAPU) occurrences are essential for improving patient outcomes and advancing our understanding the science behind this clinically relevant issue. Using an established electronic health record system at a large, urban, tertiary-care academic medical center, we investigated the process required for taking raw data of HAPU outcomes and submitting these data to a normalization process. We extracted data from 1.5 million patient shifts and filtered observations to those with a Braden score and linked tables in the electronic health record, including (1) Braden scale scores, (2) laboratory outcomes data, (3) surgical time, (4) provider orders, (5) medications, and (6) discharge diagnoses. Braden scores are important measures specific to HAPUs since these scores clarify the daily risk of a hospitalized patient for developing a pressure ulcer. The other more common measures that may be associated with HAPU outcomes are important to organize in a single data frame with Braden scores according to each patient. Primary keys were assigned to each table, and the data were processed through 3 normalization steps and 1 denormalization step. These processes created 8 tables that can be stored efficiently in a clinical database of HAPU outcomes. As hospitals focus on organizing data for review of HAPUs and other types of hospital-acquired conditions, the normalization process we describe in this article offers directions for collaboration between providers and informatics teams using a common language and structure.

KW - Data mining

KW - Data warehouse

KW - Informatics

KW - Pressure ulcer

UR - http://www.scopus.com/inward/record.url?scp=84954176335&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84954176335&partnerID=8YFLogxK

U2 - 10.1097/WON.0000000000000185

DO - 10.1097/WON.0000000000000185

M3 - Article

C2 - 26727681

AN - SCOPUS:84954176335

VL - 43

SP - 39

EP - 45

JO - Journal of Wound, Ostomy and Continence Nursing

JF - Journal of Wound, Ostomy and Continence Nursing

SN - 1071-5754

IS - 1

ER -