Health Care Systems for Tracking Colorectal Cancer Screening Tests

Table 4.14. Screening Rate Multivariable Logistic Regression, Intervention Practices Only: Reduced Model

Characteristics Stool Testa Colonoscopyb Any Screening Testc
OR 95% CI p OR 95% CI p OR 95% CI p
 
Patient Characteristics
Gender
Male                
Female         1.20 1.01 1.44 0.043        
Age
50-54        
55-64 1.38 0.88 2.17 0.159         1.09 0.80 1.47 0.596
65-69 1.98 1.36 2.87 0.000         1.45 1.09 1.95 0.012
70-79 2.22 1.25 3.96 0.007         1.60 1.15 2.22 0.005
Insurance
Commercial and other
Medicare 0.96 0.80 1.15 0.673 0.86 0.67 1.09 0.213 0.86 0.76 0.97 0.016
Self/Medicaid 0.71 0.52 0.96 0.028 0.63 0.29 1.39 0.255 0.64 0.40 1.01 0.055
Practice Characteristics
Specialty
Family medicine        
General internal medicine         1.28 1.09 1.49 0.003 1.16 0.98 1.36 0.080
Size
Small
Large 0.69 0.58 0.82 0.000 0.50 0.24 1.06 0.070 0.58 0.43 0.78 0.000
Location
Urban
Rural 1.52 0.91 2.55 0.110 2.65 1.61 4.35 0.000 2.01 1.62 2.49 0.000
Suburban 1.18 0.80 1.73 0.410 3.74 2.29 6.13 0.000 2.23 1.79 2.78 0.000
Data sources usedd
Least complete
Moderately complete 1.91 0.58 6.29 0.287 18.56 6.11 56.31 0.000 4.88 1.71 13.92 0.003
Most complete 1.94 0.96 3.93 0.066 19.32 8.84 42.23 0.000 4.62 3.00 7.12 0.000

a. Hosmer-Lemeshow Χ2 = 24.69, p = 0.002.
b. Hosmer-Lemeshow Χ2 = 14.33, p = 0.074.
c Hosmer-Lemeshow Χ2 = 13.38, p = 0.099. Includes stool test, colonoscopy, flexible sigmoidoscopy, or barium enema x ray. Persons with screens by multiple modalities are counted only once.
d The data sources used affects the amount of screening information available for patients of a given practice. It is not an indicator of predicting screening. Instead, this variable is used to control for the fact that some practices had more complete data than others.

Note: Confidence intervals and p values have been adjusted using cluster variance estimators.

Return to Document

Page last reviewed October 2014
Internet Citation: Table 4.14. Screening Rate Multivariable Logistic Regression, Intervention Practices Only: Reduced Model. October 2014. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/findings/final-reports/crcscreeningrpt/crcscreentab4-14.html