The recommended interval for the next colonoscopy is based on the endoscopic and pathologic findings of the initial colonoscopy. These intervals range from 3 to 10 years. Recent questionnaire studies report that physicians in the community setting recommend repeat colonoscopies sooner than recommended by published guidelines. Adherence to colonoscopy guidelines in the VA healthcare system is unknown. Physician recommendations of procedures are a key factor in health care utilization. Therefore, recommendations for colonoscopies that exceed guidelines may lead to overuse. Overuse of colonoscopy is expensive, reduces the capacity to perform initial screening or diagnostic colonoscopies, and places veterans at unjustified risk.
The overall goal of this study was to determine if VA physicians follow published guidelines in their recommendations for when colonoscopy should be repeated, and to identify productive targets for modifying provider behavior.
Aim 1: To document the current patterns of repeat colonoscopy recommendation in the VA healthcare system with respect to findings at index colonoscopy, and to compare provider recommendations for repeat colonoscopies to the published multi-society guidelines.
Aim 2: To identify patient, physician, and organizational factors associated with any departure from guidelines.
Data abstracted from the VA electronic medical record (VistAWeb), VA administrative claims, and physician databases were analyzed for a national random sample of VA patients, aged 50-64 years, who had recently undergone colonoscopy and had had no prior colonoscopies for at least 10 years. We estimated the proportion of colonoscopies for which the endoscopist followed guidelines in recommending the interval until the next colonoscopy. Logistic regression modeling, with evaluation of clustering, was used to assess predictors of guideline adherence.
Rates of non-adherence were 37% (532/1455) overall, 26% (236/893) for normal colonoscopy, 61% (123/203) for hyperplastic polyps, 46% (107/231) for low risk adenomas, and 52% (66/128) for higher risk adenomas. The non-adherent follow-up interval was earlier than recommended by guidelines for 100% of the normal, hyperplastic and low risk adenoma patients and for 86% of the higher risk adenoma patients. Adjusted modeling indicated clinical group and geographic region were associated with non-adherence.
There were no significant associations between the following variables and non-adherence to the published guidelines: years in practice, physician specialty, and physician board certification. Compared to male physicians, female physicians were 2.27 (95% CI: 1.01, 5.09) times more likely to be non-adherent to the published guidelines in the "no polyp/normal tissue" group. However, the same was not true for the "hyperplastic" group (OR = 1.30, 95% CI: 0.48, 3.55) and the "1-2 adenomas < 1 cm in size" group (OR = 0.96, 95% CI: 0.47, 1.97). Out of the 25 facilities, adherence ranged from 2.7% - 100%; 12 facilities (48%) had adherence rates of > 80%, 8 (32%) had adherence rates of 60-79%, and 5 (24%) had adherence rates of < 60%.
Among the 2443 colonoscopies, 476 (19%) were missing documentation of prep adequacy. Missing documentation among the 25 facilities ranged from 0 to 70%. Among the 1967 patients with documentation of prep quality, 145 (7%) were reported as inadequate. Proportion of patients at each facility with inadequate colon cleansing ranged from 1% to 48%. Cecal intubation was achieved in 64% and 94% of patients with inadequate and adequate bowel preparation, respectively, p<0.001. VA facilities in the Northeast and Midwest had a greater proportion of colonoscopies with missing bowel prep documentation as well as inadequate colon cleansing (p<0.0001 and p<0.001, respectively). Predictors of missing prep quality documentation on multivariate analysis were geographic region, non-gastroenterologist specialty, and trainee involvement.
Our study identified potential targets for quality improvement measures in the VA healthcare system. Future intervention studies are needed to address potential barriers to guideline-concordance at the facility level, as our study found that there was significant variation on guideline adherence at both the facility level and across geographic regions. Colonoscopy capacity within the VA may be increased by interventions to reduce overuse for polyp surveillance. Future research is also needed to identify ways to improve consistency across VA facilities with regards to clinical indicators such as adequacy of bowel prep, which was also found to vary significantly by facility and geographic region. Clinical documentation of determinants of the diagnostic accuracy of colonoscopy such as documentation of bowel prep quality also varied by geographic region, specialty, and trainee involvement. Potential clinical applications such as a system-wide endoscopy software program that would facilitate improved documentation of key performance variables would facilitate improved documentation of key performance variables and allow real-time assessment of quality measures and quality improvement after interventions.
External Links for this Project
None at this time.