Statistical Analysis
For the 190 veterans who had a colonoscopy performed through VCP, a 1-sample Wilcoxon signed rank test was used with a null hypothesis that the median difference in days between first available VAPHS colonoscopy and community colonoscopy dates was 0. For the utilization sample of 350 veterans, an independent samples median test was used to compare the median wait times for colonoscopy procedures performed at the VA and those performed through VCP. IBM SPSS Version 25 was used for all statistical analysis.
Results
Of the 190 identified colonoscopies completed in the community utilizing VCP, scanned records could not be found for 29 procedures (15.3%) (Table). VCP procedures were performed a median 2 days earlier than the first available VAPHS procedure, but this difference was not statistically significant (P = .62) (Figure 2). Although 52% of colonoscopies occurred sooner through VCP than the initially scheduled VAPHS date, 44% were performed later, and there was wide variability in the difference between these dates, ranging from 49 days sooner to 165 days later.
Pathology results from VCP procedures for which tissue samples were obtained were absent in 11.9% (14 of 118) of procedures. There were no clear follow-up recommendations to referring VA health care providers in the 18% (29 of 161) of available procedure reports. In VCP procedures, documentation of selected quality metrics: bowel preparation, cecal withdrawal time, and rectal retroflexion, were deficient in 27.3%, 70.2%, and 32.9%, respectively (Figure 3).
The utilization dataset sample included 350 veterans who were offered a VCP colonoscopy because the first available VAPHS procedure could not be scheduled for > 30 days. Of these patients, 231 (66%) ultimately had their colonoscopy performed at VAPHS. An additional 26.6% of the patients in the utilization sample were lost in the scheduling process (ie, could not be contacted, cancelled and could not be rescheduled, or were a “no show” their scheduled VAPHS procedure). An unknown number of these patients may have had a procedure outside of the VA, but there are no records to confirm or exclude this possibility. Ultimately, there were only 26 (7.4%) confirmed VCP colonoscopy procedures within the utilization sample (Figure 4). The median actual wait time for colonoscopy was 61 days for VA procedures and 66 days for procedures referred through the VCP, which was not statistically significant (P = .15).
Discussion
This is the first study to evaluate the performance of the VCP for colonoscopy referrals. Consistent with recently reported data in other specialties, colonoscopy referrals through VCP did not lead to more timely procedures overall, although there was wide variation.8 The use of VCP for veteran referral to the community for colonoscopy led to fragmentation of care—with 15% of records for VCP colonoscopies unavailable in the VA EHR 6 months after the procedure. In addition, there were 45 pre- or postprocedure visits in the community, which is not standard practice at VAPHS, and therefore may add to the cost of care for veterans.
Documentation of selected colonoscopy quality metrics were deficient in 27.3% to 70.2% of available VCP procedure reports. Although many veterans were eligible for VCP referral for colonoscopy, only 7.4% had a documented procedure through VCP, and two-thirds of veterans eligible for VCP participation had their colonoscopy performed at the VAPHS, reflecting overall low utilization of the program.
The national average wait time for VCP referrals for multiple specialties was estimated to be 51 days in a 2018 Government Accountability Office (GAO) report, which is similar to our findings.9 The GAO report also concluded that the VCP does not have timeliness standards and notes missed opportunities to develop a mechanism for record transfer between the community and the VA. Our finding of missing colonoscopy procedure and pathology reports within the VA EHR is consistent with this claim. Our analysis revealed that widely accepted quality standards for colonoscopy, those that are required at the VA and monitored for quality assurance at the VAPHS, are not being consistently reported for veterans who undergo procedures in the community. Last, the overall low utilization rate, combined with overall similar wait times for colonoscopies referred through the VCP vs those done at the VA, should lead to reconsideration of offering community care referral to all veterans based solely on static wait time cutoffs.