- Email: [email protected]
Colonoscopy practice for veterans within and outside the Veterans Affairs setting: a matched cohort study
ORIGINAL ARTICLE
Colonoscopy practice for veterans within and outside the Veterans Affairs setting: a matched cohort study Michael J. Bartel, MD,1 Douglas J. Robertson, MD,2 Heiko Pohl, MD2,3 White River Junction, Vermont, USA
Background and Aims: To minimize delays for colonoscopy within Veterans Affairs (VA) facilities, veterans may receive care at non-VA facilities based on fee-for-service contracts, and more recently, through the Veterans Access, Choice, and Accountability Act. The impact of diverting care from VA to non-VA facilities on quality of colonoscopy practice is unknown. Methods: We identified all veterans aged 50 to 85 years who received a fee-basis colonoscopy for colorectal cancer screening or surveillance at non-VA facilities in 2007 to 2010. These patients were matched for sex, age, and year of procedure to veterans who underwent colonoscopies at VA medical centers. The outcomes of interest were the adenoma detection rates (ADR) and compliance with surveillance guidelines. Results: During the observation period, 409 veterans (mean age 64 years; 94% men) underwent a fee-basis colonoscopy at 30 nonacademic (54%) and 2 academic (46%) facilities. Compared with colonoscopies performed at VA facilities, fee-basis colonoscopy patients had lower ADRs (38% vs 52%; P < .001), lower mean number of adenomas per procedure (0.72 vs 1.41; P < .001), and lower number of advanced ADRs (13% vs 22%; P < .001). Colonoscopies done at non-VA facilities were associated with lower ADRs in multivariate regression analysis (odds ratio 0.64; 95% CI, 0.44-0.92), whereas colonoscopies done in nonacademic settings or by colonoscopists who were not gastroenterologists were not. Compliance with surveillance guidelines was lower for colonoscopies performed outside VA facilities (80% vs 87%; P Z .03). Conclusions: In this regional study (Northern New England), compliance with colonoscopy surveillance guidelines was high in both VA and non-VA settings; however, lower ADRs raise concern that referring veterans outside the VA system may impact colonoscopy quality. (Gastrointest Endosc 2016;-:1-7.)
Abbreviations: ADR, adenoma detection rate; MAP, mean number of adenomas per procedure; SSA/P, sessile serrated adenoma and/or polyp; VA, Veterans Affairs; VHA, Veterans Health Administration. DISCLOSURE: This material is the result of work supported in part by resources from The Veterans Health Administration. The views expressed in this publication are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. All authors disclosed no financial relationships relevant to this publication. Copyright ª 2016 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2016.01.017 Received August 20, 2015. Accepted January 7, 2016. Current affiliations: Department of Gastroenterology and Hepatology, Mayo Clinic, Jacksonville, Florida (1), Department of Gastroenterology, White River Junction VA Medical Center, White River Junction, Vermont (2), Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire (3). Reprint requests: Michael J. Bartel, MD, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL 32224.
www.giejournal.org
The Veterans Health Administration (VHA) endorsed screening colonoscopy as a primary screening option for all veterans in 2007.1 Although fecal occult blood tests and sigmoidoscopy have been alternative screening options, colonoscopy has become the primary screening modality in and outside the Veterans Affairs (VA) system.2-4 However, as the single largest health care provider in the United States, the VHA has been unable to meet the demands for screening and surveillance colonoscopies. In order to ensure access to care, VA facilities occasionally use fee-for-service contracts with non-VA facilities. Through such mechanisms, veterans are able to receive colonoscopies outside of a VA facility at VHA expense. Of approximately 300,000 colonoscopies performed annually under VHA care, 100,000 procedures are completed by non-VA providers on a fee-for-service basis (personal communication with Jason Dominitz, VHA National Gastroenterology Program Director, July 2015). In August 2014, Congress passed The Veterans Access, Choice, and Accountability Act of 2014 (Choice Act), which Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 1
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
allows for care outside of the VHA for veterans “who either cannot be seen within the wait-time goals of the VHA or who qualify based on their place of residence.”5 Although they are similar in some aspects, the national VA health care system differs from the private practice setting in many ways. In both systems, physician practice is dictated by societal guidelines with respect to colorectal cancer screening and surveillance. In addition, the VHA, as a centralized health care provider, issues directives to guide care and mandate quality benchmarks.1 In a private practice setting, care is based on local preference at the facility. Financial incentives also play a greater role in private practice than within a VA setting. In addition, the veteran population differs from the general population with regard to adenoma risk factors (eg, male sex), and a higher adenoma prevalence should be expected.6 It is possible, therefore, that the system in which a physician works may have an effect on his or her practice, and that colonoscopy performance may differ between private practice and that done at VA medical centers. It is unclear whether colonoscopy quality is affected by referring patients to non-VA providers. Therefore, we examined colonoscopy practice performed outside of and within a VA setting and measured 2 main quality metrics: adenoma detection and compliance with surveillance guidelines.7 Adenoma detection is strongly associated with interval colorectal cancers, and, therefore, is considered a main quality indicator.8,9 Previous studies have shown that compliance with surveillance colonoscopy guidelines was poor in a non-VA setting and showed overutilization for patients with low-risk findings and underutilization for those with high-risk findings.10-14 Our objective was to compare adenoma detection and compliance with surveillance guidelines between veterans undergoing a fee-for-service colonoscopy outside of the VA system and those who received a colonoscopy at a VA medical center.
performed by trainees were supervised by attending gastroenterologists. Patients with concomitant lower GI symptoms, positive fecal occult blood test results, or histories of inflammatory bowel disease were excluded in order to avoid substantial variations of the adenoma detection rate (ADR). Patients who underwent fee-basis colonoscopies were matched for sex, age (according to the year of birth), and year of procedure, 1-to-1 with patients who underwent screening or surveillance colonoscopies at VA medical centers (VA cohort). The proportion of fee-basis colonoscopies between 2007 and 2010 was approximately 10%. We assessed, therefore, only every fifth consecutive patient undergoing a colonoscopy at a VA facility in each of the 4 years of eligibility (age and sex, in addition to year). If a patient did not meet matching criteria, the next fifth consecutive patient was identified. This process was repeated until the VA-matched control group was complete. The study was approved by the institutional review board at Dartmouth College.
Data collection
METHODS
Data from the VA cohort were abstracted by retrospective chart review for patient age, sex, demographics, procedure indication, colonoscopy findings, polyp characteristics, and histology, as well as recommended surveillance interval given by the endoscopist. Similar data were abstracted in the fee-basis cohort from available colonoscopy, pathology reports, and postprocedure correspondence. In the case of >1 polyp, detailed polyp characteristics (ie, size and histology) were recorded for the most advanced neoplastic polyp. Based on indication and colonoscopy findings, the appropriate surveillance interval was determined, applying the U.S. Multi-Society Task Force guidelines current at the time of the procedure.15,16 For all surveillance colonoscopies, previous colonoscopy findings were also considered when the appropriate surveillance interval was determined.
Patients
Outcomes
We identified all patients aged 50 to 85 years who were referred for a screening or a surveillance fee-basis colonoscopy from the White River Junction VA Medical Center, White River Junction, Vermont, to non-VA facilities between 2007 and 2010. The decision for fee-basis referral was based on increased demand at 3 time points during these years. Only then were patients offered a referral to an outside provider. Referred patients underwent a feebasis colonoscopy based on personal preferences in selfchosen local facilities typically close to their home addresses. These non-VA facilities included both academic and non-academic settings, and both gastroenterologists and surgeons performed colonoscopies (fee-basis cohort). Involvement of gastroenterology trainees at academic centers was not an exclusion criterion. All colonoscopies
As the primary outcome of interest, we obtained the ADR, defined as the proportion of patients with at least 1 adenoma. In addition, we calculated the mean number of adenomas per procedure (MAP). We further obtained the advanced ADR, defined as patients with 1 large adenomas (10 mm), 1 adenomas with advanced histology (villous component, high-grade dysplasia, or cancer), or 3 small adenomas. For the purpose of this study, sessile serrated adenomas and/or polyps (SSA/P) and tubular adenomas were combined as adenomas. The second main outcome of interest was the proportion of recommendations compliant with the U.S. Multi-Society Task Force guidelines.15,16 We first obtained the proportion of patients who received a surveillance recommendation. We then calculated the compliance rate by correctly
2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
www.giejournal.org
Bartel et al
assigned surveillance colonoscopy intervals among all patients who received a recommendation. We further obtained the proportion of patients who received early or late surveillance interval recommendations. In addition, outcomes were stratified by academic affiliation (academic vs non-academic) and endoscopist specialty (gastroenterologist vs surgeon).
Statistical analysis Continuous variables are presented as means with standard deviation (SD) if normally distributed or as medians with interquartile range (IQR) if not normally distributed. Proportions are presented as percentage with a 95% confidence interval (CI). Categorical data for both cohorts were analyzed by using a chi-square test. Categorical variables (including the ADR) were analyzed by using the chi-square test. Continuous variables were compared by using the non-paired t test if normally distributed and the Wilcoxon Mann-Whitney test if not normally distributed. We further examined possible factors associated with adenoma detection by using logistic regression, considering age, sex, setting, endoscopist specialty, and adenoma detection. Stepwise multivariable models were used to account for potentially confounding relationships between variables. Results were rounded to the nearest percentage point to facilitate text presentation; the tables provide more precision.
RESULTS Patient characteristics Between 2007 and 2010, a total of 409 patients (mean age 64, 95% male) underwent a fee-basis colonoscopy (fee-basis cohort) for either screening (68%) or surveillance (32%). Colonoscopies were performed by gastroenterologists (63%) or surgeons (37%) at 30 nonacademic (59%) and 2 academic facilities (41%). The control group (VA cohort) comprised 409 patients of similar age and sex (mean age 64 years, 95% male). Colonoscopies in the control group were more often performed for surveillance (39%) than in the non-VA group (32%; P < .05), with the remaining being performed for screening (61%). Eleven gastroenterologists with academic affiliations performed all colonoscopies in the control group (Table 1).
Adenoma detection Table 2 details colonoscopy findings. Patients undergoing colonoscopies at non-VA facilities had lower ADRs (38%) than those receiving colonoscopies at the VA medical center (52%; P < .001). Only about half as many adenomas were found per procedure in the fee-basis cohort compared with the VA cohort (0.7 vs 1.4; P < .001). Similarly, the advanced ADR was lower outside of the VA facilities than within the VA facilities (13% vs 22%; P < .001). Additional subgroup analysis www.giejournal.org
Colonoscopy outcomes in Veterans Affairs setting
TABLE 1. Patient baseline characteristics VA Fee-basis P (n [ 409) (n [ 409) value Patients Age, mean ( SD), y
63.7 (7.9)
63.7 (8.1)
NS
Male, no. (%)
387 (94.6) 387 (94.6)
NS
Colonoscopy indication, no. (%)
.041
Screening
250 (61.1)
278 (68)
Surveillance
159 (38.9)
131 (32) < .001
Endoscopist specialty, no. (%) Gastroenterologist
409 (100) 256 (62.6)
Surgeon
0
153 (37.4) < .001
Academic affiliation, no. (%) Academic
409 (100) 168 (41.1)
Nonacademic Polyps, no. Adenomas, no. Adenoma size, median (IQR), mm Adenoma 10 mm, no. (%)
0
241 (58.9)
901
520
< .001
578
296
< .001
4 (3-6)
5 (3-7)
NS
36 (16.8)
29 (18.5)
NS
Adenoma histology, no. (%) 563 (64.5) 275 (52.9) < .001
Tubular adenoma Villous component*
11 (1.2)
11 (2.1)
NS
Sessile serrated adenoma and/or polyp
1 (0.1)
9 (1.7)
.02
Polyp with high-grade dysplasia or cancer
3 (0.3)
2 (0.4)
NS
VA, Veterans Affairs; SD, standard deviation; NS, not significant; IQR, interquartile range. *Tubulovillous adenoma and villous adenoma.
demonstrated that gastroenterologists from VA facilities yielded significantly higher ADRs (52.3% vs 38.3%), MAP (1.41 vs 0.73), and advanced ADRs (22.2% vs 13.2%) compared with their fee-basis counterparts. Comparable results were found when only endoscopists with academic affiliations were taken into account. No significant differences for ADR, MAP, and advanced ADR were detected between gastroenterologists and surgeons or between academic and non-academic affiliated practices within the fee-basis cohort. In multivariate analysis, indication and colonoscopy performance at non-VA facilities were independently associated with adenoma detection (Table 3). Veterans who underwent colonoscopies at non-VA facilities had a 36% lower chance of being diagnosed with an adenoma (odds ratio [OR] 0.64; 95% CI, 0.44-0.92) compared with those undergoing colonoscopies at VA facilities. Although endoscopist specialty and academic affiliation were associated with adenoma detection in univariate analysis, these factors were not significant after adjusting for indication, sex, and referral setting. Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 3
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
TABLE 2. Polyp characteristics VA (n [ 409)
Fee-basis (n [ 409)
P value
Polyp detection rate, %
69.2
51.6
< .001
Adenoma detection rate, %
52.3
38.4
< .001 < .001
Specialty Gastroenterologist
52.3
38.3*
–
38.6
52.3
40.5*
Nonacademic
–
36.9
Mean adenomas per procedure, no.
1.41
0.73
1.41
0.78*
–
0.64
1.41
0.84*
Nonacademic
–
0.65
Advanced adenoma detection rate, %
22.2
13.2
22.2
14.4*
–
11.1
Surgeon Setting
< .01
Academic < .001 < .01
Specialty Gastroenterologist Surgeon
< .001
Setting Academic
< .001 < .02
Specialty Gastroenterologist Surgeon Setting Academic Nonacademic
.066 22.2
15.5*
–
13.1
VA, Veterans Affairs. *No significant difference between gastroenterologist and surgeon or academic and nonacademic affiliation.
Compliance with surveillance guidelines Recommendations for surveillance examination were less frequently documented for non-VA than for VA colonoscopies (74% vs 93%; P < .001). Among patients who received recommendations, those provided at non-VA sites were less frequently compliant with guidelines than those given at the VA site (80% vs 87%; P Z .03). This difference was primarily related to recommending a shorter interval, which was observed for 13% and 6% in the non-VA and the VA group, respectively (P Z .004). A longer interval than suggested by guidelines was similar (7%) in both settings. Among all non-VA colonoscopies, recommendations were less frequently compliant with guidelines if the procedure was performed in a non-academic setting or by a surgeon (Table 4). In multivariate analysis, an indication of screening was the only factor independently associated with stronger compliance with guidelines (OR 2.03; 95% CI, 1.33-3.11). Although academic affiliation, gastroenterology specialty, and colonoscopy performance at a non-VA site were associated with compliance with univariate analysis, these 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
factors were no longer independently associated with compliance after adjustment in multivariate analysis (Table 5).
DISCUSSION In order to meet colonoscopy demand, the VHA allows veterans to obtain care outside of the VA system based on a fee-for-service contract, and more recently as part of the Choice Act.5 Focusing on adenoma detection and compliance with surveillance guidelines, we found a significantly lower ADR, MAP, and advanced ADR for veterans who received a fee-basis colonoscopy outside of the VA system in comparison with veterans who underwent colonoscopies within the VA system. Similarly, surveillance recommendations were given less frequently after colonoscopies performed outside of the VA and were less frequently compliant with U.S. Multi-Society Guidelines; however, this association was not significant in multivariate analysis. Differences in adenoma detection within the same health care systems or local practices are well-described and felt to be primarily attributed to the patient population and the quality of the technical equipment and skill of the endoscopists.17-19 We found a significant difference in adenoma detection in the same patient population dependent on where patients had the colonoscopydat a non-VA practice or a VA medical center. Adenoma detection is an important marker of colonoscopy quality.8,9,19,20 Recent studies have shown a strong association between adenoma detection and the risk of interval cancer.8,9 In a study by Corley et al,8 each 1% increase in ADR was associated with a 3% decrease in the risk of interval cancer. In our study, we found a surprisingly high 14% absolute difference between the non-VA and the VA groups. Some of the observed difference may be attributed to nonmeasured confounders. However, the groups were well-matched, and it seems more likely that something other than patient factors are responsible for the differences in adenoma detection. Academic affiliation or endoscopist specialty (ie, surgeon) were not associated with adenoma detection. Broad variation in adenoma detection across endoscopists has been described before, and is likely responsible for some of the observed difference in adenoma detection.21,22 Further, it is possible that system factors may contribute to adenoma detection. For instance, veterans have a higher adenoma risk than the general population.6 Therefore, endoscopists performing colonoscopies at VA medical centers can expect a higher ADR than endoscopists practicing in the community. This may set different baseline expectations (VA vs non-VA) and affect adenoma detection in the same risk group (veterans). Another important difference is practice organization. VA directives guide care and set benchmarks for quality.1 In contrast, www.giejournal.org
Bartel et al
Colonoscopy outcomes in Veterans Affairs setting
TABLE 3. Associations with adenoma detection rate Univariate
Multivariate
Odds ratio
P value
Odds ratio
P value
Male
1.33 (0.72-2.5)
NS
1.25 (0.66-2.35)
NS
Colonoscopy indication (screening)
0.62 (0.46-0.82)
.001
0.64 (0.48-0.86)
.003
Facility (academic)
1.63 (1.2-2.22)
.002
1.29 (0.78-2.15)
NS
Endoscopist (gastroenterologist)
1.41 (0.98-2.02)
.061
0.87 (0.52-1.45)
NS
Referral (fee-basis)
0.57 (0.43-0.75)
< .001
0.64 (0.44-0.92)
.017
NS, Not significant.
care in private practice is largely based on local preferences. Financial incentives also play a greater role for the endoscopists in the private setting than within the VA setting. The culture in which colonoscopies are performed may, therefore, result in different outcomes. Although our study reflects only regional practice within a VA referral area in New England, it should alert policy makers that referring veterans outside the VHA may negatively impact colonoscopy quality. In addition to adenoma detection, documentation of surveillance recommendations and compliance with surveillance guidelines are considered important quality metrics.7 Documentation of surveillance recommendations in our study would have reached the recently suggested 90% quality benchmark for colonoscopies performed within the VHA (93%), but not outside the VHA (74%). We were surprised to find an overall high compliance with surveillance guidelines in >80% of colonoscopies when documented. This is a higher rate than reported in previous studies within and outside of the VA system.10,12,14,23,24 Recently, Johnson et al23 examined 25 VA facilities and found an overall compliance of 64%, ranging from 20% to 97%. In our study, recommendations outside of the VHA appeared to be slightly less compliant than those given within the VHA; however, this difference was no longer significant in adjusted analysis. Similar to those of previous studies, noncompliant recommendations were evenly balanced between a too-short and a too-long interval.11,13,23-26 A factor that is frequently associated with poor compliance with guidelines is inadequate bowel preparation.27 More concerning, Saini et al28 reported that the endoscopists contribute significantly to poor compliance; 47.5% did not know the guidelines, and 51% to 76% disagreed with guidelines and recommended earlier surveillance colonoscopies. In a VA study, depending on the clinical question, 4% to 42% of gastroenterologists were not familiar with surveillance guidelines.13 We did not analyze whether knowledge of guidelines contributed to the observed high compliance rate in our study. We found indication for screening as the sole factor associated with high compliance. This is not surprising because the absence of a history of adenoma simplifies recommendations. www.giejournal.org
TABLE 4. Compliance with surveillance guidelines VA Fee basis (n [ 409) (n [ 409)
P value
Recommendations available, no. (%)
380 (92.9)
301 (73.6)
< .001
Surveillance interval in agreement with guidelines, no. (%)
329 (86.6)
242 (80.4)
.03
Compliance
Specialty
NS
Gastroenterologist, % Surgeon, %
86.6
84
–
71.9
86.6
83.8
Setting
< .001* NS
Academic, %
–
77.4
< .001y
Shorter surveillance interval recommended, no. (%)
24 (6.3)
38 (12.6)
.004
Longer surveillance interval recommended, no. (%)
27 (7.1)
21 (7)
NS
Nonacademic, %
NS, Not significant. *Compared with fee-basis gastroenterologists. yCompared with fee-basis academic affiliation.
Several limitations of our study should be noted. First, the retrospective character and its narrow regional restriction raise concerns for selection bias. We minimized this bias by matching controls by age, sex, and procedure year. We further adjusted our analysis for other factors, including indication, with a potential effect on the outcome. However, it cannot be excluded that unmeasured patient risk factors (eg, history of smoking) may account for some of the observed differences. Some may argue that patients referred for a fee-basis colonoscopy may be different from those receiving a colonoscopy at a VA facility. However, the decision was not based on patient preference (and self-selection) but was made by the VA medical center at specific time points during the observation period. When the waiting time for routine colonoscopies became too long, patients were offered a colonoscopy outside the VA system. Outside of these times, fee-basis referral was generally not offered to patients. Likely, longer wait times for referred patients may result in an increased adenoma prevalence in this group. The obtained difference in adenoma detection between the groups may, Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 5
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
TABLE 5. Associations with compliance with surveillance guidelines Univariate
Multivariate
Odds ratio
P value
Odds ratio
P value
1.08 (0.44-2.66)
NS
1.37 (0.55-3.44)
NS
Colonoscopy indication (screening)
1.96 (1.3-2.96)
.001
2.04 (1.33-3.11)
.001
Facility (academic)
1.77 (1.14-2.77)
.011
1 (0.48-2.08)
NS
Endoscopist (gastroenterologist)
2.33 (1.39-3.9)
.001
1.94 (0.92-4.1)
NS
Referral (fee-basis)
0.64 (0.42-0.96)
.03
0.74 (0.43-1.28)
NS
1 (0.67-1.5)
NS
1 (0.65-1.53)
NS
Male
Adenoma detection NS, Not significant.
therefore, be biased toward a lower difference. A selection bias seems less likely. Second, we compared colonoscopy practice at multiple non-VA sites with colonoscopy practice at 1 VA medical center. It is possible that unmeasured factors contributed to the obtained differences in ADR, including individual endoscopists’ ADRs or participation in quality improvement projects.6,9,29 Our results therefore provide only an overall impression on regional colonoscopy practice and are not generalizable to other regions or VA facilities. However, this study can be seen as a natural experiment. Patients who were referred for a fee-basis colonoscopy chose where to have the procedure. To the extent that patient choice is simply driven by geography (and not quality), fee-basis colonoscopy should reflect the average clinical practice in the region. Third, other colonoscopy quality metrics were not analyzed, including cecal intubation rate, colonoscopy withdrawal time, or bowel preparation. These factors might have provided additional insights for the observed differences in adenoma detection in particular and colonoscopy performance in general. With passage of the Choice Act in 2014, VA centers are encouraged to outsource colonoscopies to local non-VA providers to improve access. Our study reflects regional practice in a VA referral region in New England. Veterans who were referred to local non-VA providers on a fee-forservice basis had a lower ADR, MAP, and advanced ADR than veterans undergoing a colonoscopy at their local VA medical centers. Compliance with surveillance guidelines was high in both settings. These findings raise concerns that referring veterans outside the VA may impact quality.
REFERENCES 1. Department of Veterans AffairsdColorectal Cancer Screening, VHA Directive 1015. 2014. Available at: http://www.va.gov/vhapublications/ ViewPublication.asp?pub_IDZ3068. Accessed July 1, 2015. 2. El-Serag HB, Petersen L, Hampel H, et al. The use of screening colonoscopy for patients cared for by the Department of Veterans Affairs. Arch Intern Med 2006;166:2202-8. 3. Harewood GC, Lieberman DA. Colonoscopy practice patterns since introduction of medicare coverage for average-risk screening. Clin Gastroenterol Hepatol 2004;2:72-7.
6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
4. Seeff LC, Richards TB, Shapiro JA, et al. How many endoscopies are performed for colorectal cancer screening? Results from CDC’s survey of endoscopic capacity. Gastroenterology 2004;127:1670-7. 5. Department of Veterans Affairs. Expanded access to non-VA care through the Veterans Choice Program. Interim final rule. Fed Regist 2014;79:65571-87. 6. Kahi CJ, Ballard D, Shah AS, et al. Impact of a quarterly report card on colonoscopy quality measures. Gastrointest Endosc 2013;77:925-31. 7. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for colonoscopy. Am J Gastroenterol 2015;110:72-90. 8. Corley DA, Jensen CD, Marks AR, et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014;370: 1298-306. 9. Kaminski MF, Regula J, Kraszewska E, et al. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010;362: 1795-803. 10. Iskandar H, Yan Y, Elwing J, et al. Predictors of poor adherence of US gastroenterologists with colonoscopy screening and surveillance guidelines. Dig Dis Sci 2015;60:971-8. 11. Menees SB, Elliott E, Govani S, et al. Adherence to recommended intervals for surveillance colonoscopy in average-risk patients with 1 to 2 small (<1 cm) polyps on screening colonoscopy. Gastrointest Endosc 2014;79:551-7. 12. Schreuders E, Sint Nicolaas J, de Jonge V, et al. The appropriateness of surveillance colonoscopy intervals after polypectomy. Can J Gastroenterol 2013;27:33-8. 13. Shah TU, Voils CI, McNeil R, et al. Understanding gastroenterologist adherence to polyp surveillance guidelines. Am J Gastroenterol 2012;107:1283-7. 14. Sint Nicolaas J, de Jonge V, van Baalen O, et al. Optimal resource allocation in colonoscopy: timing of follow-up colonoscopies in relation to adenoma detection rates. Endoscopy 2013;45:545-52. 15. Rex DK, Kahi CJ, Levin B, et al. Guidelines for colonoscopy surveillance after cancer resection: a consensus update by the American Cancer Society and the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2006;130:1865-71. 16. Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US MultiSociety Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 2008;134:1570-95. 17. Benson ME, Reichelderfer M, Said A, et al. Variation in colonoscopic technique and adenoma detection rates at an academic gastroenterology unit. Dig Dis Sci 2010;55:166-71. 18. Shaukat A, Oancea C, Bond JH, et al. Variation in detection of adenomas and polyps by colonoscopy and change over time with a performance improvement program. Clin Gastroenterol Hepatol 2009;7: 1335-40. 19. Coe SG, Crook JE, Diehl NN, et al. An endoscopic quality improvement program improves detection of colorectal adenomas. Am J Gastroenterol 2013;108:219-26; quiz 27.
www.giejournal.org
Bartel et al 20. Lieberman DA, Rex DK, Winawer SJ, et al. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2012;143:844-57. 21. Chen SC, Rex DK. Endoscopist can be more powerful than age and male gender in predicting adenoma detection at colonoscopy. Am J Gastroenterol 2007;102:856-61. 22. Pohl H, Bensen SP, Toor A, et al. Cap-assisted Colonoscopy and Detection of Adenomatous Polyps (CAP) study: a randomized trial. Endoscopy 2015;47:891-7. 23. Johnson MR, Grubber J, Grambow SC, et al. Physician non-adherence to colonoscopy interval guidelines in the Veterans Affairs Healthcare System. Gastroenterology 2015;149:938-51. 24. Schoen RE, Pinsky PF, Weissfeld JL, et al. Utilization of surveillance colonoscopy in community practice. Gastroenterology 2010;138: 73-81.
www.giejournal.org
Colonoscopy outcomes in Veterans Affairs setting 25. Partin MR, Noorbaloochi S, Grill J, et al. The interrelationships between and contributions of background, cognitive, and environmental factors to colorectal cancer screening adherence. Cancer Causes Control 2010;21:1357-68. 26. Radaelli F, Paggi S, Bortoli A, et al. Overutilization of post-polypectomy surveillance colonoscopy in clinical practice: a prospective, multicentre study. Dig Liver Dis 2012;44:748-53. 27. Menees SB, Elliott E, Govani S, et al. The impact of bowel cleansing on follow-up recommendations in average-risk patients with a normal colonoscopy. Am J Gastroenterol 2014;109:148-54. 28. Saini SD, Nayak RS, Kuhn L, et al. Why don’t gastroenterologists follow colon polyp surveillance guidelines? Results of a national survey. J Clin Gastroenterol 2009;43:554-8. 29. Ussui V, Coe S, Rizk C, et al. Stability of increased adenoma detection at colonoscopy. Follow-up of an endoscopic quality improvement program-EQUIP-II. Am J Gastroenterol 2015;110:489-96.
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 7
Colonoscopy practice for veterans within and outside the Veterans Affairs setting: a matched cohort study Michael J. Bartel, MD,1 Douglas J. Robertson, MD,2 Heiko Pohl, MD2,3 White River Junction, Vermont, USA
Background and Aims: To minimize delays for colonoscopy within Veterans Affairs (VA) facilities, veterans may receive care at non-VA facilities based on fee-for-service contracts, and more recently, through the Veterans Access, Choice, and Accountability Act. The impact of diverting care from VA to non-VA facilities on quality of colonoscopy practice is unknown. Methods: We identified all veterans aged 50 to 85 years who received a fee-basis colonoscopy for colorectal cancer screening or surveillance at non-VA facilities in 2007 to 2010. These patients were matched for sex, age, and year of procedure to veterans who underwent colonoscopies at VA medical centers. The outcomes of interest were the adenoma detection rates (ADR) and compliance with surveillance guidelines. Results: During the observation period, 409 veterans (mean age 64 years; 94% men) underwent a fee-basis colonoscopy at 30 nonacademic (54%) and 2 academic (46%) facilities. Compared with colonoscopies performed at VA facilities, fee-basis colonoscopy patients had lower ADRs (38% vs 52%; P < .001), lower mean number of adenomas per procedure (0.72 vs 1.41; P < .001), and lower number of advanced ADRs (13% vs 22%; P < .001). Colonoscopies done at non-VA facilities were associated with lower ADRs in multivariate regression analysis (odds ratio 0.64; 95% CI, 0.44-0.92), whereas colonoscopies done in nonacademic settings or by colonoscopists who were not gastroenterologists were not. Compliance with surveillance guidelines was lower for colonoscopies performed outside VA facilities (80% vs 87%; P Z .03). Conclusions: In this regional study (Northern New England), compliance with colonoscopy surveillance guidelines was high in both VA and non-VA settings; however, lower ADRs raise concern that referring veterans outside the VA system may impact colonoscopy quality. (Gastrointest Endosc 2016;-:1-7.)
Abbreviations: ADR, adenoma detection rate; MAP, mean number of adenomas per procedure; SSA/P, sessile serrated adenoma and/or polyp; VA, Veterans Affairs; VHA, Veterans Health Administration. DISCLOSURE: This material is the result of work supported in part by resources from The Veterans Health Administration. The views expressed in this publication are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. All authors disclosed no financial relationships relevant to this publication. Copyright ª 2016 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2016.01.017 Received August 20, 2015. Accepted January 7, 2016. Current affiliations: Department of Gastroenterology and Hepatology, Mayo Clinic, Jacksonville, Florida (1), Department of Gastroenterology, White River Junction VA Medical Center, White River Junction, Vermont (2), Geisel School of Medicine at Dartmouth, Lebanon, New Hampshire (3). Reprint requests: Michael J. Bartel, MD, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL 32224.
www.giejournal.org
The Veterans Health Administration (VHA) endorsed screening colonoscopy as a primary screening option for all veterans in 2007.1 Although fecal occult blood tests and sigmoidoscopy have been alternative screening options, colonoscopy has become the primary screening modality in and outside the Veterans Affairs (VA) system.2-4 However, as the single largest health care provider in the United States, the VHA has been unable to meet the demands for screening and surveillance colonoscopies. In order to ensure access to care, VA facilities occasionally use fee-for-service contracts with non-VA facilities. Through such mechanisms, veterans are able to receive colonoscopies outside of a VA facility at VHA expense. Of approximately 300,000 colonoscopies performed annually under VHA care, 100,000 procedures are completed by non-VA providers on a fee-for-service basis (personal communication with Jason Dominitz, VHA National Gastroenterology Program Director, July 2015). In August 2014, Congress passed The Veterans Access, Choice, and Accountability Act of 2014 (Choice Act), which Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 1
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
allows for care outside of the VHA for veterans “who either cannot be seen within the wait-time goals of the VHA or who qualify based on their place of residence.”5 Although they are similar in some aspects, the national VA health care system differs from the private practice setting in many ways. In both systems, physician practice is dictated by societal guidelines with respect to colorectal cancer screening and surveillance. In addition, the VHA, as a centralized health care provider, issues directives to guide care and mandate quality benchmarks.1 In a private practice setting, care is based on local preference at the facility. Financial incentives also play a greater role in private practice than within a VA setting. In addition, the veteran population differs from the general population with regard to adenoma risk factors (eg, male sex), and a higher adenoma prevalence should be expected.6 It is possible, therefore, that the system in which a physician works may have an effect on his or her practice, and that colonoscopy performance may differ between private practice and that done at VA medical centers. It is unclear whether colonoscopy quality is affected by referring patients to non-VA providers. Therefore, we examined colonoscopy practice performed outside of and within a VA setting and measured 2 main quality metrics: adenoma detection and compliance with surveillance guidelines.7 Adenoma detection is strongly associated with interval colorectal cancers, and, therefore, is considered a main quality indicator.8,9 Previous studies have shown that compliance with surveillance colonoscopy guidelines was poor in a non-VA setting and showed overutilization for patients with low-risk findings and underutilization for those with high-risk findings.10-14 Our objective was to compare adenoma detection and compliance with surveillance guidelines between veterans undergoing a fee-for-service colonoscopy outside of the VA system and those who received a colonoscopy at a VA medical center.
performed by trainees were supervised by attending gastroenterologists. Patients with concomitant lower GI symptoms, positive fecal occult blood test results, or histories of inflammatory bowel disease were excluded in order to avoid substantial variations of the adenoma detection rate (ADR). Patients who underwent fee-basis colonoscopies were matched for sex, age (according to the year of birth), and year of procedure, 1-to-1 with patients who underwent screening or surveillance colonoscopies at VA medical centers (VA cohort). The proportion of fee-basis colonoscopies between 2007 and 2010 was approximately 10%. We assessed, therefore, only every fifth consecutive patient undergoing a colonoscopy at a VA facility in each of the 4 years of eligibility (age and sex, in addition to year). If a patient did not meet matching criteria, the next fifth consecutive patient was identified. This process was repeated until the VA-matched control group was complete. The study was approved by the institutional review board at Dartmouth College.
Data collection
METHODS
Data from the VA cohort were abstracted by retrospective chart review for patient age, sex, demographics, procedure indication, colonoscopy findings, polyp characteristics, and histology, as well as recommended surveillance interval given by the endoscopist. Similar data were abstracted in the fee-basis cohort from available colonoscopy, pathology reports, and postprocedure correspondence. In the case of >1 polyp, detailed polyp characteristics (ie, size and histology) were recorded for the most advanced neoplastic polyp. Based on indication and colonoscopy findings, the appropriate surveillance interval was determined, applying the U.S. Multi-Society Task Force guidelines current at the time of the procedure.15,16 For all surveillance colonoscopies, previous colonoscopy findings were also considered when the appropriate surveillance interval was determined.
Patients
Outcomes
We identified all patients aged 50 to 85 years who were referred for a screening or a surveillance fee-basis colonoscopy from the White River Junction VA Medical Center, White River Junction, Vermont, to non-VA facilities between 2007 and 2010. The decision for fee-basis referral was based on increased demand at 3 time points during these years. Only then were patients offered a referral to an outside provider. Referred patients underwent a feebasis colonoscopy based on personal preferences in selfchosen local facilities typically close to their home addresses. These non-VA facilities included both academic and non-academic settings, and both gastroenterologists and surgeons performed colonoscopies (fee-basis cohort). Involvement of gastroenterology trainees at academic centers was not an exclusion criterion. All colonoscopies
As the primary outcome of interest, we obtained the ADR, defined as the proportion of patients with at least 1 adenoma. In addition, we calculated the mean number of adenomas per procedure (MAP). We further obtained the advanced ADR, defined as patients with 1 large adenomas (10 mm), 1 adenomas with advanced histology (villous component, high-grade dysplasia, or cancer), or 3 small adenomas. For the purpose of this study, sessile serrated adenomas and/or polyps (SSA/P) and tubular adenomas were combined as adenomas. The second main outcome of interest was the proportion of recommendations compliant with the U.S. Multi-Society Task Force guidelines.15,16 We first obtained the proportion of patients who received a surveillance recommendation. We then calculated the compliance rate by correctly
2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
www.giejournal.org
Bartel et al
assigned surveillance colonoscopy intervals among all patients who received a recommendation. We further obtained the proportion of patients who received early or late surveillance interval recommendations. In addition, outcomes were stratified by academic affiliation (academic vs non-academic) and endoscopist specialty (gastroenterologist vs surgeon).
Statistical analysis Continuous variables are presented as means with standard deviation (SD) if normally distributed or as medians with interquartile range (IQR) if not normally distributed. Proportions are presented as percentage with a 95% confidence interval (CI). Categorical data for both cohorts were analyzed by using a chi-square test. Categorical variables (including the ADR) were analyzed by using the chi-square test. Continuous variables were compared by using the non-paired t test if normally distributed and the Wilcoxon Mann-Whitney test if not normally distributed. We further examined possible factors associated with adenoma detection by using logistic regression, considering age, sex, setting, endoscopist specialty, and adenoma detection. Stepwise multivariable models were used to account for potentially confounding relationships between variables. Results were rounded to the nearest percentage point to facilitate text presentation; the tables provide more precision.
RESULTS Patient characteristics Between 2007 and 2010, a total of 409 patients (mean age 64, 95% male) underwent a fee-basis colonoscopy (fee-basis cohort) for either screening (68%) or surveillance (32%). Colonoscopies were performed by gastroenterologists (63%) or surgeons (37%) at 30 nonacademic (59%) and 2 academic facilities (41%). The control group (VA cohort) comprised 409 patients of similar age and sex (mean age 64 years, 95% male). Colonoscopies in the control group were more often performed for surveillance (39%) than in the non-VA group (32%; P < .05), with the remaining being performed for screening (61%). Eleven gastroenterologists with academic affiliations performed all colonoscopies in the control group (Table 1).
Adenoma detection Table 2 details colonoscopy findings. Patients undergoing colonoscopies at non-VA facilities had lower ADRs (38%) than those receiving colonoscopies at the VA medical center (52%; P < .001). Only about half as many adenomas were found per procedure in the fee-basis cohort compared with the VA cohort (0.7 vs 1.4; P < .001). Similarly, the advanced ADR was lower outside of the VA facilities than within the VA facilities (13% vs 22%; P < .001). Additional subgroup analysis www.giejournal.org
Colonoscopy outcomes in Veterans Affairs setting
TABLE 1. Patient baseline characteristics VA Fee-basis P (n [ 409) (n [ 409) value Patients Age, mean ( SD), y
63.7 (7.9)
63.7 (8.1)
NS
Male, no. (%)
387 (94.6) 387 (94.6)
NS
Colonoscopy indication, no. (%)
.041
Screening
250 (61.1)
278 (68)
Surveillance
159 (38.9)
131 (32) < .001
Endoscopist specialty, no. (%) Gastroenterologist
409 (100) 256 (62.6)
Surgeon
0
153 (37.4) < .001
Academic affiliation, no. (%) Academic
409 (100) 168 (41.1)
Nonacademic Polyps, no. Adenomas, no. Adenoma size, median (IQR), mm Adenoma 10 mm, no. (%)
0
241 (58.9)
901
520
< .001
578
296
< .001
4 (3-6)
5 (3-7)
NS
36 (16.8)
29 (18.5)
NS
Adenoma histology, no. (%) 563 (64.5) 275 (52.9) < .001
Tubular adenoma Villous component*
11 (1.2)
11 (2.1)
NS
Sessile serrated adenoma and/or polyp
1 (0.1)
9 (1.7)
.02
Polyp with high-grade dysplasia or cancer
3 (0.3)
2 (0.4)
NS
VA, Veterans Affairs; SD, standard deviation; NS, not significant; IQR, interquartile range. *Tubulovillous adenoma and villous adenoma.
demonstrated that gastroenterologists from VA facilities yielded significantly higher ADRs (52.3% vs 38.3%), MAP (1.41 vs 0.73), and advanced ADRs (22.2% vs 13.2%) compared with their fee-basis counterparts. Comparable results were found when only endoscopists with academic affiliations were taken into account. No significant differences for ADR, MAP, and advanced ADR were detected between gastroenterologists and surgeons or between academic and non-academic affiliated practices within the fee-basis cohort. In multivariate analysis, indication and colonoscopy performance at non-VA facilities were independently associated with adenoma detection (Table 3). Veterans who underwent colonoscopies at non-VA facilities had a 36% lower chance of being diagnosed with an adenoma (odds ratio [OR] 0.64; 95% CI, 0.44-0.92) compared with those undergoing colonoscopies at VA facilities. Although endoscopist specialty and academic affiliation were associated with adenoma detection in univariate analysis, these factors were not significant after adjusting for indication, sex, and referral setting. Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 3
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
TABLE 2. Polyp characteristics VA (n [ 409)
Fee-basis (n [ 409)
P value
Polyp detection rate, %
69.2
51.6
< .001
Adenoma detection rate, %
52.3
38.4
< .001 < .001
Specialty Gastroenterologist
52.3
38.3*
–
38.6
52.3
40.5*
Nonacademic
–
36.9
Mean adenomas per procedure, no.
1.41
0.73
1.41
0.78*
–
0.64
1.41
0.84*
Nonacademic
–
0.65
Advanced adenoma detection rate, %
22.2
13.2
22.2
14.4*
–
11.1
Surgeon Setting
< .01
Academic < .001 < .01
Specialty Gastroenterologist Surgeon
< .001
Setting Academic
< .001 < .02
Specialty Gastroenterologist Surgeon Setting Academic Nonacademic
.066 22.2
15.5*
–
13.1
VA, Veterans Affairs. *No significant difference between gastroenterologist and surgeon or academic and nonacademic affiliation.
Compliance with surveillance guidelines Recommendations for surveillance examination were less frequently documented for non-VA than for VA colonoscopies (74% vs 93%; P < .001). Among patients who received recommendations, those provided at non-VA sites were less frequently compliant with guidelines than those given at the VA site (80% vs 87%; P Z .03). This difference was primarily related to recommending a shorter interval, which was observed for 13% and 6% in the non-VA and the VA group, respectively (P Z .004). A longer interval than suggested by guidelines was similar (7%) in both settings. Among all non-VA colonoscopies, recommendations were less frequently compliant with guidelines if the procedure was performed in a non-academic setting or by a surgeon (Table 4). In multivariate analysis, an indication of screening was the only factor independently associated with stronger compliance with guidelines (OR 2.03; 95% CI, 1.33-3.11). Although academic affiliation, gastroenterology specialty, and colonoscopy performance at a non-VA site were associated with compliance with univariate analysis, these 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
factors were no longer independently associated with compliance after adjustment in multivariate analysis (Table 5).
DISCUSSION In order to meet colonoscopy demand, the VHA allows veterans to obtain care outside of the VA system based on a fee-for-service contract, and more recently as part of the Choice Act.5 Focusing on adenoma detection and compliance with surveillance guidelines, we found a significantly lower ADR, MAP, and advanced ADR for veterans who received a fee-basis colonoscopy outside of the VA system in comparison with veterans who underwent colonoscopies within the VA system. Similarly, surveillance recommendations were given less frequently after colonoscopies performed outside of the VA and were less frequently compliant with U.S. Multi-Society Guidelines; however, this association was not significant in multivariate analysis. Differences in adenoma detection within the same health care systems or local practices are well-described and felt to be primarily attributed to the patient population and the quality of the technical equipment and skill of the endoscopists.17-19 We found a significant difference in adenoma detection in the same patient population dependent on where patients had the colonoscopydat a non-VA practice or a VA medical center. Adenoma detection is an important marker of colonoscopy quality.8,9,19,20 Recent studies have shown a strong association between adenoma detection and the risk of interval cancer.8,9 In a study by Corley et al,8 each 1% increase in ADR was associated with a 3% decrease in the risk of interval cancer. In our study, we found a surprisingly high 14% absolute difference between the non-VA and the VA groups. Some of the observed difference may be attributed to nonmeasured confounders. However, the groups were well-matched, and it seems more likely that something other than patient factors are responsible for the differences in adenoma detection. Academic affiliation or endoscopist specialty (ie, surgeon) were not associated with adenoma detection. Broad variation in adenoma detection across endoscopists has been described before, and is likely responsible for some of the observed difference in adenoma detection.21,22 Further, it is possible that system factors may contribute to adenoma detection. For instance, veterans have a higher adenoma risk than the general population.6 Therefore, endoscopists performing colonoscopies at VA medical centers can expect a higher ADR than endoscopists practicing in the community. This may set different baseline expectations (VA vs non-VA) and affect adenoma detection in the same risk group (veterans). Another important difference is practice organization. VA directives guide care and set benchmarks for quality.1 In contrast, www.giejournal.org
Bartel et al
Colonoscopy outcomes in Veterans Affairs setting
TABLE 3. Associations with adenoma detection rate Univariate
Multivariate
Odds ratio
P value
Odds ratio
P value
Male
1.33 (0.72-2.5)
NS
1.25 (0.66-2.35)
NS
Colonoscopy indication (screening)
0.62 (0.46-0.82)
.001
0.64 (0.48-0.86)
.003
Facility (academic)
1.63 (1.2-2.22)
.002
1.29 (0.78-2.15)
NS
Endoscopist (gastroenterologist)
1.41 (0.98-2.02)
.061
0.87 (0.52-1.45)
NS
Referral (fee-basis)
0.57 (0.43-0.75)
< .001
0.64 (0.44-0.92)
.017
NS, Not significant.
care in private practice is largely based on local preferences. Financial incentives also play a greater role for the endoscopists in the private setting than within the VA setting. The culture in which colonoscopies are performed may, therefore, result in different outcomes. Although our study reflects only regional practice within a VA referral area in New England, it should alert policy makers that referring veterans outside the VHA may negatively impact colonoscopy quality. In addition to adenoma detection, documentation of surveillance recommendations and compliance with surveillance guidelines are considered important quality metrics.7 Documentation of surveillance recommendations in our study would have reached the recently suggested 90% quality benchmark for colonoscopies performed within the VHA (93%), but not outside the VHA (74%). We were surprised to find an overall high compliance with surveillance guidelines in >80% of colonoscopies when documented. This is a higher rate than reported in previous studies within and outside of the VA system.10,12,14,23,24 Recently, Johnson et al23 examined 25 VA facilities and found an overall compliance of 64%, ranging from 20% to 97%. In our study, recommendations outside of the VHA appeared to be slightly less compliant than those given within the VHA; however, this difference was no longer significant in adjusted analysis. Similar to those of previous studies, noncompliant recommendations were evenly balanced between a too-short and a too-long interval.11,13,23-26 A factor that is frequently associated with poor compliance with guidelines is inadequate bowel preparation.27 More concerning, Saini et al28 reported that the endoscopists contribute significantly to poor compliance; 47.5% did not know the guidelines, and 51% to 76% disagreed with guidelines and recommended earlier surveillance colonoscopies. In a VA study, depending on the clinical question, 4% to 42% of gastroenterologists were not familiar with surveillance guidelines.13 We did not analyze whether knowledge of guidelines contributed to the observed high compliance rate in our study. We found indication for screening as the sole factor associated with high compliance. This is not surprising because the absence of a history of adenoma simplifies recommendations. www.giejournal.org
TABLE 4. Compliance with surveillance guidelines VA Fee basis (n [ 409) (n [ 409)
P value
Recommendations available, no. (%)
380 (92.9)
301 (73.6)
< .001
Surveillance interval in agreement with guidelines, no. (%)
329 (86.6)
242 (80.4)
.03
Compliance
Specialty
NS
Gastroenterologist, % Surgeon, %
86.6
84
–
71.9
86.6
83.8
Setting
< .001* NS
Academic, %
–
77.4
< .001y
Shorter surveillance interval recommended, no. (%)
24 (6.3)
38 (12.6)
.004
Longer surveillance interval recommended, no. (%)
27 (7.1)
21 (7)
NS
Nonacademic, %
NS, Not significant. *Compared with fee-basis gastroenterologists. yCompared with fee-basis academic affiliation.
Several limitations of our study should be noted. First, the retrospective character and its narrow regional restriction raise concerns for selection bias. We minimized this bias by matching controls by age, sex, and procedure year. We further adjusted our analysis for other factors, including indication, with a potential effect on the outcome. However, it cannot be excluded that unmeasured patient risk factors (eg, history of smoking) may account for some of the observed differences. Some may argue that patients referred for a fee-basis colonoscopy may be different from those receiving a colonoscopy at a VA facility. However, the decision was not based on patient preference (and self-selection) but was made by the VA medical center at specific time points during the observation period. When the waiting time for routine colonoscopies became too long, patients were offered a colonoscopy outside the VA system. Outside of these times, fee-basis referral was generally not offered to patients. Likely, longer wait times for referred patients may result in an increased adenoma prevalence in this group. The obtained difference in adenoma detection between the groups may, Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 5
Colonoscopy outcomes in Veterans Affairs setting
Bartel et al
TABLE 5. Associations with compliance with surveillance guidelines Univariate
Multivariate
Odds ratio
P value
Odds ratio
P value
1.08 (0.44-2.66)
NS
1.37 (0.55-3.44)
NS
Colonoscopy indication (screening)
1.96 (1.3-2.96)
.001
2.04 (1.33-3.11)
.001
Facility (academic)
1.77 (1.14-2.77)
.011
1 (0.48-2.08)
NS
Endoscopist (gastroenterologist)
2.33 (1.39-3.9)
.001
1.94 (0.92-4.1)
NS
Referral (fee-basis)
0.64 (0.42-0.96)
.03
0.74 (0.43-1.28)
NS
1 (0.67-1.5)
NS
1 (0.65-1.53)
NS
Male
Adenoma detection NS, Not significant.
therefore, be biased toward a lower difference. A selection bias seems less likely. Second, we compared colonoscopy practice at multiple non-VA sites with colonoscopy practice at 1 VA medical center. It is possible that unmeasured factors contributed to the obtained differences in ADR, including individual endoscopists’ ADRs or participation in quality improvement projects.6,9,29 Our results therefore provide only an overall impression on regional colonoscopy practice and are not generalizable to other regions or VA facilities. However, this study can be seen as a natural experiment. Patients who were referred for a fee-basis colonoscopy chose where to have the procedure. To the extent that patient choice is simply driven by geography (and not quality), fee-basis colonoscopy should reflect the average clinical practice in the region. Third, other colonoscopy quality metrics were not analyzed, including cecal intubation rate, colonoscopy withdrawal time, or bowel preparation. These factors might have provided additional insights for the observed differences in adenoma detection in particular and colonoscopy performance in general. With passage of the Choice Act in 2014, VA centers are encouraged to outsource colonoscopies to local non-VA providers to improve access. Our study reflects regional practice in a VA referral region in New England. Veterans who were referred to local non-VA providers on a fee-forservice basis had a lower ADR, MAP, and advanced ADR than veterans undergoing a colonoscopy at their local VA medical centers. Compliance with surveillance guidelines was high in both settings. These findings raise concerns that referring veterans outside the VA may impact quality.
REFERENCES 1. Department of Veterans AffairsdColorectal Cancer Screening, VHA Directive 1015. 2014. Available at: http://www.va.gov/vhapublications/ ViewPublication.asp?pub_IDZ3068. Accessed July 1, 2015. 2. El-Serag HB, Petersen L, Hampel H, et al. The use of screening colonoscopy for patients cared for by the Department of Veterans Affairs. Arch Intern Med 2006;166:2202-8. 3. Harewood GC, Lieberman DA. Colonoscopy practice patterns since introduction of medicare coverage for average-risk screening. Clin Gastroenterol Hepatol 2004;2:72-7.
6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
4. Seeff LC, Richards TB, Shapiro JA, et al. How many endoscopies are performed for colorectal cancer screening? Results from CDC’s survey of endoscopic capacity. Gastroenterology 2004;127:1670-7. 5. Department of Veterans Affairs. Expanded access to non-VA care through the Veterans Choice Program. Interim final rule. Fed Regist 2014;79:65571-87. 6. Kahi CJ, Ballard D, Shah AS, et al. Impact of a quarterly report card on colonoscopy quality measures. Gastrointest Endosc 2013;77:925-31. 7. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for colonoscopy. Am J Gastroenterol 2015;110:72-90. 8. Corley DA, Jensen CD, Marks AR, et al. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014;370: 1298-306. 9. Kaminski MF, Regula J, Kraszewska E, et al. Quality indicators for colonoscopy and the risk of interval cancer. N Engl J Med 2010;362: 1795-803. 10. Iskandar H, Yan Y, Elwing J, et al. Predictors of poor adherence of US gastroenterologists with colonoscopy screening and surveillance guidelines. Dig Dis Sci 2015;60:971-8. 11. Menees SB, Elliott E, Govani S, et al. Adherence to recommended intervals for surveillance colonoscopy in average-risk patients with 1 to 2 small (<1 cm) polyps on screening colonoscopy. Gastrointest Endosc 2014;79:551-7. 12. Schreuders E, Sint Nicolaas J, de Jonge V, et al. The appropriateness of surveillance colonoscopy intervals after polypectomy. Can J Gastroenterol 2013;27:33-8. 13. Shah TU, Voils CI, McNeil R, et al. Understanding gastroenterologist adherence to polyp surveillance guidelines. Am J Gastroenterol 2012;107:1283-7. 14. Sint Nicolaas J, de Jonge V, van Baalen O, et al. Optimal resource allocation in colonoscopy: timing of follow-up colonoscopies in relation to adenoma detection rates. Endoscopy 2013;45:545-52. 15. Rex DK, Kahi CJ, Levin B, et al. Guidelines for colonoscopy surveillance after cancer resection: a consensus update by the American Cancer Society and the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2006;130:1865-71. 16. Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US MultiSociety Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 2008;134:1570-95. 17. Benson ME, Reichelderfer M, Said A, et al. Variation in colonoscopic technique and adenoma detection rates at an academic gastroenterology unit. Dig Dis Sci 2010;55:166-71. 18. Shaukat A, Oancea C, Bond JH, et al. Variation in detection of adenomas and polyps by colonoscopy and change over time with a performance improvement program. Clin Gastroenterol Hepatol 2009;7: 1335-40. 19. Coe SG, Crook JE, Diehl NN, et al. An endoscopic quality improvement program improves detection of colorectal adenomas. Am J Gastroenterol 2013;108:219-26; quiz 27.
www.giejournal.org
Bartel et al 20. Lieberman DA, Rex DK, Winawer SJ, et al. Guidelines for colonoscopy surveillance after screening and polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer. Gastroenterology 2012;143:844-57. 21. Chen SC, Rex DK. Endoscopist can be more powerful than age and male gender in predicting adenoma detection at colonoscopy. Am J Gastroenterol 2007;102:856-61. 22. Pohl H, Bensen SP, Toor A, et al. Cap-assisted Colonoscopy and Detection of Adenomatous Polyps (CAP) study: a randomized trial. Endoscopy 2015;47:891-7. 23. Johnson MR, Grubber J, Grambow SC, et al. Physician non-adherence to colonoscopy interval guidelines in the Veterans Affairs Healthcare System. Gastroenterology 2015;149:938-51. 24. Schoen RE, Pinsky PF, Weissfeld JL, et al. Utilization of surveillance colonoscopy in community practice. Gastroenterology 2010;138: 73-81.
www.giejournal.org
Colonoscopy outcomes in Veterans Affairs setting 25. Partin MR, Noorbaloochi S, Grill J, et al. The interrelationships between and contributions of background, cognitive, and environmental factors to colorectal cancer screening adherence. Cancer Causes Control 2010;21:1357-68. 26. Radaelli F, Paggi S, Bortoli A, et al. Overutilization of post-polypectomy surveillance colonoscopy in clinical practice: a prospective, multicentre study. Dig Liver Dis 2012;44:748-53. 27. Menees SB, Elliott E, Govani S, et al. The impact of bowel cleansing on follow-up recommendations in average-risk patients with a normal colonoscopy. Am J Gastroenterol 2014;109:148-54. 28. Saini SD, Nayak RS, Kuhn L, et al. Why don’t gastroenterologists follow colon polyp surveillance guidelines? Results of a national survey. J Clin Gastroenterol 2009;43:554-8. 29. Ussui V, Coe S, Rizk C, et al. Stability of increased adenoma detection at colonoscopy. Follow-up of an endoscopic quality improvement program-EQUIP-II. Am J Gastroenterol 2015;110:489-96.
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 7