Comparison of two national databases for general thoracic surgery

Mark S. Allen, Shanda Blackmon, Francis C. Nichols, Stephen D. Cassivi, K. Robert Shen, Dennis A Wigle

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Background Improving the quality of surgical care through accurate measurement of outcomes is an important endeavor. The purpose of this study was to compare data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) and The Society of Thoracic Surgeons (STS) general thoracic surgery database to determine if a sampling technique (ACS NSQIP) is as effective and accurate as the comprehensive technique (STS database). Methods A common data abstractor collected and recorded data for the ACS NSQIP and STS database from our institution for the year 2012. The data was completely deidentified and analyzed for demographics, preoperative risk factors, mortality, and morbidity. Results The STS database recorded 1,595 (100%) operations for the year 2012, whereas the ACS NSQIP by design collects a limited sample and recorded 308 (19.3%) operations. Postoperative events were recorded in 17.2% of ACS NSQIP operations and in 30.1% of operations reported in the STS database. As more specific operations are examined, errors in the NSQIP data increase significantly. For example, the ACS NSQIP underestimated the pneumonia rate for lobectomy (5.9% versus 10.9%) and overestimated the pneumonia rate for an Ivor Lewis esophagogastrectomy (23.8% vs 18.8%). When the ACS NSQIP was used to compare our institution to the ACS NSQIP national norms, our institution was ranked in the lowest eighth decile for 30-day operative mortality; however, we were better than average when using STS database data (1.2% [2 of 162 procedures] vs 1.4% [538 of 37,324 procedures]) for pulmonary resections and 3.0% (3 of 100 procedures) vs 3.6% [138 of 3,865 procedures] for esophagectomy). Conclusions Databases built on partial sampling that do not capture all patients, such as the ACS NSQIP, may be useful for global analyses, but fall short of providing a foundation for meaningful quality improvement initiatives when analyzing data for specific thoracic surgical operations. These results highlight the utility and importance of complete databases such as the STSDB. National comparisons of clinical outcomes for thoracic surgical procedures should be interpreted with caution when using partial databases.

Original languageEnglish (US)
Pages (from-to)1155-1162
Number of pages8
JournalAnnals of Thoracic Surgery
Volume100
Issue number4
DOIs
StatePublished - 2015

Fingerprint

Thoracic Surgery
Quality Improvement
Databases
Thorax
Surgeons
Pneumonia
Thoracic Surgical Procedures
Esophagectomy
Mortality
Quality of Health Care
Demography
Morbidity
Lung

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine
  • Surgery
  • Pulmonary and Respiratory Medicine

Cite this

Comparison of two national databases for general thoracic surgery. / Allen, Mark S.; Blackmon, Shanda; Nichols, Francis C.; Cassivi, Stephen D.; Shen, K. Robert; Wigle, Dennis A.

In: Annals of Thoracic Surgery, Vol. 100, No. 4, 2015, p. 1155-1162.

Research output: Contribution to journalArticle

Allen, MS, Blackmon, S, Nichols, FC, Cassivi, SD, Shen, KR & Wigle, DA 2015, 'Comparison of two national databases for general thoracic surgery', Annals of Thoracic Surgery, vol. 100, no. 4, pp. 1155-1162. https://doi.org/10.1016/j.athoracsur.2015.05.031
Allen, Mark S. ; Blackmon, Shanda ; Nichols, Francis C. ; Cassivi, Stephen D. ; Shen, K. Robert ; Wigle, Dennis A. / Comparison of two national databases for general thoracic surgery. In: Annals of Thoracic Surgery. 2015 ; Vol. 100, No. 4. pp. 1155-1162.
@article{ae07c7b0f5fe4f95af2ad58e9db8b227,
title = "Comparison of two national databases for general thoracic surgery",
abstract = "Background Improving the quality of surgical care through accurate measurement of outcomes is an important endeavor. The purpose of this study was to compare data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) and The Society of Thoracic Surgeons (STS) general thoracic surgery database to determine if a sampling technique (ACS NSQIP) is as effective and accurate as the comprehensive technique (STS database). Methods A common data abstractor collected and recorded data for the ACS NSQIP and STS database from our institution for the year 2012. The data was completely deidentified and analyzed for demographics, preoperative risk factors, mortality, and morbidity. Results The STS database recorded 1,595 (100{\%}) operations for the year 2012, whereas the ACS NSQIP by design collects a limited sample and recorded 308 (19.3{\%}) operations. Postoperative events were recorded in 17.2{\%} of ACS NSQIP operations and in 30.1{\%} of operations reported in the STS database. As more specific operations are examined, errors in the NSQIP data increase significantly. For example, the ACS NSQIP underestimated the pneumonia rate for lobectomy (5.9{\%} versus 10.9{\%}) and overestimated the pneumonia rate for an Ivor Lewis esophagogastrectomy (23.8{\%} vs 18.8{\%}). When the ACS NSQIP was used to compare our institution to the ACS NSQIP national norms, our institution was ranked in the lowest eighth decile for 30-day operative mortality; however, we were better than average when using STS database data (1.2{\%} [2 of 162 procedures] vs 1.4{\%} [538 of 37,324 procedures]) for pulmonary resections and 3.0{\%} (3 of 100 procedures) vs 3.6{\%} [138 of 3,865 procedures] for esophagectomy). Conclusions Databases built on partial sampling that do not capture all patients, such as the ACS NSQIP, may be useful for global analyses, but fall short of providing a foundation for meaningful quality improvement initiatives when analyzing data for specific thoracic surgical operations. These results highlight the utility and importance of complete databases such as the STSDB. National comparisons of clinical outcomes for thoracic surgical procedures should be interpreted with caution when using partial databases.",
author = "Allen, {Mark S.} and Shanda Blackmon and Nichols, {Francis C.} and Cassivi, {Stephen D.} and Shen, {K. Robert} and Wigle, {Dennis A}",
year = "2015",
doi = "10.1016/j.athoracsur.2015.05.031",
language = "English (US)",
volume = "100",
pages = "1155--1162",
journal = "Annals of Thoracic Surgery",
issn = "0003-4975",
publisher = "Elsevier USA",
number = "4",

}

TY - JOUR

T1 - Comparison of two national databases for general thoracic surgery

AU - Allen, Mark S.

AU - Blackmon, Shanda

AU - Nichols, Francis C.

AU - Cassivi, Stephen D.

AU - Shen, K. Robert

AU - Wigle, Dennis A

PY - 2015

Y1 - 2015

N2 - Background Improving the quality of surgical care through accurate measurement of outcomes is an important endeavor. The purpose of this study was to compare data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) and The Society of Thoracic Surgeons (STS) general thoracic surgery database to determine if a sampling technique (ACS NSQIP) is as effective and accurate as the comprehensive technique (STS database). Methods A common data abstractor collected and recorded data for the ACS NSQIP and STS database from our institution for the year 2012. The data was completely deidentified and analyzed for demographics, preoperative risk factors, mortality, and morbidity. Results The STS database recorded 1,595 (100%) operations for the year 2012, whereas the ACS NSQIP by design collects a limited sample and recorded 308 (19.3%) operations. Postoperative events were recorded in 17.2% of ACS NSQIP operations and in 30.1% of operations reported in the STS database. As more specific operations are examined, errors in the NSQIP data increase significantly. For example, the ACS NSQIP underestimated the pneumonia rate for lobectomy (5.9% versus 10.9%) and overestimated the pneumonia rate for an Ivor Lewis esophagogastrectomy (23.8% vs 18.8%). When the ACS NSQIP was used to compare our institution to the ACS NSQIP national norms, our institution was ranked in the lowest eighth decile for 30-day operative mortality; however, we were better than average when using STS database data (1.2% [2 of 162 procedures] vs 1.4% [538 of 37,324 procedures]) for pulmonary resections and 3.0% (3 of 100 procedures) vs 3.6% [138 of 3,865 procedures] for esophagectomy). Conclusions Databases built on partial sampling that do not capture all patients, such as the ACS NSQIP, may be useful for global analyses, but fall short of providing a foundation for meaningful quality improvement initiatives when analyzing data for specific thoracic surgical operations. These results highlight the utility and importance of complete databases such as the STSDB. National comparisons of clinical outcomes for thoracic surgical procedures should be interpreted with caution when using partial databases.

AB - Background Improving the quality of surgical care through accurate measurement of outcomes is an important endeavor. The purpose of this study was to compare data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) and The Society of Thoracic Surgeons (STS) general thoracic surgery database to determine if a sampling technique (ACS NSQIP) is as effective and accurate as the comprehensive technique (STS database). Methods A common data abstractor collected and recorded data for the ACS NSQIP and STS database from our institution for the year 2012. The data was completely deidentified and analyzed for demographics, preoperative risk factors, mortality, and morbidity. Results The STS database recorded 1,595 (100%) operations for the year 2012, whereas the ACS NSQIP by design collects a limited sample and recorded 308 (19.3%) operations. Postoperative events were recorded in 17.2% of ACS NSQIP operations and in 30.1% of operations reported in the STS database. As more specific operations are examined, errors in the NSQIP data increase significantly. For example, the ACS NSQIP underestimated the pneumonia rate for lobectomy (5.9% versus 10.9%) and overestimated the pneumonia rate for an Ivor Lewis esophagogastrectomy (23.8% vs 18.8%). When the ACS NSQIP was used to compare our institution to the ACS NSQIP national norms, our institution was ranked in the lowest eighth decile for 30-day operative mortality; however, we were better than average when using STS database data (1.2% [2 of 162 procedures] vs 1.4% [538 of 37,324 procedures]) for pulmonary resections and 3.0% (3 of 100 procedures) vs 3.6% [138 of 3,865 procedures] for esophagectomy). Conclusions Databases built on partial sampling that do not capture all patients, such as the ACS NSQIP, may be useful for global analyses, but fall short of providing a foundation for meaningful quality improvement initiatives when analyzing data for specific thoracic surgical operations. These results highlight the utility and importance of complete databases such as the STSDB. National comparisons of clinical outcomes for thoracic surgical procedures should be interpreted with caution when using partial databases.

UR - http://www.scopus.com/inward/record.url?scp=84952715859&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84952715859&partnerID=8YFLogxK

U2 - 10.1016/j.athoracsur.2015.05.031

DO - 10.1016/j.athoracsur.2015.05.031

M3 - Article

C2 - 26319486

AN - SCOPUS:84952715859

VL - 100

SP - 1155

EP - 1162

JO - Annals of Thoracic Surgery

JF - Annals of Thoracic Surgery

SN - 0003-4975

IS - 4

ER -