3090 — A Decision Tool for Determining Colonoscopy Preparation

Imperiale TF (COE, Indianapolis VAMC), Cardwell J (IT, Indianapolis VAMC), Khan M (Department of Medicine, VAMC), Stuart JS (COE, Indianapolis VAMC), Castelluccio P (COE, Indianapolis VAMC), Sutherland JM (COE, Indianapolis VAMC)

Objectives:
In most open access endoscopy units, polyethylene glycol (PEG) is used for colonoscopy preparation despite evidence that sodium phosphate (NaP) is better tolerated, results in a better quality prep, and costs less. Greater use of NaP in open access systems, however, requires an autonomous decision tool (DT) to accurately select patients who could safely use NaP instead of PEG. We designed and pilot tested a computer-based DT that decides how patients prep for colonoscopy.

Methods:
The DT was designed using Boolean logic to examine the electronic medical record (EMR) for colonoscopy indication, medications, diagnoses, and laboratory values, and to determine whether NaP could be used instead of PEG and whether a stimulant laxative was required. The DT was designed to identify patients with renal insufficiency, ascites, CHF, and MI < 6 months - conditions that preclude NaP use. We examined a 6-month (1/06-6/06) sample of patients scheduled for outpatient colonoscopy to determine DT performance as compared with a trained research assistant (RA), who was blinded to the DT results. Disagreements between the DT and RA were adjudicated by a board-eligible internist (MD), who was blinded to decisions made by the DT and RA.

Results:
Of 1201 patients (mean age 61.5 +/- 10 years, 95% men), the DT and RA identified 939 (73%) and 944 (74%) patients, respectively, who could receive NaP instead of PEG, with 98.8% agreement (1187/1201). Of the 14 disagreements, the DT disallowed NaP in 9 and allowed NaP in 5. Of these 5 cases, the MD agreed with the DT in 3. For the second decision, the DT and RA identified 398 (33%) and 393 (33%) of 1201 patients, respectively, with agreement of 99.4% (1195 of 1201). Of 6 disagreements between DT and RA, the DT allowed a stimulant laxative in 6 cases with MD agreement for all 6.

Implications:
The DT identified a substantial proportion of patients who could take NaP instead of PEG, patients with conditions that preclude its safe use, and those who might benefit from addition of a stimulant laxative.

Impacts:
DT has the potential to improve both patient and systems colonoscopy outcomes within VHA.