Organizational Variations in Colorectal Cancer Screening Rates

Yano EM. Organizational Variations in Colorectal Cancer Screening Rates. Sepulveda, CA: VA HSR&D Center of Excellence; 2004 Apr 1. Report No.: Final Report.

Related HSR&D Project(s)

• CRS 02-163 – Organization Variations in Colorectal Cancer Screening Rates

OBJECTIVES: The purpose of this study was to conduct secondary analyses of existing data to elucidate the environmental, organizational, practice and patient level predictors of colorectal cancer (CRC) screening performance among VA medical centers nationwide.RESEARCH PLAN: We obtained CRC screening data through the Office of Quality and Performance's (OQP) External Peer Review Program (EPRP) for FY01 and FY02. Organizational data were obtained from the VHA Survey of Primary Care Practices (1999-00), reflecting over 200 organizational and practice features among 219 geographically distinct VA primary care practices, and other administrative data. Sample measures included environmental features (e.g., region, urban/rural location, managed care penetration), organizational characteristics (e.g., academic affiliation, complexity/size, leadership characteristics), and primary care practice structure (e.g., service line organization, staffing, practice arrangements with specialists, fiscal structure and resource changes, decision support, and managed care practices). METHODOLOGY: Simple frequencies and histograms of the variability in CRC screening in VA settings were analyzed and presented for overall variation assessments. The outcome variables of interest included overall screening penetration rates (any screening modality). We then conducted multivariate analyses to examine the organizational characteristics independently associated with CRC screening rates in VA practices. We used simple linear regression, using different approaches to address the distributional characteristics of CRC screening rates, and hierarchical linear regression, assessing the contribution of organizational factors adjusting for the patient-level characteristics and potential clustering of patients within practices.FINDINGS: CRC screening rates were stable for FY01 and FY02 with screening rates of 62.6% and 61.0% respectively. Significant geographic variations in CRC screening exist, with the South/East regions performing significantly lower than other regions. CRC screening does not vary, however, by the location of the facility in a metropolitan vs. non-metropolitan area. Facility size is a strong predictor of CRC screening (i.e., smaller facilities perform better than larger facilities), while facilities with an academic affiliation are significantly less likely to perform CRC screening, even after adjusting for the size of the facility. The degree of primary care leader practice autonomy was significantly and positively associated with the delivery of CRC screening, even after adjusting for size and academic affiliation. The level of primary care resource sufficiency was also a significant independent predictor of CRC screening, even after adjusting for facility size and academic affiliation. Provider mix was significantly and positively associated with delivery of CRC screening. Primary care practice-level characteristics accounted for a much higher percentage of the variance than patient-level characteristics. Among patient characteristics, female gender and lower income were significantly associated with a lower likelihood of screening whereas older age was significantly and positively associated with a higher likelihood of screening. Patient characteristics (i.e., age, race, gender, frequency of primary care visits) were not predictive of receipt of a follow-up exam after a positive fecal occult blood test.CLINICAL SIGNIFICANCE: VA health policymakers and health care managers lack needed information about the determinants of variations in CRC screening across the VA healthcare system to design evidence-based quality improvements. This study represents the first empirical demonstration of the contribution of discrete primary care practice organizational features on prevention performance in VA settings, pointing to mutable attributes that ought to be considered in intervention design. STATUS: Complete.