Trial Information

Assessing and Addressing Patient Colorectal Cancer Screening Barriers

Background / Rationale:

Despite strong evidence for the effectiveness and cost-effectiveness of a variety of
colorectal cancer (CRC) screening methods for reducing CRC mortality, current CRC screening
rates fall far below the levels needed to significantly impact CRC mortality. Unfortunately,
however, the existing literature on patient CRC screening behavior does not yet provide a
sufficient evidence base for making sound recommendations regarding how to most effectively
improve upon these rates in the VA. This study will inform future CRC screening promotion
efforts and make important scientific contributions to existing literature by: (a)
delineating the relative contribution of patient cognitive, environmental and background
factors to CRC screening behavior using a multi-level, theory driven analysis approach on a
nationally representative sample, and (b) identifying the determinants of variation in CRC
screening behavior across vulnerable population subgroups.

Objective(s):

The overall goal of this study was to address significant gaps in the existing evidence base
in order to inform the development of effective patient-directed interventions to increase
CRC screening among veterans age 50 and older. This was accomplished by using data collected
from a mailed patient survey and theory-based analysis approaches to uncover key barriers to
screening adherence and to identify fruitful intervention approaches for modifying them. The
specific primary objectives of this study were to: (1) Estimate the relative effect of
patient cognitive (knowledge, attitudes, and self-efficacy), environmental (social network
and medical care characteristics), and background (demographics, health status, prior
screening experiences) factors on CRC screening behavior; (2) Identify factors that
contribute to any disparities in CRC screening behavior by race/ethnicity or other patient
characteristics; (3) Identify from these analyses: (a) priority population subgroups to
target in future interventions (i.e., those at the greatest risk of failing to be screened),
and (b) priority factors to target in future interventions (i.e., those that are not only
strongly associated with CRC screening but also prevalent in the target population and
amenable to intervention, as well as those that are most likely to ameliorate race and other
disparities). Secondary objectives included: (1) assessing patient values and preferences
regarding the various CRC screening modality options, (2) estimating stage of readiness to
adopt CRC screening in the study population, and (3) validating measures of CRC knowledge
and self-reported screening behavior.

Methods:

This is an observational study based on a nationally representative, cross-sectional mailed
survey of 3,744 male and female veterans age 50-75 who have had one or more primary care
visits at a VA Medical facility in the past two years. The survey sample was drawn using a
two stage procedure where we first randomly select 24 VA facilities stratified by size and
racial mix and then select a simple random sample of 156 eligible veterans from each sampled
facility. Prior to the national survey, a pilot survey was conducted with the purpose of
refining both the study instruments and the study protocol. The sample consisted of 900
randomly selected veterans from the Minneapolis VA Medical Center meeting the same sampling
eligibility criteria used for the national mailed survey. The mailed patient questionnaire,
made up primarily of previously validated measures, included measures of self-reported CRC
screening behavior; patient demographic, health, social network and medical care
characteristics; CRC screening knowledge, attitudes, social norms and self-efficacy; and
attitudes toward medical care. Additional measures of organizational-level CRC screening
practices from a recently completed VA facility survey were linked to the patient survey.The
primary outcome is whether the patient is currently compliant with CRC screening guidelines
(i.e., received either a fecal occult blood test in the past year, a sigmoidoscopy or double
contrast barium enema in the past five years, or a colonoscopy in the past ten years). The
primary analyses tested (using logistic regression and a multi-level, structural equation
modeling approach) specific hypotheses about the association between this measure and
patient background, cognitive and environmental factors and their interactions. Additional
analyses conducted include a multinomial logistic regression to assess patient screening
mode preferences and their determinants, and logistic and multinomial logistics regression
analyses with interactions to determine whether and why any observed patterns in CRC
screening behavior vary by race.

Status:

Completed