Delivering colonoscopy screening for low-income populations in Suffolk County
Strategies, outcomes, and benchmarks
The articles in this supplement were commissioned based on participation in evaluating the Centers for Disease Control and Prevention-funded Colorectal Cancer Screening Demonstration Program.
We also thank all the collaborators of the Stony Brook University Medical Center SCOPE (Suffolk County Preventive Endoscopy) Project, who are listed elsewhere in this supplement.
The opinions or views expressed in this supplement are those of the authors and do not necessarily reflect the opinions or recommendations of the journal editors, the American Cancer Society, John Wiley & Sons Inc, or the Centers for Disease Control and Prevention.
Current and pending legislation provides colorectal cancer screening reimbursement for previously uninsured populations. Colonoscopy is currently the screening method most frequently recommended by physicians for insured patients. The experience of the SCOPE (Suffolk County Preventive Endoscopy) demonstration project (Project SCOPE) at Stony Brook University Medical Center provides a model for delivering colonoscopy screening to low-income populations to meet anticipated increasing demands.
Project SCOPE, based in the Department of Preventive Medicine, featured internal collaboration with the academic medical center's large gastroenterology practice and external collaboration with the Suffolk County Department of Health Services' network of community health centers. Colonoscopies were performed by faculty gastroenterologists or supervised fellows. Measures of colonoscopy performance were compared with quality indicators and differences between faculty and supervised fellows were identified.
During a 40-month screening period, 800 initial colonoscopies were performed. Approximately 21% of women screened were found to have adenomatous polyps compared with 36% of men. Five cancers were detected. The majority of the population screened (70%) were members of minority populations. African American individuals had a higher percentage of proximally located adenomas (78%) compared with white individuals (65%) and Hispanics (49%), based on the location of the most advanced lesion. Hispanic individuals had a 36% lower risk of adenomas compared with white individuals. Performance measures including the percentage of procedures with adequate bowel preparation, cecum reached, scope withdrawal time, and adenoma detection rate met quality benchmarks when performed by either faculty or supervised fellows.
Project SCOPE's operational strategies demonstrated a feasible method for an academic medical center to provide high-quality screening colonoscopy for low-income populations. Cancer 2013;119(15 suppl):2842–8. © 2013 American Cancer Society.
At the time that the Centers for Disease Control and Prevention (CDC)'s Colorectal Cancer Screening Demonstration Program (CRCSDP) commenced in 2005, national guidelines for colorectal cancer screening recommended any of the following 4 methods: fecal occult blood testing (FOBT), sigmoidoscopy, colonoscopy, and double-contrast barium enema.[1, 2] However, given its low sensitivity for large polyps, the use of double-contrast barium enema for colorectal cancer screening is rare. In its latest guidelines, the U.S. Preventive Services Task Force reduced the number of recommended screening tests to the remaining suggestion of 3. The most recent joint guidelines from the American Cancer Society, the U.S. Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology divide the screening methods into cancer prevention, such as colonoscopy, and cancer detection tests. The use of colonoscopy increased rapidly after Medicare began providing coverage for colorectal cancer screening in 2001, by any of the recommended methods, including colonoscopy.[7, 8] The increase in colonoscopy use was accompanied by a nearly equivalent decline in the use of sigmoidoscopy for screening,[4, 7, 8] whereas the use of FOBT also decreased more gradually. Despite the finding that national guidelines continue to offer a choice and that the several options continue to be used, common reasons given for the preference for colonoscopy among referring physicians include it is a single-step process that can encompass screening, diagnosis, and resection of polyps; a longer length of the colon can be observed through the colonoscope than through the sigmoidoscope; it has the longest interval between recommended screening for patients at average risk (10 years) and the related lack of dependence on patient adherence to annual FOBT; freeing the patient from handling stool specimens; and recognition of the inadequate tracking of FOBT kits that are not returned by the patient. The higher recommendation rate along with patient completion and consent suggest that these advantages of colonoscopy are believed by physicians and patients to outweigh its higher cost, the greater risk of complications from the procedure, the need for arduous patient bowel preparation (bowel preparation is also required for sigmoidoscopy), and sedation. Patient preferences for colonoscopy have been found to be attributed most frequently to its sensitivity. In part because of the cost, there are disparities in referral for colonoscopy between insured individuals and those without insurance; those who lack insurance coverage are commonly denied access to this colorectal cancer screening option.[10, 11] For this reason in particular, our project selected colonoscopy as our screening method.
The passage of the Patient Protection and Affordable Care Act, which provides health care coverage for currently uninsured low-income populations, coupled with the proposed national legislation that would add colorectal cancer screening to CDC-funded state programs for breast and cervical cancer screening, would make colonoscopy for colorectal cancer screening more widely accessible, and within the reach of the disadvantaged. Our CRCSDP, which is based in an academic medical center, provides a model for delivering colonoscopy for colorectal cancer screening to low-income populations as well as strategies, outcomes, and benchmarks that can be used by other endoscopy practices that are planning to meet similarly increasing demands for colonoscopy screening services from underserved patients.
MATERIALS AND METHODS
As described in greater detail elsewhere, the SCOPE (Suffolk County Preventive Endoscopy) Project (Project SCOPE) provided screening colonoscopy during a single visit to the gastroenterologist in the endoscopy suite.[12, 13] This was preceded by a telephone “visit” conducted by a preventive medicine physician (lead public health clinician) prior to colonoscopy to determine medical eligibility, ensure that the patient's status was appropriate with respect to comorbidity and medications, and provide patient education and scheduling for the procedure. A copy of the assessment tool for medical eligibility is available elsewhere. Colonoscopy was delayed for patients determined to have medical problems based on discussions between the lead public health clinician and the gastroenterologist and/or the referring physician (eg, for current comorbid symptoms such as chest pain). Bilingual patient navigators facilitated patient completion of colonoscopy by providing language and culturally appropriate support services including arrangements for transportation, having the patient identify an escort to accompany them home after sedation, reinforcing instructions for the bowel preparation, and navigating the patient to the endoscopy suite on the day of screening.
The primary target population were uninsured and underinsured registrants of the Suffolk County Department of Health Services 10 community health centers, which are located in underserved communities throughout the county. As described elsewhere, a system for referral was developed with the lead public health clinician serving as the liaison between the primary care physician and endoscopy services.[12, 13] This system included direct scheduling of colonoscopy at the time of the precolonoscopy telephone visit within a block of appointments reserved for project patients for a time-limited period.
Project-funded screening was limited to asymptomatic individuals without personal risk factors for colorectal cancer. Those with a family history of colorectal cancer or polyps were considered eligible but those with high-risk factors (eg, genetic syndromes such as familial adenomatous polyposis, a personal history of polyps and colorectal cancer) or those with gastrointestinal symptoms were referred directly to the Division of Gastroenterology and Hepatology for surveillance or diagnosis.
Approximately 90% of the uninsured and underinsured patients screened were aged between 50 years and 64 years, 7% were aged ≥ 65 years, and 3% were aged ≤ 49 years (nearly all of whom had a family history of polyps). There were 840 screening colonoscopies completed during the 40 months of screening, 800 of which were initial colonoscopies and the remaining 40 of which were repeat colonoscopies (the majority performed for surveillance after positive findings). As described elsewhere,[12, 13] the total number of patients screened was limited by the total number of patients referred, the percentage of patients referred who responded to calls from the project, the percentage of respondents who were eligible, and the level of CDC funding. All data presented subsequently are limited to these 800 initial colonoscopies. During the first 18 months of screening, 81% of the 320 colonoscopies (260 colonoscopies) were performed by 1 gastroenterologist (J.A.) as the project's lead endoscopist, and the remainder were performed by gastrointestinal fellows under direct supervision of the same faculty gastroenterologist. During the remaining 22 months of screening, colonoscopies were performed by 11 gastroenterologists, with 1 of them alone performing the largest number (100 colonoscopies) of the remaining 480 procedures.
In addition, all participating gastroenterologists directly supervised colonoscopies performed by fellows. Continuous mentoring during the procedure was achieved by the supervising gastroenterologist who, in the same examination room with the fellow, viewed the procedure on a monitor that projected the images captured by the camera at the end of the colonoscope as it traversed the colon. The fellows performed 391 colonoscopies under the direct supervision of the faculty gastroenterologists during the full 40-month screening period. There were 9 fellows performing screenings during the project period, with 2 of the fellows performing the bulk of these procedures (101 colonoscopies and 147 colonoscopies, respectively) and the remainder performing from 2 colonoscopies to 40 colonoscopies each. Differences between faculty and supervised fellows in terms of quality measures of performance were explored using frequency distributions and cross-tabular analyses. A multivariate logistic regression model was performed to estimate the probability of having an adenoma by race/ethnicity after adjusting for age, sex, and family history of colorectal cancer.
To accommodate the Project SCOPE patients, an average of 4 half-day sessions per week were reserved in the endoscopy schedule by the lead public health clinician. For patient recruitment, it was necessary to offer both morning and afternoon options to accommodate the work schedules of patients and/or their escorts. Conscious sedation was routinely used for Project SCOPE patients as recommended for the screening of average-risk patients. During this CRCSDP, the CDC covered expenses for standard sedation (meperidine and midazolam) but not for propofol. Because anesthesia with propofol was used for other patients undergoing colonoscopy at the study institution, Project SCOPE patients were scheduled at either the beginning or the end of a gastroenterologist's schedule so that the anesthesia team would not have to be in the endoscopy suite.
As is customary with all patients receiving colonoscopy services at Stony Brook University Medical Center, the results were directly communicated to the Project SCOPE patients and to the primary care physician by the gastroenterologist performing the procedure. To further ensure and track appropriate follow-up, project staff confirmed the receipt of the colonoscopy report by the health center provider and sent another copy if needed.
Among the 800 patients who completed their first colonoscopy screening, 37% (298 patients) were found to have polyps. A total of 691 polyps were detected among these 298 patients. Approximately 68% (204 patients) of those patients were found to have 1 to 2 polyps. Among those with polyps, 70% (208 patients) had adenomatous polyps, or 26% of the total number of patients screened. Five cancers were detected: 3 were stage I and 2 were stage III [TNM Tumor Node Mass Classification (American Joint Committee on Cancer: AJCC Staging Manual 7th ed. New York: Springer; 2010).].
Adenoma Detection Rate
Our overall rates of adenoma detection by sex were 21% among women and 35% among men. The rates of adenomas detected by faculty and supervised fellows did not vary significantly (P = .30):for those on the faculty, the rates were 22.5% among women and 39% among men, whereas for fellows, the detection rates were 18% among women and 34% among men.
Consistent with the CDC project reporting system, the Project SCOPE colonoscopy findings shown in Table 1 characterize the patients screened by their most advanced lesion found on endoscopy, and all hyperplastic and proximal serrated polyps (except sessile serrated adenomas) were combined under the 1 hyperplastic category. Among the 208 patients who had neoplastic (adenomatous) polyps, 20 had advanced neoplasia (or 2.5% of total patients screened). These included large (measuring ≥ 1 cm in diameter) adenomas, villous adenomas (any size), polyps with high-grade dysplasia (any size), and adenocarcinoma. Among all 347 adenomas detected, 143 (41%) were distal lesions and 204 adenomas (59%) were proximal. Adenoma location was classified either as proximal (cecum, appendix, ascending colon, hepatic flexure, transverse colon, and splenic flexure) or distal (descending colon, sigmoid, rectosigmoid junction, and rectum). There was some variation in adenoma location by ethnicity. Approximately 78% of the 59 adenomas detected in African American patients were proximal, compared with 65% of the 114 adenomas detected in white patients and 49% of the 129 adenomas detected among the Hispanic patients. The percentage of patients with advanced adenoma also varied by race and ethnicity. Approximately 5% of Hispanic patients (4 of 83 patients) had advanced adenomas compared with 12.5% of African American patients (4 of 32 patients) and 13% of white patients (9 of 71 patients). After multivariate analysis was conducted to control for age, sex, and family history, Hispanic individuals were found to have a reduced risk of having any adenoma (36% lower risk) compared with white individuals, but the difference between African Americans and whites was not significant.
Table 1. Colonoscopy Findings for Initial Examination in 800 Patients
|No polyps||420 (52.5%)|| |
|Normal or other nonpolyp histology/nonadenomatous polypsa||86 (10.7%)|| |
|Hyperplastic polyps||79 (9.9%)|| |
|Tubular adenomas||189 (23.6%)|| |
|Tubular adenoma, <1 cm|| || |
|1-2 tubular adenomas||152|| |
|2 tubular adenomas||30|| |
|Tubular adenoma, ≥1 cm|| || |
|1-2 tubular adenomas||4|| |
|>2 tubular adenomas||0|| |
|Tubular adenoma, size unknown|| || |
|1-2 tubular adenomas||3|| |
|Tubular villous||7 (<1%)|| |
|<1 cm||5|| |
|Size unknown||2|| |
|Villous||2 (<1%)|| |
|<1 cm||1|| |
|≥1 cm||1|| |
|Serrated||3 (<1%)|| |
|<1 cm||1|| |
|≥1 cm||1|| |
|Size unknown||1|| |
|High-grade dysplasia||3 (<1%)|| |
|<1 cm||2|| |
|≥1 cm||0|| |
|Size unknown||1|| |
|Cancer||5 (<1%)|| |
|Adenocarcinoma, invasive||3|| |
|Carcinoma, other||2|| |
|Unknown/other lesions ablated, not retrieved, or confirmed||6 (<1%)|| |
Two of the 5 cancers detected were removed during the screening colonoscopy. The other 3 cancers were referred for surgical resection. All of the cancers detected were located in the distal colon.
Adequacy of Bowel Preparation
Before the existence of Project SCOPE, the perception of the endoscopy service was that health center patients rarely kept their appointments and that inadequate bowel preparation was the norm for these individuals. The services of the patient navigator and lead public health clinician reversed that perception, with a low no-show rate of 3% and > 90% adequate bowel preparations noted.
For 92% of the 800 colonoscopies, the endoscopy reports indicated an adequate bowel preparation (noted as good or excellent in the report). There was little variation noted among faculty providers, although fellows more frequently indicated a lower bowel preparation adequacy rate (86%) compared with faculty (97%) (P < .001). However, this difference was largely due to 1 fellow's assessment of the bowel preparation of the 40 patients he examined (which was the third largest number of colonoscopies performed by a fellow in the current study).
As confirmed by photo documentation, the cecum was reached in 100% of completed colonoscopies performed by faculty gastroenterologists alone and in 95% of those performed by supervised fellows (P < .001). As expected, reaching the cecum was correlated with the quality of the bowel preparation. Among those examinations with adequate bowel preparation, the cecum was reached 99% of the time (725 examinations). When the bowel preparation was not considered to be adequate, the cecum was reached 95% of the time (35 examinations).
Scope Withdrawal Time
For patients with an adequate bowel preparation, no polyps detected, no biopsies, and no complications in whom the cecum was reached, the withdrawal time ranged from 8 to 53 minutes (median, 21 minutes) for faculty gastroenterologists, whereas supervised fellow withdrawal times were between 8 and 68 minutes (median, 20 minutes).
Complete Removal of Polyps
Complete polyp removal could not be determined from 22 endoscopy reports and therefore these were excluded from this analysis. As indicated on the remaining endoscopy and/or pathology reports, 96.4% of polyps were completely removed. The rate of complete polyp removal was 98.7% for faculty and 93.2% for fellows. The majority of the incompletely removed polyps (10 polyps) occurred in patients with multiple polyps who required rescoping within a few months for the complete removal of all polyps. One of the patients had an incomplete removal of a large polyp that required rescheduling for surgical resection. The remaining patients with incomplete removal were rescheduled to return at an interval corresponding to the polyp's pathologic diagnosis (eg, hyperplastic or serrated).
Screening or Surveillance Interval
We compared the Project SCOPE-recommended interval on the endoscopy reports with national guidelines for patients whose colonoscopy was complete to the cecum, who had an adequate bowel preparation, and for whom all identified polyps had been completely removed. Only cases that met those criteria were considered in the following analysis. Patients who were noted to be at high risk also were removed from this analysis.
Of those patients with no polyps, a normal examination, nonpolyp histology, nonadenomatous polyps, or other hyperplastic polyps, 81% were recommended for a 10-year screening or surveillance interval by faculty, and 65% were recommended for the same by the supervised fellows. Nearly all the remaining patients were recommended to return for screening in 5 years to 7 years, and among faculty, one-half of those shortened intervals were recommended by a single endoscopist. For patients with 1 to 2 tubular adenomas (or adenoma not otherwise specified) measuring < 1 cm, with no high-grade dysplasia, the screening or surveillance interval recommendation was 5 years to 10 years for 80% of such patients screened by faculty and 60% screened by supervised fellows. The remaining patients were recommended to return at a shorter interval. For the 19 patients with 3 to 10 tubular adenomas measuring < 1 cm or any adenoma measuring ≥ 1 cm or those with villous features or high-grade dysplasia, a 3-year screening or surveillance interval was recommended by 59% of faculty and 57% of supervised fellows. All the remaining patients were recommended to return at a shorter interval.
Patient Tolerance of the Procedure
Patient tolerance was usually noted on the endoscopy report as poor, good, or excellent. Patient tolerance of the procedure was also assessed during the endoscopy staff's 24-hour call back to the patient to check for complications and the patient navigator's call to the patient 30 days after the procedure to check for any late complications of the procedure and to ask about satisfaction with the colonoscopy screening experience. Overall, there was 1 complication reported among the 800 procedures and, as discussed elsewhere in this supplement, patient satisfaction was overwhelmingly positive.
The experience reported in Project SCOPE demonstrated that an academic gastroenterology practice could accommodate an increase in low-income patients for colonoscopy screening by scheduling appointments 4 half-days a week, accommodating approximately 15 screenings per week, of which we filled a maximum of 14. Our capacity always exceeded the number of actual screening colonoscopies performed. Reliance on predominantly 1 to 3 gastroenterologists, assisted by fellows, was the pattern and appeared to function effectively, with additional faculty coverage needed when the participating gastroenterologist was on hospital service, on vacation, or attending national meetings, or when the patient could only come on a day or time outside the regular 4 half-day sessions.
Our overall rates of adenomatous polyp detection by sex were 21% among women and 35% among men, and compared favorably with a quality indicator of at least 15% adenomatous polyp detection among women and 25% in men, regardless of whether they were performed by faculty or supervised fellows.[14, 15] The higher percentage of adenomas with a proximal location noted among African American patients has also been reported in the literature and has suggested the desirability of colonoscopy versus sigmoidoscopy screening in this minority population[16, 17] We found a lower percentage of advanced adenomas among Hispanic individuals compared with whites, which persisted when controlled for age, sex, and family history through multivariate analysis. This lower total risk among Hispanics has also been reported by others. Reports of colonoscopy screening of asymptomatic patients that included a broader age range and fewer eligibility criteria also found a higher rate of advanced adenomas among men than among women.[19, 20] Imperiale et al found a relative risk of 3.3 for male sex, which we also found.
In addition to the detection rate for adenomas, the performance by both endoscopy faculty and supervised fellows met the quality indicator recommendations of the U.S. Multi-Society Task Force on Colorectal Cancer for bowel preparation, cecum reached, and withdrawal time that follow.[14, 21] A bowel preparation is considered to be adequate if it is possible to detect polyps measuring > 5 mm in size. Our 92% rate of adequate bowel preparation met the quality indicator of ≥ 90%. Reaching the cecum as evidenced by photo documentation is a quality indicator for an adequate depth of insertion, and suggests that a cecal intubation rate of ≥ 95% is a quality benchmark, which was met by our rate of 97%. The quality indicator for withdrawal time is a 6-minute minimum[15, 23] which was met by our 8-minute minimum.
Although complete removal of polyps is a quality indicator, to the best of our knowledge a benchmark for an expected percentage of polyps completely removed in a quality endoscopy practice has not been provided. When complete polyp removal could be determined from the endoscopy and/or pathology reports, 96% of polyps were removed. Patients with incomplete removal who did not require rescoping to remove multiple polyps or surgical resection were rescheduled to return at an interval corresponding to the polyp's pathologic diagnosis, as recommended in the U.S. Multi-Society Task Force on Colorectal Cancer/American Cancer Society guidelines.
The U.S. Multi-Society Task Force on Colorectal Cancer guidelines for screening and surveillance after colonoscopy apply to patients whose colonoscopy was complete to the cecum, who had an adequate bowel preparation, for whom all identified polyps were completely removed, and who were not at high risk due to a family or personal history of polyps.[15, 24, 25] These guidelines recommend a 10-year interval for patients with no polyps, a normal examination, nonpolyp histology, nonadenomatous polyps, or other hyperplastic polyps. A 5-year to 10-year screening and surveillance interval is the recommendation for patients with 1 to 2 tubular adenomas (or adenoma not otherwise specified) measuring < 1 cm, with no high-grade dysplasia. A 3-year screening and surveillance interval is the recommendation for patients with 3 to 10 tubular adenomas measuring < 1 cm or any adenoma measuring ≥ 1 cm or with villous features or high-grade dysplasia. Because the number of patients with advanced adenomas was small overall (20 patients) and particularly the numbers seen by individual endoscopists, our rates for their recommended screening and surveillance interval should be viewed with caution. Although the bulk of recommended intervals for screening and surveillance based on colonoscopy findings were in keeping with the above-mentioned guidelines, the shortened interval recommended for some patients deserves further investigation to uncover reasons beyond those traditionally identified that influence endoscopist decision-making. For example, scheduling patients with serrated adenomas for return screening in 3 years to 5 years is consistent with the recent literature.[26-28] Schreiner et al concluded that the detection of proximal and large hyperplastic and nondysplastic serrated polyps at a screening colonoscopy is associated with an increased risk of synchronous advanced neoplasia. An increased risk of colorectal cancer was also found by Hiraoka et al for large serrated polyps, particularly of the proximal colon. Recognizing the importance of these specific types of serrated polyps, recommendations have been made to update guidelines to indicate a surveillance period of 3 years to 5 years.[28, 29] Although to the best of our knowledge no quality indicator yet exists for the detection of significant serrated lesions, there is some suggestion that the detection of proximal serrated polyps may be correlated with endoscopist and adenoma detection rates. Previously published data demonstrate rates of serrated lesions of 7.9% in the United States. Using the same method for calculating the detection rate as these prior studies, which counted all of the detected serrated lesions (nonadenomatous, proximal, serrated polyps and sessile, serrated adenomas), our detection rate is 11.7% (35 of 298 lesions). A logical reason for the higher rate of shorter screening intervals noted among supervised fellows compared with faculty may be lower confidence related to less experience.
In addition to meeting the quality indicators overall, the small differences in the rates noted between faculty and fellows were not clinically significant. Nevertheless, tracking the quality indicator performance of faculty and fellows is an excellent method for identifying individual “outliers” who need further education about guideline recommendations and should justify their reasons for diverging from them in the endoscopy report.[15, 21] Further discussion of quality performance within the entire CRCSDP is provided elsewhere in this supplement.
The fellows did not compromise the quality of the services delivered and are now better trained to perform colonoscopy when they are in independent practice. These findings should give comfort to patients and program administrators that use or contract with academic medical centers at which fellows receive training. In addition to the clinical skills learned, this demonstration project for an uninsured multiethnic population using county health centers also provided the opportunity for training fellows in 3 of the 6 core competencies required by the Accreditation Council for Graduate Medical Education: systems-based practice, interpersonal communication skills, and professionalism.
In addition to the high-quality performance of the endoscopists, the success of the project was in great measure due to the facilitating role of the lead public health clinician and the patient navigators, who removed barriers by selecting patients who were medically appropriate, decreasing the gastroenterologist visits to one visit, providing intensive training in bowel preparation, and delivering language-appropriate services to overcome health illiteracy. An important positive outcome was our success in reaching the minority population, which made up 70% (562 patients) of the patients screened and 64% (362 patients) of the minority patients seen were Hispanic. The literature documents lower use of colorectal cancer screening in general and specifically the use of colonoscopy by African American and Hispanic patients when compared with white patients. Our experience documents that the removal of access, financial, linguistic, educational, and transportation barriers can reduce, if not eliminate, these disparities. The mission and resources of the academic medical center are ideally suited to meet these needs and to increase colonoscopy screening for low-income populations in their community.
The Colorectal Cancer Screening Demonstration Program evaluated in this supplement was funded by the Centers for Disease Control and Prevention Funding Opportunity Number RFA AA030. The current study was supported by the Centers for Disease Control and Prevention Co-operative Agreement U55/CCU22501701.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.