- Email: [email protected]
Timing of Repeat Colonoscopy
Timing of Repeat Colonoscopy Disparity Between Guidelines and Endoscopists’ Recommendation Alex H. Krist, MD, MPH, Resa M. Jones, MPH, PhD, Steven H. Woolf, MD, MPH, Sarah E. Woessner, MD, Daniel Merenstein, MD, J. William Kerns, MD, Walter Foliaco, MD, Paul Jackson, MD Background: Colonoscopy possesses the highest sensitivity of available screening tests for colorectal cancer and polyps, but it also carries risks. Appropriate intervals for repeating colonoscopy are important to ensure that the benefits of screening and surveillance are not offset by harms. The study objective was to examine whether endoscopists’ recommendations for repeat colonoscopy, as communicated to primary care clinicians after the procedure, adhered to published guidelines. Methods:
Analysts abstracted medical records at ten primary care practices in Virginia and Maryland in 2006. The records of a random sample of men and women (300 per practice) aged 50 to 70 years were reviewed. The sample included patients who had a colonoscopy and a written report from an endoscopist, and who lacked designated risk factors. The main outcome was concordance between endoscopists’ recommendations and published guidelines regarding repeat colonoscopy.
Results:
Of 3000 charts reviewed, 1282 (42.7%) included records of a colonoscopy and 1021 (34%) included an endoscopist’s report. In 64.9% of communications, the endoscopist specified when retesting should occur. Recommendations were consistent with contemporaneous guidelines in only 39.2% of cases and with current guidelines in 36.7% of cases. The adjusted mean number of years in which repeat colonoscopy was recommended was 7.8 years following normal colonoscopy and 5.8 years and 4.4 years, respectively, when hyperplastic polyps or 1–2 small adenomatous polyps were found.
Conclusions: Endoscopists often recommended repeat colonoscopy at shorter intervals than are advised either by current guidelines or by guidelines in effect at the time of the procedure. Endoscopists’ communications to primary care clinicians often lacked contextual information that might explain these discrepancies. Unless appropriate caveats apply, adhering to endoscopists’ recommendations may incur unnecessary harms and costs. (Am J Prev Med 2007;33(6):471– 478) © 2007 American Journal of Preventive Medicine
Introduction
O
f the available screening and surveillance tests for colorectal cancer (CRC), colonoscopy offers the greatest single-test probability for detecting either CRC or polyps.1–5 Colonoscopy also facilitates excision of precancerous lesions, thereby potentially preventing their progression to cancer.6 –11 From the Departments of Family Medicine (Krist, Woolf, Kerns), Epidemiology and Community Health (Jones, Woolf), Virginia Commonwealth University, Richmond; Fairfax Family Practice Residency (Krist, Woessner), Fairfax; Shenandoah Valley Family Practice Residency Program (Kerns), Front Royal; Chesterfield Family Medicine Residency (Foliaco), St. Francis Family Medicine Residency (Jackson), Massey Cancer Center (Jones), Richmond, Virginia; and Department of Family Medicine (Merenstein), Georgetown University, Washington DC Address correspondence and reprint requests to: Alex H. Krist, MD, MPH, 3825 Charles Stewart Drive, Fairfax VA 22033. E-mail: ahkrist@ vcu.edu. The full text of this article is available via AJPM Online at www. ajpm-online.net; 1 unit of Category-1 CME credit is also available, with details on the website.
However, the benefits of colonoscopy are accompanied by risks for bleeding, perforation, and death secondary to the procedure.12–18 Performing colonoscopy too often not only increases patients’ exposure to procedural harm,19 but also drains limited resources that could be more effectively used to bolster the nation’s limited capacity to adequately screen those in need.20 –22 Guidelines promulgated in the early 1990s advocated short follow-up intervals after an abnormal colonoscopy. That changed in 1997, when the profession adopted new consensus guidelines, issued by the United States Multi-Society Task Force on Colorectal Cancer and the American Cancer Society.23 The Task Force issued guideline updates in 200324 and 200625 (see Table 1), and with each iteration the guidelines proposed longer follow-up intervals for some abnormal findings and clarified risk stratification categories. Despite the evidence and consensus of experts articulated in these guidelines, a survey of endoscopists in 2004 found that
Am J Prev Med 2007;33(6) © 2007 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/07/$–see front matter doi:10.1016/j.amepre.2007.07.039
471
Table 1. Recommended intervals for repeat colonoscopy based on colonoscopic findings Finding
1997 recommendations
2003 recommendations24
2006 recommendations25
Normal Hyperplastic polyps 1–2 small (⬍1 cm) tubular adenoma 3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia ⬎10 adenomas on one examination Adenomas that are removed piecemeal Family history of colon cancerd
10 years N.R. N.R.a 3 yearsa
10 years N.R. 5 yearsb 3 years
10 years 10 years 5–10 years 3 years
N.R.c N.R. 5 years
N.R.c N.R. 5 years
⬍3 years 2–6 months 5 years
a
The guideline directly addresses patients with large (⬎1 cm) and “multiple adenomatous polyps.” However, no specific recommendation is made for 1–2 small adenomas. b Recommendation states, “It is not unreasonable to choose even longer intervals.” c Recommendation made only for patients with “multiple adenomatous polyps” without a separate recommendation for ⬎10 adenomatous polyps. d In 2003 and 2006, family history is defined as, “a first-degree relative diagnosed prior to the age of 60 years or two first-degree relatives diagnosed at any age.” In 1997, family history is defined as, “a close relative (sibling parent, or child) who has had colorectal cancer or an adenomatous polyp.” N.R., no recommendation appears in the guideline.
more than 50% of respondents were repeating colonoscopies at shorter intervals than recommended in the 1997 and 2003 guidelines.26 Frequently, the endoscopist’s follow-up recommendations are communicated to the patient’s primary care clinician (PCC) in a note or letter sent after the procedure. The correspondence may be sent before the endoscopist has received all of the final pathology reports on biopsy specimens. Some endoscopists do not update the PCC when the final results become available. Since the PCC is more likely than the endoscopist to have an ongoing relationship with the patient,27 the PCC is often ultimately responsible for ensuring the delivery of appropriate follow-up care years later, when repeat testing is due.28 Not having personally examined the bowel, the PCC must rely on the information and recommendations received from the endoscopist. This communication therefore becomes crucial for ensuring high-quality, long-term care. The objective of this study was to evaluate the extent to which endoscopists’ recommendations on repeat colonoscopy, as conveyed in communications to PCCs, were concordant with contemporaneous and current (2006) guidelines on screening and postpolypectomy surveillance.
Methods Abstractors reviewed the medical records of 3000 randomly selected patients at ten primary care practices. The purpose was to observe the endoscopists’ recommendations through the lens of the PCC by reviewing all communications and information that PCCs received from the endoscopist. The study was not designed to evaluate the information contained in endoscopists’ records, nor was it intended to document when retesting actually occurred.
Setting Five primary care practices from the Virginia Ambulatory Care Outcomes Research Network (ACORN) and five prac-
472
tices from the Capital Area Primary Care Research Network (CAPRICORN) participated in the study. Four of the practices maintained family medicine residency programs, while the others were nonteaching practices. The practices cared for patients in a large geographic area: five in northern VA, one in Front Royal VA, two in Richmond VA, and two in southern MD. No PCCs at the study sites performed colonoscopies themselves. All practices referred patients to endoscopists, the majority of whom were in private practice. The endoscopists were selected based on the preferences of patients and PCCs, convenience to the patient, and insurance restrictions.
Study Sample To derive the sample, the 10 practices generated a list of all patients aged 50 to 70 years who made an office visit between February 2005 and August 2006. Of these, each practice selected 300 patients for chart review using a random sample generator program in STATA. From this sampling frame of 3000 patients, subjects were included in the final study sample if they had ever had a documented colonoscopy (irrespective of date performed) and a communication (e.g., procedure note, office visit note, or letter) from the endoscopist to the PCC reporting the findings. High-risk patients were excluded from the final sample because determining the appropriate interval for repeat colonoscopy required access to historic data that were not collected. Specifically, patients with a personal history of CRC, inflammatory bowel disease, large adenoma, ⬎10 adenomas, villous adenoma, or a prior adenoma with high-grade dysplasia were excluded.
Data Collection Each chart from the sample was examined by one of nine reviewers, who used a standardized abstraction form to record the study data. Reviewers were instructed by the investigators and by each practice on where in a patient’s record to find information about CRC screening. Resources did not permit dual review of charts. All relevant information contained in the PCC record—including correspondence from the endoscopist, pathologist, or hospital—was examined.
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
Outcomes
Table 2. Characteristics of study population (N⫽1006)
Data abstracted from the charts included the patient’s age (at the time of the procedure) and gender, documentation of a prior colonoscopy, any record of family history of CRC in the patient’s chart, presence of a communication from the endoscopist about the procedure, and pathology reports for patients who had a biopsy. When the information was available in the chart or in the endoscopist’s correspondence, reviewers recorded the indication for colonoscopy, the colonoscopy findings, the recommended follow-up tests and testing interval advised by the endoscopist, and (when stated) mitigating factors (e.g., inadequate bowel preparation) that might have influenced the recommendation. If the record included multiple recommendations related to previous colonoscopies, the most recent retesting recommendation from the endoscopist was used.
Characteristic
Findings
Male (%) Mean age (standard error) Proportion of sample who had received prior colorectal cancer screening or diagnostic testing (%) Any test Colonoscopy Sigmoidoscopy Fetal occult blood test Type of provider performing colonoscopy (Nⴝ130 unique endoscopists) (%) Gastroenterologist General surgeon Colorectal surgeon Primary care physician/internist Unknown Indication for performing colonoscopy (%) Screening Surveillance Diagnostic
49.8 57.9 years (0.189)
Analysis Colonoscopy findings were classified as: normal examination; hyperplastic polyp(s); 1–2 small adenoma(s); 3–10 adenomas, large adenoma (ⱖ1 cm in size), or high-grade dysplasia; ⬎10 adenomas; CRC; or unknown polyp type. Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. Similarly, a family history of CRC requires a shorter retesting interval than is indicated after a normal examination or after the detection of hyperplastic polyps or 1–2 adenomas. Therefore, patients with the combination of these findings or unknown polyp type and a family history were evaluated separately from those with the same findings but no family history. Endoscopist recommendations on repeat colonoscopy were compared to both contemporaneous guidelines and to current (2006) guidelines. Colonoscopies performed prior to 1997 were excluded from the contemporaneous comparison, because a consensus guideline from the gastroenterology organizations was not available at that time, but they were included in the comparison with current guidelines. Findings for which contemporaneous guidelines made no explicit recommendations were excluded from the analysis. For example, because the 2003 guideline made no explicit recommendations about follow-up after detecting hyperplastic polyps, the contemporaneous comparison would exclude advice given in 2004 following a colonoscopy that detected such polyps. For all colonoscopies performed across these time periods, the mean length of time (years) in which endoscopists recommended reexamination was calculated for each category of colonoscopy findings. Time to recommended follow-up did not demonstrate departure from normal so linear regression was utilized to determine the adjusted mean years to follow-up by colonoscopy findings.29 Physicians were nested within each practice as a random effect and the means were adjusted for four fixed effects: patient age and gender, primary care practice, and the endoscopist’s specialty. The adjusted means were also stratified by the indication for performing the test (screening, surveillance, or diagnostic). Statistical analyses were performed using SAS, version 9.1.3. This study was reviewed and approved by the Virginia Commonwealth University Institutional Review Board.
December 2007
54.6 20.1 12.3 37.8
86.2 5.4 3.3 0.9 4.3 67.1 9.5 23.4
Results Of the 3000 charts reviewed at the 10 practices, 1282 (42.7%) contained documentation of a prior colonoscopy and 1021 (34.0% of all abstracted charts, 79.6% of those in which a colonoscopy occurred) included correspondence from the endoscopist. A total of 15 patients from the sampling frame met exclusion criteria due to a prior history or new diagnosis of CRC (n⫽5), high-grade dysplasia or villous polyps (n⫽7), or inflammatory bowel disease (n⫽3). The characteristics of the remaining 1006 patients who constituted the study sample are presented in Table 2. While 55% of patients had some form of prior CRC screening, only 32% had a prior endoscopy, significantly lower than the national average of 53% in 2004.30 The indication for the most recent colonoscopy was predominantly screening (67.1%). A family history of CRC or polyps was recorded in 120 records. The colonoscopies were performed by 130 unique endoscopists, the majority of whom (86.2%) were gastroenterologists (Table 2). The colonoscopies were performed between 1992 and 2006, with a median procedure year of 2004. The proportion of colonoscopies performed before February 1997, in March 1997–February 2003, in March 2003–June 2006, and in July 2006 were 1.0%, 24.7%, 63.6%, and 10.7%, respectively. The colonoscopy results were normal in 54.8% of cases; 10.3% of patients had hyperplastic polyps, 7.6% had 1–2 small adenomas, and 3.3% had 3–10 adenomas, a large adenoma, or high-grade dysplasia. In 12.1% of cases, correspondence from the endoscopist indicated that the colonoscopy detected an unspecified type of polyp and no pathology information was found in the Am J Prev Med 2007;33(6)
473
Table 3. Frequency of endoscopists’ follow-up recommendations based on colonoscopy findings Colonoscopy findings with report (Nⴝ1006)a
Repeat colonoscopy recommendations (Nⴝ663 with report and colonoscopy recommendation)b
Normal
n⫽551 (54.8%)
Hyperplastic polyp(s)
n⫽104 (10.3%)
1–2 small (⬍1 cm) tubular adenoma
n⫽76 (7.6%)
3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia
n⫽40 (3.9%)
Family history of colorectal cancer or of polypsd
n⫽113 (11.2%)
Unknown polyp type
n⫽122 (12.1%)
Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽351) — 7.1 yrs/7.5 yrs [7.1– 8.6 years]c Repeat in ⬍5 years — 7.4% Repeat in 5–9 years — 56.4% Repeat in 10 years — 36.2% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽65) — 4.9 yrs/5.8 yrs [4.9 – 6.7 years]c Repeat in ⬍5 years — 43.1% Repeat in 5–9 years — 46.1% Repeat in 10 years — 10.8% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽47) — 3.6 yrs/4.4 yrs [3.5–5.3 years]c Repeat in ⬍3 years — 6.4% Repeat in 3–⬍5 years — 61.7% Repeat in 5 years — 31.9% Repeat in 6 –10 years — 0% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽30) — 2.9 yrs/3.5 yrs [2.5– 4.5 years]c Repeat in ⬍3 years — 30.0% Repeat in 3 years — 50.0% Repeat in 4 –5 years — 20.0% Repeat in ⬎5 years — 0% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽90) — 4.8 yrs/5.8 yrs [5.0 – 6.6 years]c Repeat in ⬍3 years — 2.2% Repeat in 3– 4 years — 33.3% Repeat in 5 years — 55.6% Repeat in 6 –9 years — 1.1% Repeat in 10 years — 7.8% Mean years to repeat colonoscopy (n⫽80) — 4.0 yrs Repeat in ⬍3 years — 10.0% Repeat in 3 years — 36.3% Repeat in 4 years — 11.3% Repeat in 5 years — 33.8% Repeat in 6 –9 years — 3.8% Repeat in 10 years — 2.5%
colonoscopy
colonoscopy
colonoscopy
colonoscopy
colonoscopy
a
Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. Bold font designates the follow-up interval recommended in the 2006 guidelines.25 (see Table 1). Adjusted for random effect of providers nested within practice as well as patient age and gender, primary care practice, and the endoscopist’s specialty. d Patients with a family history as an indication for performing the test and a result of normal colonoscopy, hyperplastic polyp(s), 1–2 adenomatous polyps or unknown polyp type were included only in this row. b c
chart. Multiple findings were reported in 2.9% of charts. Although the frequency of hyperplastic polyps in the study sample was consistent with prior literature, the prevalence of precancerous adenomas reported above and cancers was below average.12,31 Correspondence from the endoscopist included no guidance on follow-up testing in 33.5% of cases. Of the remaining 743 patients in whom retesting was discussed, colonoscopy was predominantly recommended (66.0% of patients). Other recommendations included sigmoidoscopy (0.4%), fecal occult blood testing alone (0.5%), and fecal occult blood testing in combination with colonoscopy (6.9%). Endoscopists generally advocated a specific number of years in which retesting should occur; in only 9.6% of cases was a range of years stated. The advice given for 474
the 663 patients for whom repeat colonoscopy was recommended is presented in detail in Table 3. The retesting interval advocated by the endoscopists was consistent with contemporaneous guidelines in only 39.2% of cases and in only 36.7% of cases when all findings, irrespective of the year of the colonoscopy, were compared against current (2006) guidelines. Adherence to guidelines was greatest for patients at increased risk (family history or multiple adenomas) and was lowest when hyperplastic polyps were detected (Table 4). After adjustment, the mean recommended years for retesting, regardless of indication, was 7.8 years for a normal colonoscopy, 5.8 years for hyperplastic polyps, 4.4 years for 1–2 small adenomas, 3.5 years for 3–10 adenomas, and 5.8 years for a family history of CRC (Table 3).
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
Table 4. Percentage of patients with repeat colonoscopy intervals consistent with contemporaneous guidelinesa Referent guidelineb
Findingc Normal Hyperplastic polyps 1–2 small (⬍1 cm) tubular adenoma 3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia Family history of colorectal cancere Total consistent with guideline
1997 2003 2006 guideline23 guideline24 guideline25 (nⴝ113) (nⴝ328) (nⴝ74) 40.7 N.R. N.R.
34.7 N.R. 25.0
33.3 27.3 50.0d
37.5
53.3
57.1
54.2
55.4
62.5
43.4
38.1
37.8
a
Adjusted for random effect of providers nested within practice as well as patient age and gender, primary care practice, and the endoscopist’s specialty. b Referent guidelines: the 1997, 2003, and 2006 guidelines were referent, respectively, for colonoscopies performed in 1997–2003 (n⫽113), 2003–2006 (n⫽328), and 2006 (n⫽74). c Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. d Assumes an acceptable retesting interval could include 5, 5–10, or 10 years, as recommended in current (2006) guidelines (see Table 1). e Patients with a family history as an indication for performing the test and a result of normal colonoscopy, hyperplastic polyp(s), 1–2 adenomatous polyps, or unknown polyp type were included only in this row. N.R., no recommendation appears in the guideline.
Discussion Based on the information provided in this sample of correspondence with PCCs, the endoscopists often recommended follow-up colonoscopy at shorter intervals than those advocated at the time of the procedure, or currently, by gastroenterologic societies. Guidelines issued in 1997, 2003, and 2006 have stated consistently that colonoscopy should be repeated 10 years after a normal examination and the 2003 and 2006 guidelines recommended a 5-year follow-up interval,4,23–25 yet the adjusted mean follow-up interval recommended by endoscopists was significantly shorter in both cases. This study adds to an extensive literature on overuse of healthcare services, a phenomenon in the U.S. that fuels higher healthcare costs, consumes resources that can be put to more effective use, and exposes patients to unnecessary risks.32–34 The inference that this study documents overuse of colonoscopy requires the assumptions that the shorter retesting intervals advocated by endoscopists were (1) inappropriate and (2) implemented. Neither of these assumptions can be confirmed by this study, which was designed to examine the advice that endoscopists communicated to PCCs rather than the testing that actually occurred. Clinical considerations known to the endoscopist but not reflected in the records of the PCC could explain the appropriateness of recommended retesting intervals that appear to December 2007
depart from those advocated in practice guidelines. For example, endoscopists may favor a shorter follow-up interval if they question the quality of their examinations (due to incomplete bowel preparation or inadequate luminal views) or are cognizant of relevant symptoms or pathology detected on prior colonoscopies that heighten the patient’s CRC risk. While these undisclosed details may justify the recommendations of endoscopists, the failure to cite them explicitly as a rationale to accompany recommendations compromises the ability of PCCs to vet the appropriateness of the advice. In the correspondence with PCCs examined in this study, there were few notations about the adequacy of the bowel preparation or luminal visualization. Beyond a statement of polyp findings or the rare report of not completing the examination, little justification for repeat-testing intervals was offered. Apart from the appropriateness of the advice given by the endoscopists, the question of whether the frequent retesting actually occurred cannot be confirmed by this study, which neither audited the records or insurance claims of endoscopists or PCCs nor surveyed patients to document their experience. In many instances, however, the advice given in the correspondence analyzed in this study is what will dictate the timing of future follow-up colonoscopy. Whether or not they agree with the recommended follow-up interval or can find in the communication a convincing rationale for the advice, PCCs face medicolegal consequences if they depart from the recommendations documented in the medical record. Informed patients who recall the advice given by the endoscopist also may hold expectations about the timing of follow-up. What is communicated to the PCC and what appears in the medical record of the patient—the data substrate for this study—therefore can supersede the above clinical caveats in determining whether colonoscopy is overused. A longstanding body of work documents the difficulties faced by physicians in adhering to clinical practice guidelines and the various factors that account for this tendency.35– 40 Similar factors may account for the disparity between colonoscopy guidelines and the advice endoscopists appear to be giving PCCs. For example, in many cases, the tendency of endoscopists to advocate retesting at short intervals may relate less to legitimate clinical considerations than to an entrenched misunderstanding about current recommendations or the data on which they are based. Many endoscopists may dispute that evidence, disagree with the guidelines, or favor more aggressive approaches after having experienced adverse outcomes with previous patients in whom CRC escaped detection. Endoscopists who maintain concerns about medicolegal liability41 also may prefer shorter follow-up intervals. Financial incentives also may motivate some endoscopists to endorse shorter retesting intervals. Am J Prev Med 2007;33(6)
475
The literature documents that physicians require time to adapt to new guidelines. Colonoscopy guidelines have changed since 1997 (Table 1), but many of today’s endoscopists may have formed their views about retesting intervals in the era before 1997, when frequent retesting was advocated. Since 1997, guidelines have consistently recommended a 3-year follow-up interval for patients with 3–10 adenomas (or any adenoma ⱖ1 cm or with high-grade dysplasia), and the temporal trend in Table 4 suggests that adherence to that guideline has gradually improved over time, consistent with prevailing theories about the diffusion of new knowledge. Table 4 displays less encouraging trends, however, for adherence to policy for normal examinations. While it might come as little surprise that recommendations made by endoscopists years ago do not adhere to current (2006) guidelines, the data in Table 4 finds the same to be true for adherence to contemporaneous guidelines as well. Whatever the reasons, the overuse of colonoscopy carries profound public health implications. As the guideline panels have documented, there is little evidence that patients benefit by undergoing colonoscopies at shorter intervals than the periods recommended in the guidelines. There is measurable harm from this practice, however, because of the known complication rates associated with colonoscopy. Formal methods are necessary to make precise economic projections, but the impact can be approximated if one assumes: (1) that the rate for colonoscopy complications is: 0.5 death, nine bowel perforations, 20 major hemorrhages, and 50 serious complications per 10,000 colonoscopies12–18; (2) that 14.2 million colonoscopies are performed annually in the United States41,42; (3) that the prevalence of pathologic findings on colonoscopies nationwide resembles the distribution that was observed; (4) that the retesting intervals advocated in the 2006 guidelines are ideal; and (5) that a colonoscopy costs $1200.43 These conditions would result in the performance of 2.85 million excess colonoscopies over the next 10 years. Given the above assumptions, this cascade of procedures would induce approximately 142 deaths, 2561 perforations, 5692 major hemorrhages, and 14,229 serious complications. The excess colonoscopies would generate $3.4 billion in direct costs, consuming limited healthcare resources that could be more effectively applied to closing the gap in needed CRC screening.44 The overuse of colonoscopy reduces the overall cost-effectiveness ratio for the procedure and places greater burden on healthcare personnel to meet heightened volume demands. This study documented poor communication between endoscopists and PCCs. Of the charts with a documented colonoscopy, 20% lacked an endoscopist’s report; the results of the colonoscopy could not be determined from the PCC chart. In 10% of cases, correspondence from the endoscopist made reference 476
to one or more polyps detected on colonoscopy, but the chart contained no biopsy results. Frequently, the chart contained only a procedure note generated immediately after the colonoscopy with the notation that a polyp was found and biopsied. In such cases, recommendations on retesting were made preliminarily by the endoscopist based on the gross appearance of the polyp but prior to receiving pathologic details. This study has several important limitations. First, as noted above, the analysis of endoscopists’ recommendations through the lens of the PCC does not provide an objective assessment of the knowledge, beliefs, records, or actions of endoscopists or of the advice that they provide directly to patients. Second, the covariables listed above that might justify a shorter screening interval were not assessed. Third, charts may be inaccurate with respect to family history, both neglecting to report a significant family history and attributing a “family history” when the family member was distantly related (i.e., not a first-degree relative) or did not have CRC. Fourth, chart audits have known limitations45– 47 and this study lacked resources for dual review and arbitration of coding disagreements. Fifth, as noted earlier, endoscopists need time to adapt to new guidelines, and this analysis held them accountable for advice given in the same year the guidelines were released. The findings of this study underscore the need for system changes to bring the practice of endoscopists and PCCs into closer alignment with published guidelines. In particular, seven system changes seem warranted: (1) Systems are needed to ensure that both endoscopists and PCCs are knowledgeable about guidelines and, in particular, are made aware of changes as new guidelines are issued. Their respective specialty societies can play an important role in facilitating diffusion; (2) Physicians should consider whether other strategies might be more effective than increasing the frequency of colonoscopy in preventing CRC. For example, emphasizing changes in technique to ensure that colonoscopies are performed well may be more effective and safer than repeating the test more frequently48; (3) Practices and healthcare systems should adopt tools to facilitate adherence to the guidelines (e.g., prompts or standing orders that note recommended follow-up intervals alongside check boxes for colonoscopy findings). These reminder systems must be updated as new guidelines are issued; (4) Practices and healthcare systems should enhance communication between the endoscopist and the PCC, in particular to ensure that both the procedure note and the pathology results, which may reach the endoscopist later, are always forwarded to the PCC;
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
(5) PCCs need systems to flag missing information in the records, such as failure to document whether a screening or surveillance examination occurred, the specific examination findings, and copies of pathology reports. In particular, the failure to collect a complete and accurate family history influences the degree to which PCCs can ensure adequate retesting for those with genetic predispositions for CRC; (6) PCCs should not rely solely on the endoscopist for guidance on follow-up nor should they assume that recommended retesting intervals are consistent with either current guidelines or the patient’s pathologic findings. Physicians who will be engaged in the ongoing care of patients, usually PCCs, are responsible for ensuring that the advice they give patients about follow-up is appropriate for their test results and consistent with recent evidence-based guidelines. As needed, PCCs should amend the recommendations of endoscopists and document the rationale for the change; and (7) Patients should be provided current consumeroriented guidelines and be informed of the specific findings on their examinations, thereby empowering them as change agents to ensure that they are aware of the appropriate follow-up interval for their next test and can advocate on their own behalf. The findings reported here reflect conditions at 10 practices in Virginia and Maryland. Studies currently in progress elsewhere in the United States will expand the evidence base and help confirm whether the phenomenon that was observed is widespread in primary care. Further research will also help to identify and evaluate strategies to change systems of care to help the population avert the public health consequences of overuse of colonoscopy. We would like to thank the Virginia Commonwealth University Department of Family Medicine for supporting this project, the practices that facilitated data collection (Fairfax Family Practice, Front Royal Family Practice, Chesterfield Family Medicine, St. Francis Family Medicine, Vienna Family Medicine, Potomac Physician Associates, Little Falls Practice, Premier Primary Care Physicians, Franconia Family Medicine), and the faculty, residents, and medical students who collected the data (Maria Gaspar; Britta Johnson, MD; Michael Johnson; Katherine Jones; Christine Kerns, RN; Haewon Park; Kelley Prince, MD; Walter Waugh, MD; and Ericka Young, MD). This project could not have been accomplished without their generous assistance. Dr. Krist and Dr. Kerns are faculty members, practicing physicians, and partial owners of Fairfax Family Practice and Front Royal Family Practice, respectively. Otherwise, no authors have any conflicts of interest to report. No financial disclosures were reported by the authors of this paper.
December 2007
References 1. Pignone M, Rich M, Teutsch SM, Berg AO, Lohr KN. Screening for colorectal cancer in adults at average risk: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:132– 41. 2. Colorectal cancer screening. Recommendation statement from the Canadian Task Force on Preventive Health Care. CMAJ 2006;165:206 – 8. 3. Pignone M, Saha S, Hoerger T, Mandelblatt J. Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:96 –104. 4. Screening for colorectal cancer: recommendation and rationale. Ann Intern Med 2002;137:129 –31. 5. Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA 2003;289:1288 –96. 6. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977– 81. 7. Citarda F, Tomaselli G, Capocaccia R, Barcherini S, Crespi M. Efficacy in standard clinical practice of colonoscopic polypectomy in reducing colorectal cancer incidence. Gut 2001;48:753– 4. 8. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 1992;326:658 – 62. 9. Thiis-Evensen E, Hoff GS, Sauar J, Langmark F, Majak BM, Vatn MH. Population-based surveillance by colonoscopy: effect on the incidence of colorectal cancer. Telemark Polyp Study I. Scand J Gastroenterol 1999; 34:414 –20. 10. Newcomb PA, Storer BE, Morimoto LM, Templeton A, Potter JD. Longterm efficacy of sigmoidoscopy in the reduction of colorectal cancer incidence. J Natl Cancer Inst 2003;95:622–5. 11. Bertario L, Russo A, Sala P, et al. Predictors of metachronous colorectal neoplasms in sporadic adenoma patients. Int J Cancer 2003;105:82–7. 12. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000; 343:162– 8. 13. Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169 –74. 14. Yousfi M, Gostout CJ, Baron TH, et al. Postpolypectomy lower gastrointestinal bleeding: potential role of aspirin. Am J Gastroenterol 2004;99: 1785–9. 15. Cobb WS, Heniford BT, Sigmon LB, et al. Colonoscopic perforations: incidence, management, and outcomes. Am Surg 2004;70:750 –7; discussion 757– 8. 16. Misra T, Lalor E, Fedorak RN. Endoscopic perforation rates at a Canadian university teaching hospital. Can J Gastroenterol 2004;18:221– 6. 17. Nelson DB, McQuaid KR, Bond JH, Lieberman DA, Weiss DG, Johnston TK. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55:307–14. 18. Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an integrated health care delivery system. Ann Intern Med 2006;145:880 – 6. 19. Ransohoff DF. Colon cancer screening in 2005: status and challenges. Gastroenterology 2005;128:1685–95. 20. Sirovich BE, Schwartz LM, Woloshin S. Screening men for prostate and colorectal cancer in the United States: does practice reflect the evidence? JAMA 2003;289:1414 –20. 21. Vijan S, Inadomi J, Hayward RA, Hofer TP, Fendrick AM. Projections of demand and capacity for colonoscopy related to increasing rates of colorectal cancer screening in the United States. Aliment Pharmacol Ther 2004;20:507–15. 22. Ladabaum U, Song K. Projected national impact of colorectal cancer screening on clinical and economic outcomes and health services demand. Gastroenterology 2005;129:1151– 62. 23. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594 – 642. 24. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale–Update based on new evidence. Gastroenterology 2003;124:544 – 60. 25. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the U.S. MultiSociety Task Force on Colorectal Cancer and the American Cancer Society. CA Cancer J Clin 2006;56:143–59; quiz 184 –5. 26. Mysliwiec PA, Brown ML, Klabunde CN, Ransohoff DF. Are physicians doing too much colonoscopy? A national survey of colorectal surveillance after polypectomy. Ann Intern Med 2004;141:264 –71.
Am J Prev Med 2007;33(6)
477
27. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q 2005;83:457–502. 28. Tyler CV Jr, Snyder CW. Cancer risk assessment: examining the family physician’s role. J Am Board Fam Med 2006;19:468 –77. 29. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and multivariable methods. 3rd edn. Boston: Duxbury Press, 1998. 30. Behavioral Risk Factor Surveillance System: Prevalence Data: National Center for Chronic Disease Prevention and Health Promotion, 2007. Available online at: http://apps.nccd.cdc.gov/brfss/index.asp. 31. Schoenfeld P, Cash B, Flood A, et al. Colonoscopic screening of averagerisk women for colorectal neoplasia. N Engl J Med 2005;352:2061– 8. 32. Woolf SH. Potential health and economic consequences of misplaced priorities. JAMA 2007;297:523– 6. 33. Schwartz LM, Woloshin S, Fowler FJ Jr, Welch HG. Enthusiasm for cancer screening in the United States. JAMA 2004;291:71– 8. 34. Benefits foregone: too much of the wrong and too little of the right. Based on a presentation by Bernard S. Bloom PhD., Am J Manag Care 1999;5(6 Suppl):S355– 60. 35. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med 2003;348:2635– 45. 36. Woolf SH, Krist A. The liability of giving patients a choice: shared decision making and prostate cancer. Am Fam Physician 2005;71:1871–2. 37. Ardery G, Carter BL, Milchak JL, et al. Explicit and implicit evaluation of physician adherence to hypertension guidelines. J Clin Hypertens (Greenwich) 2007;9:113–9. 38. Flanders SA, Halm EA. Guidelines for community-acquired pneumonia: are they reflected in practice? Treat Respir Med 2004;3:67–77.
478
39. Schers H, Wensing M, Huijsmans Z, van Tulder M, Grol R. Implementation barriers for general practice guidelines on low back pain a qualitative study. Spine 2001;26:E348 –53. 40. McAlister FA, Campbell NR, Zarnke K, Levine M, Graham ID. The management of hypertension in Canada: a review of current guidelines, their shortcomings and implications for the future. CMAJ 2001;164:517–22. 41. Seeff LC, Manninen DL, Dong FB, et al. Is there endoscopic capacity to provide colorectal cancer screening to the unscreened population in the United States? Gastroenterology 2004;127:1661–9. 42. 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. 43. Colorectal cancer: health professionals facts on screening. Washington DC: Centers for Disease Control and Prevention, 2000. Available online at: http://www.cdc.gov/cancer/colorectal/pdf/fs-professional.pdf. 44. Butterly L, Olenec C, Goodrich M, Carney P, Dietrich A. Colonoscopy demand and capacity in New Hampshire. Am J Prev Med 2007;32:25–31. 45. Murff HJ, Patel VL, Hripcsak G, Bates DW. Detecting adverse events for patient safety research: a review of current methodologies. J Biomed Inform 2003;36:131– 43. 46. Worster A, Haines T. Advanced statistics: understanding medical record review (MRR) studies. Acad Emerg Med 2004;11:187–92. 47. Camacho LA, Rubin HR. Assessment of the validity and reliability of three systems of medical record screening for quality of care assessment. Med Care 1998;36:748 –51. 48. Barclay RL, Vicari JJ, Doughty AS, Johanson JF, Greenlaw RL. Colonoscopic withdrawal times and adenoma detection during screening colonoscopy. N Engl J Med 2006;355:2533– 41.
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
Analysts abstracted medical records at ten primary care practices in Virginia and Maryland in 2006. The records of a random sample of men and women (300 per practice) aged 50 to 70 years were reviewed. The sample included patients who had a colonoscopy and a written report from an endoscopist, and who lacked designated risk factors. The main outcome was concordance between endoscopists’ recommendations and published guidelines regarding repeat colonoscopy.
Results:
Of 3000 charts reviewed, 1282 (42.7%) included records of a colonoscopy and 1021 (34%) included an endoscopist’s report. In 64.9% of communications, the endoscopist specified when retesting should occur. Recommendations were consistent with contemporaneous guidelines in only 39.2% of cases and with current guidelines in 36.7% of cases. The adjusted mean number of years in which repeat colonoscopy was recommended was 7.8 years following normal colonoscopy and 5.8 years and 4.4 years, respectively, when hyperplastic polyps or 1–2 small adenomatous polyps were found.
Conclusions: Endoscopists often recommended repeat colonoscopy at shorter intervals than are advised either by current guidelines or by guidelines in effect at the time of the procedure. Endoscopists’ communications to primary care clinicians often lacked contextual information that might explain these discrepancies. Unless appropriate caveats apply, adhering to endoscopists’ recommendations may incur unnecessary harms and costs. (Am J Prev Med 2007;33(6):471– 478) © 2007 American Journal of Preventive Medicine
Introduction
O
f the available screening and surveillance tests for colorectal cancer (CRC), colonoscopy offers the greatest single-test probability for detecting either CRC or polyps.1–5 Colonoscopy also facilitates excision of precancerous lesions, thereby potentially preventing their progression to cancer.6 –11 From the Departments of Family Medicine (Krist, Woolf, Kerns), Epidemiology and Community Health (Jones, Woolf), Virginia Commonwealth University, Richmond; Fairfax Family Practice Residency (Krist, Woessner), Fairfax; Shenandoah Valley Family Practice Residency Program (Kerns), Front Royal; Chesterfield Family Medicine Residency (Foliaco), St. Francis Family Medicine Residency (Jackson), Massey Cancer Center (Jones), Richmond, Virginia; and Department of Family Medicine (Merenstein), Georgetown University, Washington DC Address correspondence and reprint requests to: Alex H. Krist, MD, MPH, 3825 Charles Stewart Drive, Fairfax VA 22033. E-mail: ahkrist@ vcu.edu. The full text of this article is available via AJPM Online at www. ajpm-online.net; 1 unit of Category-1 CME credit is also available, with details on the website.
However, the benefits of colonoscopy are accompanied by risks for bleeding, perforation, and death secondary to the procedure.12–18 Performing colonoscopy too often not only increases patients’ exposure to procedural harm,19 but also drains limited resources that could be more effectively used to bolster the nation’s limited capacity to adequately screen those in need.20 –22 Guidelines promulgated in the early 1990s advocated short follow-up intervals after an abnormal colonoscopy. That changed in 1997, when the profession adopted new consensus guidelines, issued by the United States Multi-Society Task Force on Colorectal Cancer and the American Cancer Society.23 The Task Force issued guideline updates in 200324 and 200625 (see Table 1), and with each iteration the guidelines proposed longer follow-up intervals for some abnormal findings and clarified risk stratification categories. Despite the evidence and consensus of experts articulated in these guidelines, a survey of endoscopists in 2004 found that
Am J Prev Med 2007;33(6) © 2007 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/07/$–see front matter doi:10.1016/j.amepre.2007.07.039
471
Table 1. Recommended intervals for repeat colonoscopy based on colonoscopic findings Finding
1997 recommendations
2003 recommendations24
2006 recommendations25
Normal Hyperplastic polyps 1–2 small (⬍1 cm) tubular adenoma 3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia ⬎10 adenomas on one examination Adenomas that are removed piecemeal Family history of colon cancerd
10 years N.R. N.R.a 3 yearsa
10 years N.R. 5 yearsb 3 years
10 years 10 years 5–10 years 3 years
N.R.c N.R. 5 years
N.R.c N.R. 5 years
⬍3 years 2–6 months 5 years
a
The guideline directly addresses patients with large (⬎1 cm) and “multiple adenomatous polyps.” However, no specific recommendation is made for 1–2 small adenomas. b Recommendation states, “It is not unreasonable to choose even longer intervals.” c Recommendation made only for patients with “multiple adenomatous polyps” without a separate recommendation for ⬎10 adenomatous polyps. d In 2003 and 2006, family history is defined as, “a first-degree relative diagnosed prior to the age of 60 years or two first-degree relatives diagnosed at any age.” In 1997, family history is defined as, “a close relative (sibling parent, or child) who has had colorectal cancer or an adenomatous polyp.” N.R., no recommendation appears in the guideline.
more than 50% of respondents were repeating colonoscopies at shorter intervals than recommended in the 1997 and 2003 guidelines.26 Frequently, the endoscopist’s follow-up recommendations are communicated to the patient’s primary care clinician (PCC) in a note or letter sent after the procedure. The correspondence may be sent before the endoscopist has received all of the final pathology reports on biopsy specimens. Some endoscopists do not update the PCC when the final results become available. Since the PCC is more likely than the endoscopist to have an ongoing relationship with the patient,27 the PCC is often ultimately responsible for ensuring the delivery of appropriate follow-up care years later, when repeat testing is due.28 Not having personally examined the bowel, the PCC must rely on the information and recommendations received from the endoscopist. This communication therefore becomes crucial for ensuring high-quality, long-term care. The objective of this study was to evaluate the extent to which endoscopists’ recommendations on repeat colonoscopy, as conveyed in communications to PCCs, were concordant with contemporaneous and current (2006) guidelines on screening and postpolypectomy surveillance.
Methods Abstractors reviewed the medical records of 3000 randomly selected patients at ten primary care practices. The purpose was to observe the endoscopists’ recommendations through the lens of the PCC by reviewing all communications and information that PCCs received from the endoscopist. The study was not designed to evaluate the information contained in endoscopists’ records, nor was it intended to document when retesting actually occurred.
Setting Five primary care practices from the Virginia Ambulatory Care Outcomes Research Network (ACORN) and five prac-
472
tices from the Capital Area Primary Care Research Network (CAPRICORN) participated in the study. Four of the practices maintained family medicine residency programs, while the others were nonteaching practices. The practices cared for patients in a large geographic area: five in northern VA, one in Front Royal VA, two in Richmond VA, and two in southern MD. No PCCs at the study sites performed colonoscopies themselves. All practices referred patients to endoscopists, the majority of whom were in private practice. The endoscopists were selected based on the preferences of patients and PCCs, convenience to the patient, and insurance restrictions.
Study Sample To derive the sample, the 10 practices generated a list of all patients aged 50 to 70 years who made an office visit between February 2005 and August 2006. Of these, each practice selected 300 patients for chart review using a random sample generator program in STATA. From this sampling frame of 3000 patients, subjects were included in the final study sample if they had ever had a documented colonoscopy (irrespective of date performed) and a communication (e.g., procedure note, office visit note, or letter) from the endoscopist to the PCC reporting the findings. High-risk patients were excluded from the final sample because determining the appropriate interval for repeat colonoscopy required access to historic data that were not collected. Specifically, patients with a personal history of CRC, inflammatory bowel disease, large adenoma, ⬎10 adenomas, villous adenoma, or a prior adenoma with high-grade dysplasia were excluded.
Data Collection Each chart from the sample was examined by one of nine reviewers, who used a standardized abstraction form to record the study data. Reviewers were instructed by the investigators and by each practice on where in a patient’s record to find information about CRC screening. Resources did not permit dual review of charts. All relevant information contained in the PCC record—including correspondence from the endoscopist, pathologist, or hospital—was examined.
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
Outcomes
Table 2. Characteristics of study population (N⫽1006)
Data abstracted from the charts included the patient’s age (at the time of the procedure) and gender, documentation of a prior colonoscopy, any record of family history of CRC in the patient’s chart, presence of a communication from the endoscopist about the procedure, and pathology reports for patients who had a biopsy. When the information was available in the chart or in the endoscopist’s correspondence, reviewers recorded the indication for colonoscopy, the colonoscopy findings, the recommended follow-up tests and testing interval advised by the endoscopist, and (when stated) mitigating factors (e.g., inadequate bowel preparation) that might have influenced the recommendation. If the record included multiple recommendations related to previous colonoscopies, the most recent retesting recommendation from the endoscopist was used.
Characteristic
Findings
Male (%) Mean age (standard error) Proportion of sample who had received prior colorectal cancer screening or diagnostic testing (%) Any test Colonoscopy Sigmoidoscopy Fetal occult blood test Type of provider performing colonoscopy (Nⴝ130 unique endoscopists) (%) Gastroenterologist General surgeon Colorectal surgeon Primary care physician/internist Unknown Indication for performing colonoscopy (%) Screening Surveillance Diagnostic
49.8 57.9 years (0.189)
Analysis Colonoscopy findings were classified as: normal examination; hyperplastic polyp(s); 1–2 small adenoma(s); 3–10 adenomas, large adenoma (ⱖ1 cm in size), or high-grade dysplasia; ⬎10 adenomas; CRC; or unknown polyp type. Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. Similarly, a family history of CRC requires a shorter retesting interval than is indicated after a normal examination or after the detection of hyperplastic polyps or 1–2 adenomas. Therefore, patients with the combination of these findings or unknown polyp type and a family history were evaluated separately from those with the same findings but no family history. Endoscopist recommendations on repeat colonoscopy were compared to both contemporaneous guidelines and to current (2006) guidelines. Colonoscopies performed prior to 1997 were excluded from the contemporaneous comparison, because a consensus guideline from the gastroenterology organizations was not available at that time, but they were included in the comparison with current guidelines. Findings for which contemporaneous guidelines made no explicit recommendations were excluded from the analysis. For example, because the 2003 guideline made no explicit recommendations about follow-up after detecting hyperplastic polyps, the contemporaneous comparison would exclude advice given in 2004 following a colonoscopy that detected such polyps. For all colonoscopies performed across these time periods, the mean length of time (years) in which endoscopists recommended reexamination was calculated for each category of colonoscopy findings. Time to recommended follow-up did not demonstrate departure from normal so linear regression was utilized to determine the adjusted mean years to follow-up by colonoscopy findings.29 Physicians were nested within each practice as a random effect and the means were adjusted for four fixed effects: patient age and gender, primary care practice, and the endoscopist’s specialty. The adjusted means were also stratified by the indication for performing the test (screening, surveillance, or diagnostic). Statistical analyses were performed using SAS, version 9.1.3. This study was reviewed and approved by the Virginia Commonwealth University Institutional Review Board.
December 2007
54.6 20.1 12.3 37.8
86.2 5.4 3.3 0.9 4.3 67.1 9.5 23.4
Results Of the 3000 charts reviewed at the 10 practices, 1282 (42.7%) contained documentation of a prior colonoscopy and 1021 (34.0% of all abstracted charts, 79.6% of those in which a colonoscopy occurred) included correspondence from the endoscopist. A total of 15 patients from the sampling frame met exclusion criteria due to a prior history or new diagnosis of CRC (n⫽5), high-grade dysplasia or villous polyps (n⫽7), or inflammatory bowel disease (n⫽3). The characteristics of the remaining 1006 patients who constituted the study sample are presented in Table 2. While 55% of patients had some form of prior CRC screening, only 32% had a prior endoscopy, significantly lower than the national average of 53% in 2004.30 The indication for the most recent colonoscopy was predominantly screening (67.1%). A family history of CRC or polyps was recorded in 120 records. The colonoscopies were performed by 130 unique endoscopists, the majority of whom (86.2%) were gastroenterologists (Table 2). The colonoscopies were performed between 1992 and 2006, with a median procedure year of 2004. The proportion of colonoscopies performed before February 1997, in March 1997–February 2003, in March 2003–June 2006, and in July 2006 were 1.0%, 24.7%, 63.6%, and 10.7%, respectively. The colonoscopy results were normal in 54.8% of cases; 10.3% of patients had hyperplastic polyps, 7.6% had 1–2 small adenomas, and 3.3% had 3–10 adenomas, a large adenoma, or high-grade dysplasia. In 12.1% of cases, correspondence from the endoscopist indicated that the colonoscopy detected an unspecified type of polyp and no pathology information was found in the Am J Prev Med 2007;33(6)
473
Table 3. Frequency of endoscopists’ follow-up recommendations based on colonoscopy findings Colonoscopy findings with report (Nⴝ1006)a
Repeat colonoscopy recommendations (Nⴝ663 with report and colonoscopy recommendation)b
Normal
n⫽551 (54.8%)
Hyperplastic polyp(s)
n⫽104 (10.3%)
1–2 small (⬍1 cm) tubular adenoma
n⫽76 (7.6%)
3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia
n⫽40 (3.9%)
Family history of colorectal cancer or of polypsd
n⫽113 (11.2%)
Unknown polyp type
n⫽122 (12.1%)
Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽351) — 7.1 yrs/7.5 yrs [7.1– 8.6 years]c Repeat in ⬍5 years — 7.4% Repeat in 5–9 years — 56.4% Repeat in 10 years — 36.2% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽65) — 4.9 yrs/5.8 yrs [4.9 – 6.7 years]c Repeat in ⬍5 years — 43.1% Repeat in 5–9 years — 46.1% Repeat in 10 years — 10.8% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽47) — 3.6 yrs/4.4 yrs [3.5–5.3 years]c Repeat in ⬍3 years — 6.4% Repeat in 3–⬍5 years — 61.7% Repeat in 5 years — 31.9% Repeat in 6 –10 years — 0% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽30) — 2.9 yrs/3.5 yrs [2.5– 4.5 years]c Repeat in ⬍3 years — 30.0% Repeat in 3 years — 50.0% Repeat in 4 –5 years — 20.0% Repeat in ⬎5 years — 0% Mean years (unadjusted/adjusted [95% CI]) to repeat (n⫽90) — 4.8 yrs/5.8 yrs [5.0 – 6.6 years]c Repeat in ⬍3 years — 2.2% Repeat in 3– 4 years — 33.3% Repeat in 5 years — 55.6% Repeat in 6 –9 years — 1.1% Repeat in 10 years — 7.8% Mean years to repeat colonoscopy (n⫽80) — 4.0 yrs Repeat in ⬍3 years — 10.0% Repeat in 3 years — 36.3% Repeat in 4 years — 11.3% Repeat in 5 years — 33.8% Repeat in 6 –9 years — 3.8% Repeat in 10 years — 2.5%
colonoscopy
colonoscopy
colonoscopy
colonoscopy
colonoscopy
a
Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. Bold font designates the follow-up interval recommended in the 2006 guidelines.25 (see Table 1). Adjusted for random effect of providers nested within practice as well as patient age and gender, primary care practice, and the endoscopist’s specialty. d Patients with a family history as an indication for performing the test and a result of normal colonoscopy, hyperplastic polyp(s), 1–2 adenomatous polyps or unknown polyp type were included only in this row. b c
chart. Multiple findings were reported in 2.9% of charts. Although the frequency of hyperplastic polyps in the study sample was consistent with prior literature, the prevalence of precancerous adenomas reported above and cancers was below average.12,31 Correspondence from the endoscopist included no guidance on follow-up testing in 33.5% of cases. Of the remaining 743 patients in whom retesting was discussed, colonoscopy was predominantly recommended (66.0% of patients). Other recommendations included sigmoidoscopy (0.4%), fecal occult blood testing alone (0.5%), and fecal occult blood testing in combination with colonoscopy (6.9%). Endoscopists generally advocated a specific number of years in which retesting should occur; in only 9.6% of cases was a range of years stated. The advice given for 474
the 663 patients for whom repeat colonoscopy was recommended is presented in detail in Table 3. The retesting interval advocated by the endoscopists was consistent with contemporaneous guidelines in only 39.2% of cases and in only 36.7% of cases when all findings, irrespective of the year of the colonoscopy, were compared against current (2006) guidelines. Adherence to guidelines was greatest for patients at increased risk (family history or multiple adenomas) and was lowest when hyperplastic polyps were detected (Table 4). After adjustment, the mean recommended years for retesting, regardless of indication, was 7.8 years for a normal colonoscopy, 5.8 years for hyperplastic polyps, 4.4 years for 1–2 small adenomas, 3.5 years for 3–10 adenomas, and 5.8 years for a family history of CRC (Table 3).
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
Table 4. Percentage of patients with repeat colonoscopy intervals consistent with contemporaneous guidelinesa Referent guidelineb
Findingc Normal Hyperplastic polyps 1–2 small (⬍1 cm) tubular adenoma 3–10 adenomas or any adenoma ⱖ1 cm or high grade dysplasia Family history of colorectal cancere Total consistent with guideline
1997 2003 2006 guideline23 guideline24 guideline25 (nⴝ113) (nⴝ328) (nⴝ74) 40.7 N.R. N.R.
34.7 N.R. 25.0
33.3 27.3 50.0d
37.5
53.3
57.1
54.2
55.4
62.5
43.4
38.1
37.8
a
Adjusted for random effect of providers nested within practice as well as patient age and gender, primary care practice, and the endoscopist’s specialty. b Referent guidelines: the 1997, 2003, and 2006 guidelines were referent, respectively, for colonoscopies performed in 1997–2003 (n⫽113), 2003–2006 (n⫽328), and 2006 (n⫽74). c Multiple findings found on colonoscopy were classified by the finding that would justify the shortest interval for retesting. d Assumes an acceptable retesting interval could include 5, 5–10, or 10 years, as recommended in current (2006) guidelines (see Table 1). e Patients with a family history as an indication for performing the test and a result of normal colonoscopy, hyperplastic polyp(s), 1–2 adenomatous polyps, or unknown polyp type were included only in this row. N.R., no recommendation appears in the guideline.
Discussion Based on the information provided in this sample of correspondence with PCCs, the endoscopists often recommended follow-up colonoscopy at shorter intervals than those advocated at the time of the procedure, or currently, by gastroenterologic societies. Guidelines issued in 1997, 2003, and 2006 have stated consistently that colonoscopy should be repeated 10 years after a normal examination and the 2003 and 2006 guidelines recommended a 5-year follow-up interval,4,23–25 yet the adjusted mean follow-up interval recommended by endoscopists was significantly shorter in both cases. This study adds to an extensive literature on overuse of healthcare services, a phenomenon in the U.S. that fuels higher healthcare costs, consumes resources that can be put to more effective use, and exposes patients to unnecessary risks.32–34 The inference that this study documents overuse of colonoscopy requires the assumptions that the shorter retesting intervals advocated by endoscopists were (1) inappropriate and (2) implemented. Neither of these assumptions can be confirmed by this study, which was designed to examine the advice that endoscopists communicated to PCCs rather than the testing that actually occurred. Clinical considerations known to the endoscopist but not reflected in the records of the PCC could explain the appropriateness of recommended retesting intervals that appear to December 2007
depart from those advocated in practice guidelines. For example, endoscopists may favor a shorter follow-up interval if they question the quality of their examinations (due to incomplete bowel preparation or inadequate luminal views) or are cognizant of relevant symptoms or pathology detected on prior colonoscopies that heighten the patient’s CRC risk. While these undisclosed details may justify the recommendations of endoscopists, the failure to cite them explicitly as a rationale to accompany recommendations compromises the ability of PCCs to vet the appropriateness of the advice. In the correspondence with PCCs examined in this study, there were few notations about the adequacy of the bowel preparation or luminal visualization. Beyond a statement of polyp findings or the rare report of not completing the examination, little justification for repeat-testing intervals was offered. Apart from the appropriateness of the advice given by the endoscopists, the question of whether the frequent retesting actually occurred cannot be confirmed by this study, which neither audited the records or insurance claims of endoscopists or PCCs nor surveyed patients to document their experience. In many instances, however, the advice given in the correspondence analyzed in this study is what will dictate the timing of future follow-up colonoscopy. Whether or not they agree with the recommended follow-up interval or can find in the communication a convincing rationale for the advice, PCCs face medicolegal consequences if they depart from the recommendations documented in the medical record. Informed patients who recall the advice given by the endoscopist also may hold expectations about the timing of follow-up. What is communicated to the PCC and what appears in the medical record of the patient—the data substrate for this study—therefore can supersede the above clinical caveats in determining whether colonoscopy is overused. A longstanding body of work documents the difficulties faced by physicians in adhering to clinical practice guidelines and the various factors that account for this tendency.35– 40 Similar factors may account for the disparity between colonoscopy guidelines and the advice endoscopists appear to be giving PCCs. For example, in many cases, the tendency of endoscopists to advocate retesting at short intervals may relate less to legitimate clinical considerations than to an entrenched misunderstanding about current recommendations or the data on which they are based. Many endoscopists may dispute that evidence, disagree with the guidelines, or favor more aggressive approaches after having experienced adverse outcomes with previous patients in whom CRC escaped detection. Endoscopists who maintain concerns about medicolegal liability41 also may prefer shorter follow-up intervals. Financial incentives also may motivate some endoscopists to endorse shorter retesting intervals. Am J Prev Med 2007;33(6)
475
The literature documents that physicians require time to adapt to new guidelines. Colonoscopy guidelines have changed since 1997 (Table 1), but many of today’s endoscopists may have formed their views about retesting intervals in the era before 1997, when frequent retesting was advocated. Since 1997, guidelines have consistently recommended a 3-year follow-up interval for patients with 3–10 adenomas (or any adenoma ⱖ1 cm or with high-grade dysplasia), and the temporal trend in Table 4 suggests that adherence to that guideline has gradually improved over time, consistent with prevailing theories about the diffusion of new knowledge. Table 4 displays less encouraging trends, however, for adherence to policy for normal examinations. While it might come as little surprise that recommendations made by endoscopists years ago do not adhere to current (2006) guidelines, the data in Table 4 finds the same to be true for adherence to contemporaneous guidelines as well. Whatever the reasons, the overuse of colonoscopy carries profound public health implications. As the guideline panels have documented, there is little evidence that patients benefit by undergoing colonoscopies at shorter intervals than the periods recommended in the guidelines. There is measurable harm from this practice, however, because of the known complication rates associated with colonoscopy. Formal methods are necessary to make precise economic projections, but the impact can be approximated if one assumes: (1) that the rate for colonoscopy complications is: 0.5 death, nine bowel perforations, 20 major hemorrhages, and 50 serious complications per 10,000 colonoscopies12–18; (2) that 14.2 million colonoscopies are performed annually in the United States41,42; (3) that the prevalence of pathologic findings on colonoscopies nationwide resembles the distribution that was observed; (4) that the retesting intervals advocated in the 2006 guidelines are ideal; and (5) that a colonoscopy costs $1200.43 These conditions would result in the performance of 2.85 million excess colonoscopies over the next 10 years. Given the above assumptions, this cascade of procedures would induce approximately 142 deaths, 2561 perforations, 5692 major hemorrhages, and 14,229 serious complications. The excess colonoscopies would generate $3.4 billion in direct costs, consuming limited healthcare resources that could be more effectively applied to closing the gap in needed CRC screening.44 The overuse of colonoscopy reduces the overall cost-effectiveness ratio for the procedure and places greater burden on healthcare personnel to meet heightened volume demands. This study documented poor communication between endoscopists and PCCs. Of the charts with a documented colonoscopy, 20% lacked an endoscopist’s report; the results of the colonoscopy could not be determined from the PCC chart. In 10% of cases, correspondence from the endoscopist made reference 476
to one or more polyps detected on colonoscopy, but the chart contained no biopsy results. Frequently, the chart contained only a procedure note generated immediately after the colonoscopy with the notation that a polyp was found and biopsied. In such cases, recommendations on retesting were made preliminarily by the endoscopist based on the gross appearance of the polyp but prior to receiving pathologic details. This study has several important limitations. First, as noted above, the analysis of endoscopists’ recommendations through the lens of the PCC does not provide an objective assessment of the knowledge, beliefs, records, or actions of endoscopists or of the advice that they provide directly to patients. Second, the covariables listed above that might justify a shorter screening interval were not assessed. Third, charts may be inaccurate with respect to family history, both neglecting to report a significant family history and attributing a “family history” when the family member was distantly related (i.e., not a first-degree relative) or did not have CRC. Fourth, chart audits have known limitations45– 47 and this study lacked resources for dual review and arbitration of coding disagreements. Fifth, as noted earlier, endoscopists need time to adapt to new guidelines, and this analysis held them accountable for advice given in the same year the guidelines were released. The findings of this study underscore the need for system changes to bring the practice of endoscopists and PCCs into closer alignment with published guidelines. In particular, seven system changes seem warranted: (1) Systems are needed to ensure that both endoscopists and PCCs are knowledgeable about guidelines and, in particular, are made aware of changes as new guidelines are issued. Their respective specialty societies can play an important role in facilitating diffusion; (2) Physicians should consider whether other strategies might be more effective than increasing the frequency of colonoscopy in preventing CRC. For example, emphasizing changes in technique to ensure that colonoscopies are performed well may be more effective and safer than repeating the test more frequently48; (3) Practices and healthcare systems should adopt tools to facilitate adherence to the guidelines (e.g., prompts or standing orders that note recommended follow-up intervals alongside check boxes for colonoscopy findings). These reminder systems must be updated as new guidelines are issued; (4) Practices and healthcare systems should enhance communication between the endoscopist and the PCC, in particular to ensure that both the procedure note and the pathology results, which may reach the endoscopist later, are always forwarded to the PCC;
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net
(5) PCCs need systems to flag missing information in the records, such as failure to document whether a screening or surveillance examination occurred, the specific examination findings, and copies of pathology reports. In particular, the failure to collect a complete and accurate family history influences the degree to which PCCs can ensure adequate retesting for those with genetic predispositions for CRC; (6) PCCs should not rely solely on the endoscopist for guidance on follow-up nor should they assume that recommended retesting intervals are consistent with either current guidelines or the patient’s pathologic findings. Physicians who will be engaged in the ongoing care of patients, usually PCCs, are responsible for ensuring that the advice they give patients about follow-up is appropriate for their test results and consistent with recent evidence-based guidelines. As needed, PCCs should amend the recommendations of endoscopists and document the rationale for the change; and (7) Patients should be provided current consumeroriented guidelines and be informed of the specific findings on their examinations, thereby empowering them as change agents to ensure that they are aware of the appropriate follow-up interval for their next test and can advocate on their own behalf. The findings reported here reflect conditions at 10 practices in Virginia and Maryland. Studies currently in progress elsewhere in the United States will expand the evidence base and help confirm whether the phenomenon that was observed is widespread in primary care. Further research will also help to identify and evaluate strategies to change systems of care to help the population avert the public health consequences of overuse of colonoscopy. We would like to thank the Virginia Commonwealth University Department of Family Medicine for supporting this project, the practices that facilitated data collection (Fairfax Family Practice, Front Royal Family Practice, Chesterfield Family Medicine, St. Francis Family Medicine, Vienna Family Medicine, Potomac Physician Associates, Little Falls Practice, Premier Primary Care Physicians, Franconia Family Medicine), and the faculty, residents, and medical students who collected the data (Maria Gaspar; Britta Johnson, MD; Michael Johnson; Katherine Jones; Christine Kerns, RN; Haewon Park; Kelley Prince, MD; Walter Waugh, MD; and Ericka Young, MD). This project could not have been accomplished without their generous assistance. Dr. Krist and Dr. Kerns are faculty members, practicing physicians, and partial owners of Fairfax Family Practice and Front Royal Family Practice, respectively. Otherwise, no authors have any conflicts of interest to report. No financial disclosures were reported by the authors of this paper.
December 2007
References 1. Pignone M, Rich M, Teutsch SM, Berg AO, Lohr KN. Screening for colorectal cancer in adults at average risk: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:132– 41. 2. Colorectal cancer screening. Recommendation statement from the Canadian Task Force on Preventive Health Care. CMAJ 2006;165:206 – 8. 3. Pignone M, Saha S, Hoerger T, Mandelblatt J. Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force. Ann Intern Med 2002;137:96 –104. 4. Screening for colorectal cancer: recommendation and rationale. Ann Intern Med 2002;137:129 –31. 5. Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA 2003;289:1288 –96. 6. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329:1977– 81. 7. Citarda F, Tomaselli G, Capocaccia R, Barcherini S, Crespi M. Efficacy in standard clinical practice of colonoscopic polypectomy in reducing colorectal cancer incidence. Gut 2001;48:753– 4. 8. Atkin WS, Morson BC, Cuzick J. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 1992;326:658 – 62. 9. Thiis-Evensen E, Hoff GS, Sauar J, Langmark F, Majak BM, Vatn MH. Population-based surveillance by colonoscopy: effect on the incidence of colorectal cancer. Telemark Polyp Study I. Scand J Gastroenterol 1999; 34:414 –20. 10. Newcomb PA, Storer BE, Morimoto LM, Templeton A, Potter JD. Longterm efficacy of sigmoidoscopy in the reduction of colorectal cancer incidence. J Natl Cancer Inst 2003;95:622–5. 11. Bertario L, Russo A, Sala P, et al. Predictors of metachronous colorectal neoplasms in sporadic adenoma patients. Int J Cancer 2003;105:82–7. 12. Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000; 343:162– 8. 13. Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343:169 –74. 14. Yousfi M, Gostout CJ, Baron TH, et al. Postpolypectomy lower gastrointestinal bleeding: potential role of aspirin. Am J Gastroenterol 2004;99: 1785–9. 15. Cobb WS, Heniford BT, Sigmon LB, et al. Colonoscopic perforations: incidence, management, and outcomes. Am Surg 2004;70:750 –7; discussion 757– 8. 16. Misra T, Lalor E, Fedorak RN. Endoscopic perforation rates at a Canadian university teaching hospital. Can J Gastroenterol 2004;18:221– 6. 17. Nelson DB, McQuaid KR, Bond JH, Lieberman DA, Weiss DG, Johnston TK. Procedural success and complications of large-scale screening colonoscopy. Gastrointest Endosc 2002;55:307–14. 18. Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an integrated health care delivery system. Ann Intern Med 2006;145:880 – 6. 19. Ransohoff DF. Colon cancer screening in 2005: status and challenges. Gastroenterology 2005;128:1685–95. 20. Sirovich BE, Schwartz LM, Woloshin S. Screening men for prostate and colorectal cancer in the United States: does practice reflect the evidence? JAMA 2003;289:1414 –20. 21. Vijan S, Inadomi J, Hayward RA, Hofer TP, Fendrick AM. Projections of demand and capacity for colonoscopy related to increasing rates of colorectal cancer screening in the United States. Aliment Pharmacol Ther 2004;20:507–15. 22. Ladabaum U, Song K. Projected national impact of colorectal cancer screening on clinical and economic outcomes and health services demand. Gastroenterology 2005;129:1151– 62. 23. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594 – 642. 24. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale–Update based on new evidence. Gastroenterology 2003;124:544 – 60. 25. Winawer SJ, Zauber AG, Fletcher RH, et al. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the U.S. MultiSociety Task Force on Colorectal Cancer and the American Cancer Society. CA Cancer J Clin 2006;56:143–59; quiz 184 –5. 26. Mysliwiec PA, Brown ML, Klabunde CN, Ransohoff DF. Are physicians doing too much colonoscopy? A national survey of colorectal surveillance after polypectomy. Ann Intern Med 2004;141:264 –71.
Am J Prev Med 2007;33(6)
477
27. Starfield B, Shi L, Macinko J. Contribution of primary care to health systems and health. Milbank Q 2005;83:457–502. 28. Tyler CV Jr, Snyder CW. Cancer risk assessment: examining the family physician’s role. J Am Board Fam Med 2006;19:468 –77. 29. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied regression analysis and multivariable methods. 3rd edn. Boston: Duxbury Press, 1998. 30. Behavioral Risk Factor Surveillance System: Prevalence Data: National Center for Chronic Disease Prevention and Health Promotion, 2007. Available online at: http://apps.nccd.cdc.gov/brfss/index.asp. 31. Schoenfeld P, Cash B, Flood A, et al. Colonoscopic screening of averagerisk women for colorectal neoplasia. N Engl J Med 2005;352:2061– 8. 32. Woolf SH. Potential health and economic consequences of misplaced priorities. JAMA 2007;297:523– 6. 33. Schwartz LM, Woloshin S, Fowler FJ Jr, Welch HG. Enthusiasm for cancer screening in the United States. JAMA 2004;291:71– 8. 34. Benefits foregone: too much of the wrong and too little of the right. Based on a presentation by Bernard S. Bloom PhD., Am J Manag Care 1999;5(6 Suppl):S355– 60. 35. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med 2003;348:2635– 45. 36. Woolf SH, Krist A. The liability of giving patients a choice: shared decision making and prostate cancer. Am Fam Physician 2005;71:1871–2. 37. Ardery G, Carter BL, Milchak JL, et al. Explicit and implicit evaluation of physician adherence to hypertension guidelines. J Clin Hypertens (Greenwich) 2007;9:113–9. 38. Flanders SA, Halm EA. Guidelines for community-acquired pneumonia: are they reflected in practice? Treat Respir Med 2004;3:67–77.
478
39. Schers H, Wensing M, Huijsmans Z, van Tulder M, Grol R. Implementation barriers for general practice guidelines on low back pain a qualitative study. Spine 2001;26:E348 –53. 40. McAlister FA, Campbell NR, Zarnke K, Levine M, Graham ID. The management of hypertension in Canada: a review of current guidelines, their shortcomings and implications for the future. CMAJ 2001;164:517–22. 41. Seeff LC, Manninen DL, Dong FB, et al. Is there endoscopic capacity to provide colorectal cancer screening to the unscreened population in the United States? Gastroenterology 2004;127:1661–9. 42. 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. 43. Colorectal cancer: health professionals facts on screening. Washington DC: Centers for Disease Control and Prevention, 2000. Available online at: http://www.cdc.gov/cancer/colorectal/pdf/fs-professional.pdf. 44. Butterly L, Olenec C, Goodrich M, Carney P, Dietrich A. Colonoscopy demand and capacity in New Hampshire. Am J Prev Med 2007;32:25–31. 45. Murff HJ, Patel VL, Hripcsak G, Bates DW. Detecting adverse events for patient safety research: a review of current methodologies. J Biomed Inform 2003;36:131– 43. 46. Worster A, Haines T. Advanced statistics: understanding medical record review (MRR) studies. Acad Emerg Med 2004;11:187–92. 47. Camacho LA, Rubin HR. Assessment of the validity and reliability of three systems of medical record screening for quality of care assessment. Med Care 1998;36:748 –51. 48. Barclay RL, Vicari JJ, Doughty AS, Johanson JF, Greenlaw RL. Colonoscopic withdrawal times and adenoma detection during screening colonoscopy. N Engl J Med 2006;355:2533– 41.
American Journal of Preventive Medicine, Volume 33, Number 6
www.ajpm-online.net