3004 — Effects of VISN 7 OncWatch Intervention on Colorectal Cancer Screening Adherence
Bian J (Atlanta VAMC), Fisher D
(Duke University and Durham VAMC), Lipscomb J
(Emory University), Ribeiro M
(Atlanta VAMC), Byrd-Sellers J
The U.S. Department of Veteran Affairs (VA) is the leader in applying health information technology (IT) for quality improvement. However, there is little direct empirical evidence linking use of IT to quality improvement. In fiscal year (FY) 2008, the Veterans Integrated Service Network (VISN) 7 implemented the colorectal cancer (CRC) Oncology Watch intervention (OncWatch), an IT system whose priorities aimed to (1) improve CRC screening adherence and (2) expand use of colonoscopy for timely CRC diagnostic and surveillance follow-ups. This study is to evaluate the effects of the OncWatch on CRC screening adherence.
We used FY 1998-2009 VA administrative data to construct two independent cohorts of average-risk, age 50-64 veterans eligible for CRC screening, one for FY 2007 and the other for FY 2009. Veterans in a cohort for a given year were considered adherent if they completed fecal occult blood test (FOBT) during that year, flexible sigmoidoscopy or double-contrast barium enema during that year or the 4 previous years, or colonoscopy during that year or the 9 previous years. Using a difference-in-differences (DD) approach (or quasi-experiment), we applied multivariable linear probability models with VA medical center (VAMC) fixed-effects for estimation. In the DD approach, VISN 7 served as the intervention site and the other 20 VISNs as the control sites; FY 2007 was designated as the pre-intervention period and FY 2009 the post-intervention period. Additional covariates included veteran demographics and VAMC-level time-varying characteristics.
There were a total of 137,225 average-risk veterans eligible for CRC screening in VISN 7 (including 9 VAMCs), consisting of 65,876 veterans in 2007 and 71,349 in 2009. The corresponding numbers for the control VISNs (including 120 VAMCs) were 2,030,842; 998,572; and 1,032,270. The proportions of veterans adherent to CRC screening in VISN 7 were 31.62% in 2007 and 34.37% in 2009. The proportions in the control VISNs were 30.27% and 32.33%, respectively. Among the screening adherent, the proportions adherent to colonoscopy in VISN 7 were 16.44% in 2007 and 24.40% in 2009; the proportions in the control VISNs were 26.16% and 38.59%, respectively. The DD analyses showed that OncWatch was associated with a one percentage point increase in the likelihood of adherence among the veterans eligible for screening and a 3 percentage point decrease in the likelihood of screening colonoscopy among the adherent (both P < 0.001). In addition, the results suggested that increases in VAMC outpatient visit volume were associated with significant reductions in the likelihood of screening adherence among the eligible and use of screening colonoscopy among the adherent (both P < 0.001).
This preliminary study suggests that OncWatch slightly increased the overall likelihood of CRC screening adherence among average-risk, age 50-64 veterans. However, the intervention may have unexpectedly reduced use of screening colonoscopy. One possible explanation could be that under the OncWatch two priorities, limited colonoscopy capacity was diverted from screening to diagnostic and surveillance indications. However, the observed shift in screening modality from FOBT to colonoscopy may represent more efficient use of limited resources.
Wide adoption of health IT in the VA may improve quality of care.