These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


Pubmed for Handhelds

PUBMED FOR HANDHELDS

Search MEDLINE/PubMed

  • Title: Missing data and interpretation of cancer surgery outcomes at the American College of Surgeons National Surgical Quality Improvement Program.
    Author: Parsons HM, Henderson WG, Ziegenfuss JY, Davern M, Al-Refaie WB.
    Journal: J Am Coll Surg; 2011 Sep; 213(3):379-91. PubMed ID: 21700480.
    Abstract:
    BACKGROUND: The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) has become an important surgical quality program in the United States, yet few studies describe their methods for handling missing data. Our study examines the impact of missing data on predictive models for short-term operative outcomes after cancer surgery in the ACS NSQIP database. STUDY DESIGN: We identified 97,230 patients who underwent oncologic resections for neoplasms in the 2005-2009 ACS NSQIP. We used multivariable logistic regression to assess the impact of pre-, intra-, and postoperative factors on short-term operative outcomes by type of procedure where missing values were included as a variable category, excluded, and imputed. RESULTS: A large proportion (72.8%) of patients had one or more missing pre-, intra-, or postoperative characteristics, particularly preoperative laboratory values. Missing data were more frequent in healthier patients and those undergoing lower-risk procedures. Although data were not missing at random, the impact of preoperative risk factors on adverse operative outcomes after cancer surgery was similar across methods for handling missing data. However, analytic approaches using only patients with complete or imputed information risk basing the analysis on a potentially nonrepresentative sample. CONCLUSIONS: Missing data present challenges to interpreting predictors of short-term operative outcomes after cancer surgery at ACS NSQIP hospitals. Similar to best practices for other data sets, this study highlights the importance of using missing values carefully when using ACS NSQIP. Given its potential to introduce bias, the approach to handling missing values should be detailed in future ACS NSQIP studies.
    [Abstract] [Full Text] [Related] [New Search]