What Can Be as a Good Predictor of a Higher ADR in GI practices (Actuarial Science)

Presenter Information

Rui Huang
Jinlong Cheng

Document Type

Oral Presentation

Campus where you would like to present

SURC 202

Start Date

16-5-2013

End Date

16-5-2013

Abstract

Our presentation deals with ongoing data collected and analyzed from the Gastroenterology Practice Management Group (GPMG) from 2007–2010. At each participating site, a research coordinator collects de-identified group-level data from ambulatory centers across the United States. We established our main variable as the adenoma detection rate (ADR), which is calculated by dividing the number of colonoscopies with adenomatous polyp by the total number of colonoscopies. ADR was not stratified by gender or indication and was adjusted for age greater than 50. Our main goal is to analyze the performance of ADR, in response to four other highly relevant variables: 1) the quality of the preparation; 2) value of the withdrawal rate; 3) changes of seasons; and 4) location. A total of 256,362 colonoscopies were conducted, of which 84,507 were initial screening (45 percent male, 78 percent over age 50, cecum reached 97 percent), and had been distributed into 13 groups representing 37 gastroenterologists. With statistical analyzing tools, SPSS and Microsoft Excel, we iteratively performed a series of data analyzing methods, including regression analysis, one-way and two-way analyses of variance, and descriptive and inferential analysis techniques. As a result, we are able to find statistically significant relationships among the variables, which can be used as predictors for potential ADRs.

Faculty Mentor(s)

Dominic Kylve

Additional Mentoring Department

Actuarial Science

This document is currently not available here.

Share

COinS
 
May 16th, 12:20 PM May 16th, 12:40 PM

What Can Be as a Good Predictor of a Higher ADR in GI practices (Actuarial Science)

SURC 202

Our presentation deals with ongoing data collected and analyzed from the Gastroenterology Practice Management Group (GPMG) from 2007–2010. At each participating site, a research coordinator collects de-identified group-level data from ambulatory centers across the United States. We established our main variable as the adenoma detection rate (ADR), which is calculated by dividing the number of colonoscopies with adenomatous polyp by the total number of colonoscopies. ADR was not stratified by gender or indication and was adjusted for age greater than 50. Our main goal is to analyze the performance of ADR, in response to four other highly relevant variables: 1) the quality of the preparation; 2) value of the withdrawal rate; 3) changes of seasons; and 4) location. A total of 256,362 colonoscopies were conducted, of which 84,507 were initial screening (45 percent male, 78 percent over age 50, cecum reached 97 percent), and had been distributed into 13 groups representing 37 gastroenterologists. With statistical analyzing tools, SPSS and Microsoft Excel, we iteratively performed a series of data analyzing methods, including regression analysis, one-way and two-way analyses of variance, and descriptive and inferential analysis techniques. As a result, we are able to find statistically significant relationships among the variables, which can be used as predictors for potential ADRs.