Predicting Prolonged Length of Stay after Elective Colectomy: Development of a Clinical Decision Support System Using ACS NSQIP Data Pubblico

Abstract

The objective of this project was to develop a Clinical Decision Support System (CDSS) tool to assist surgical teams in postoperative and discharge decision-making at the point of care. Risk factors associated with prolonged postoperative length of stay (pLOS) following colectomy have not been validated nor used in the development of predictive models. CDSS help physicians better integrate real-time clinical data when making decisions; like when and how to discharge complex surgical patients. No tools currently exist to help physicians make evidence-based decisions regarding length of stay and discharge for patients after elective colectomy. This was a retrospective analysis of American College of Surgeons National Surgical Quality Improvement Program data. We determined factors significantly associated with pLOS at our main academic hospital, tested the performance of these factors on an independent cohort via logistic regression modeling, and developed a clinical risk scoring system for pLOS (the pLOS Risk Score). Demographic variables associated with pLOS include age, disseminated cancer, ≥ 3 comorbidities, prior abdominal surgery, and preoperative admission > 1 day. Included laboratory and intraoperative risk factors were elevated international normalized ratio, operative time, blood loss, and open approach to colectomy. External validation of the model yielded an area under the ROC curve of 0.81 and allowed us to predict pLOS with 59% sensitivity, 85% specificity, and 77% accuracy using a cut point of > 24% predicted risk. Prolonged length of stay following elective colectomy can be accurately predicted and translated in to a useful CDSS tool.

Table of Contents

INTRODUCTION..................................................................1

BACKGROUND.....................................................................1

METHODS..........................................................................4

Data source and patient selection..........................................4

Outcome of interest and risk factors ......................................5

Risk prediction model...........................................................6

Points system....................................................................7

Formula 1..........................................................................8

RESULTS...........................................................................8

CONCLUSIONS...................................................................12

REFERENCES.....................................................................20

TABLES AND FIGURES.........................................................21

Figure 1...........................................................................21

Figure 2...........................................................................22

Table 1............................................................................23

Table 2............................................................................24

Table 3............................................................................25

Table 4............................................................................26

Table 5............................................................................26

Figure 2...........................................................................27

Table 6 .................................... ......... ...............................28

Table 7............................................................................28

Table 8............................................................................29

Table 9............................................................................30

Table 10..........................................................................31

Table 11..........................................................................31

Figure 3...........................................................................32

Table 12..........................................................................32

Figure 4...........................................................................33

About this Master's Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Laney Graduate School
Department	Clinical Research Curriculum Program
Degree	MS
Submission	Master's Thesis
Language	English
Research Field	Health Sciences, Medicine and Surgery
Parola chiave	Prolonged length of stay Colorectal surgery pLOS Risk Score Predictive logistic regression modeling Clinical Decision Support Systems
Committee Chair / Thesis Advisor	Sweeney, John F, Emory University Sullivan, patrick Sean, Emory University
Committee Members	Shah, Amit J, Emory University Manatunga, Amita, Emory University

Ultima modifica

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Predicting Prolonged Length of Stay after Elective Colectomy: Development of a Clinical Decision Support System Using ACS NSQIP Data ()	2018-08-28 13:50:17 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions