|Home | About | Journals | Submit | Contact Us | Français|
Primary care physicians (hereafter, physicians) play a critical role in the delivery of colorectal cancer (CRC) screening in the U.S. This study describes the CRC screening recommendations and practices of U.S. physicians and compares them to findings from a 1999–2000 national provider survey.
Data from 1266 physicians responding to the 2006–2007 National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening (cooperation rate=75%) were analyzed in 2008. Descriptive statistics were used to examine physicians' CRC screening recommendations and practices as well as the office systems used to support screening activities. Sample weights were applied in the analyses to obtain national estimates.
Ninety-five percent of physicians routinely recommend screening colonoscopy to asymptomatic, average-risk patients; 80% recommend fecal occult blood testing (FOBT). Only a minority recommend sigmoidoscopy, double-contrast barium enema, computed tomographic colonography, or fecal DNA testing. Fifty-six percent recommend two screening modalities; 17% recommend one. Nearly all physicians who recommend endoscopy refer their patients for the procedure. Four percent perform sigmoidoscopy, a 25-percentage-point decline from 1999–2000. Although 61% of physicians reported that their practice had guidelines for CRC screening, only 30% use provider reminders; 15% use patient reminders.
Physicians' CRC screening recommendations and practices have changed substantially since 1999–2000. Colonoscopy is now the most frequently recommended test. Most physicians do not recommend the full menu of test options prescribed in national guidelines. Few perform sigmoidoscopy. Office systems to support CRC screening are lacking in many physicians' practices. Given ongoing changes in CRC screening technologies and guidelines, the continued monitoring of physicians' CRC screening recommendations and practices is imperative.
Colorectal cancer (CRC) is the second leading cause of cancer death in the U.S.1 Screening for CRC is recommended by national expert groups for average-risk adults aged ≥50 years2,3 and has been designated as a high-priority preventive service because of its substantial potential impact on disease burden and its cost effectiveness.4 Screening options recommended by the U.S. Preventive Services Task Force (USPSTF) include annual fecal occult blood testing (FOBT); sigmoidoscopy every 5 years; or colonoscopy every 10 years.3 In 2005, 50% of adults aged ≥50 years in the U.S. had been screened according to these recommendations.5 This rate is notably lower than that of other recommended adult preventive services.6–8
Most cancer screening in the U.S. occurs within routine medical practice, especially primary care. By recommending CRC screening tests, performing them, or referring patients for them, primary care physicians (hereafter, physicians) play a critical role in implementing guidelines and achieving public health targets for CRC screening. Accordingly, physician recommendation has been shown to be a strong correlate of whether or not patients undergo CRC screening.9–11 Practice-level systems to support the translation of physicians' recommendations into clinical delivery are also recognized as an important influence on the utilization of CRC screening.11–13
An earlier study14 used data from a nationally representative survey conducted in 1999–2000 to assess physicians' recommendations for and means of conducting CRC screening in their clinical practices. The study found that 98% of physicians recommended CRC screening to patients, most often with FOBT, flexible sigmoidoscopy, or both, and that many recommended screening at nonstandard starting ages or too-frequent intervals. Subsequent to that study, several major developments in national policies related to CRC screening occurred. These include the expansion of coverage by the Medicare program to include screening colonoscopy for average-risk beneficiaries in 2001,15 the addition of colonoscopy and double-contrast barium enema (DCBE) to updated guidelines by the USPSTF in 2002,16 and the adoption of CRC screening as a Health Plan Employer Data and Information Set performance measure by the National Committee for Quality Assurance17 in 2003. Physicians' CRC screening recommendations and practices may have been influenced, altered, or both, by these and other developments in the rapidly evolving CRC screening field.
To assess physicians' current knowledge, attitudes, recommendations, and practices regarding CRC and three other types of cancer screening, the National Cancer Institute (NCI)—in collaboration with the Agency for Healthcare Research and Quality and the CDC—fielded in 2006–2007 the National Survey of Primary Care Physicians' Recommendations and Practices for Breast, Cervical, Colorectal, and Lung Cancer Screening (healthservices.cancer.gov/surveys/screening_rp/). The current study uses data from the new survey to characterize U.S. physicians' CRC screening recommendations and practices. Results are compared to selected findings from the earlier, 1999–2000 survey.14 The practice-level systems that physicians report having in place to support CRC screening activities, about which there is limited information at the national level, are also described.
Between September 2006 and May 2007, a nationally representative sample of physicians was surveyed. The American Medical Association's Physician Masterfile, which contains demographic and practice information on all allopathic and nearly all osteopathic physicians in the U.S., was used as the sampling frame. Eligible respondents were nonfederal, office-based family physicians, general practitioners, general internists, and obstetricians/gynecologists aged ≤75 years with patient care as their major activity. Physicians who were listed as retired, deceased, or with an address outside the U.S. were considered ineligible. Because of the complexity of the topic of cancer screening and the need for fielding relatively brief questionnaires, one half of physicians received a questionnaire on breast and cervical cancer screening and the other half a questionnaire on colorectal and lung cancer screening.
For the colorectal/lung cancer survey, a systematic, stratified random sample of 2576 physicians was selected, using the four specialty types as the sampling strata. Sampling was designed to provide national estimates of physicians' cancer-screening knowledge, attitudes, recommendations, and practices for the four specialties, combined with a margin of error of ±3% at a 95% CI. The sample was selected after sorting the sampling-frame database by physicians' age group, gender, urban versus rural practice location, U.S. census region, and age in single-year increments.
Of the 2576 physicians in the sample, 92 had requested on the Masterfile that they not be contacted and therefore were excluded from survey procedures, leaving 2484 potential respondents. Several online and directory-assistance resources were searched to obtain current contact information for approximately one third of the sample. In August 2006, telephone calls were placed by trained interviewers to the offices of physicians in the sample to verify physicians' eligibility for participation in the survey and the accuracy of contact information. Physicians were reclassified as ineligible if it was reported during these calls that they did not provide primary care, provided clinical care for <1 day per week, practiced in a federal healthcare facility, or were located outside the U.S. Despite tracing and telephone efforts, valid contact information could not be obtained for 7% of the sample.
In September 2006, a total of 1975 physicians were sent by express mail a packet containing a questionnaire; a cover letter from the NCI describing the study and requesting participation; a fact sheet with additional information about the survey; a letter of support from the American Academy of Family Physicians, the American Society of General Internal Medicine, or the American College of Obstetricians and Gynecologists, depending on the physician's specialty; a $50 honorarium check; and a postage-paid return envelope. Approximately 3 weeks later, a second mailing of the questionnaire was sent to eligible nonrespondents. Follow-up telephone calls were placed to the offices of physicians who were sent the second mailing to verify receipt of the mailed package and to encourage participation in the survey. Physicians were offered the option of completing a telephone interview; only three did so. Additional telephone calls were placed to the offices of nonresponding physicians. In February 2007, a third mailing of the questionnaire was sent by express mail to a sample of 341 nonresponding physicians. Follow-up telephone calls were placed to the offices of those physicians a few days after the mailing and continued through March 2007.
The questionnaire required about 20 minutes to complete. Items asking about physicians' CRC screening recommendations and practices were specific to asymptomatic, average-risk patients and included FOBT, sigmoidoscopy, colonoscopy, DCBE, computed tomographic (CT) colonography, and fecal DNA testing. Items assessing perceived test effectiveness were measured with a 4-point Likert scale and the categories very effective, somewhat effective, not effective, and don't know. Physicians were asked whether they routinely recommended to their patients a specific modality as a CRC screening strategy and, if so, their policies on starting age, test frequency, and stopping age. For sigmoidoscopy and colonoscopy, they were asked whether they performed the procedure or referred patients to another provider for it. Physicians also were asked about the types of systems available in their practices to support cancer-screening activities. These questions included whether the practice had implemented CRC screening guidelines, used reminder systems for CRC screening, provided reports to the physician on his or her CRC screening rates, and the type of medical-record system used. The survey instrumentation is available at healthservices.cancer.gov/surveys/screening_rp/.
Descriptive statistics were used to examine physicians' CRC screening recommendations and practices as well as the office systems available in their practices to support CRC screening activities. Physicians' modality-specific CRC screening recommendations and practices were compared to national data from a survey conducted in 1999–2000.14 A sample weight that adjusts for undercoverage, nonresponse to telephone-screener calls, and survey nonresponse was assigned to each respondent. Sample weights were applied in the statistical analyses to obtain national estimates. SAS version 9.1.3 was used to obtain unweighted sample sizes. SUDAAN version 9.0.1 was used in calculating percentages and 95% CIs to account for the complex survey design. The analyses were conducted in 2008.
A total of 1266 physicians responded to the survey. The absolute response rate was 69.3%. The cooperation rate, which excludes physicians lacking valid contact information, was 75%. Nearly half of the respondents were family physicians or general practitioners (Table 1). Approximately half were aged <50 years. The majority were men, non-Hispanic white, board-certified, graduates of U.S. medical schools, full or part owners of their practices, and without a medical school faculty appointment. More than three quarters spent ≥90% of their professional time providing patient care. The practices of most respondents were in urban locations, comprised five or fewer physicians, and were single-specialty.
Ninety-five percent of respondents expressed the belief that colonoscopy is very effective in reducing CRC mortality (Table 2). Only a minority perceived as very effective guaiac-based FOBT (12%); sigmoidoscopy (16%); DCBE (17%); CT colonography (23%); or fecal DNA testing (7%). In contrast, in 1999–2000, 86% (95% CI=83.9, 87.8) of physicians perceived colonoscopy to be very effective, while FOBT was perceived as very effective by 28% (95% CI=25.6, 30.6); sigmoidoscopy by 49% (95% CI=46.5, 52.0); and DCBE by 33% (95% CI=30.1, 35.3). The 1999–2000 survey did not collect information on CT colonography or fecal DNA testing.
Of the six CRC screening modalities included in the 2006–2007 survey, colonoscopy was routinely recommended to asymptomatic, average-risk patients by the highest percentage of physicians (95%), followed by FOBT (80%); sigmoidoscopy (26%); DCBE (9%); CT colonography (5%); and fecal DNA testing (5%; Table 2). These results contrast sharply with the 1999–2000 national survey, which showed FOBT to be the modality recommended by most physicians (95%); followed by sigmoidoscopy (78%); colonoscopy (38%); and DCBE (14%; Figure 1).
With respect to the age at which physicians recommend that asymptomatic, average-risk patients begin screening with the modality, the 2006–2007 data show that for all of the tests, the majority specify age 50 years (Table 2). However, the percentage who do so is slightly lower for FOBT (58%) and fecal DNA testing (57%) than for the other tests; nearly 30% of physicians who indicated that they recommend FOBT or fecal DNA testing to patients mentioned 40 years as their recommended starting age. In 1999–2000, the percentages of physicians recommending that screening begin at age 50 years were lower for two of the tests: FOBT (46%; 95% CI=42.8, 48.6) and colonoscopy (76%; 95% CI=71.9, 79.8).
The most commonly mentioned frequencies at which physicians recommended that their patients undergo screening with the modality were every year for FOBT (87%) and fecal DNA testing (50%); once every 5 years for sigmoidoscopy (65%), DCBE (56%), and CT colonography (37%); and once every 10 years for colonoscopy (54%; Table 2). Variation in recommended frequencies for applying each of the tests is evident. In 1999–2000, the percentages of physicians recommending guideline-consistent frequencies were lower for sigmoidoscopy (48%; 95% CI=44.5, 50.9) and colonoscopy (17%; 95% CI=14.0, 20.9).
One third or fewer of the physicians who recommended a specific modality for CRC screening indicated that they have a stopping age at which they no longer recommend screening with that test to healthy patients (Table 2). In 1999–2000, lower percentages of physicians reported having a stopping age for three of the CRC screening tests: 13% (95% CI=11.3, 15.2) for FOBT; 23% (95% CI=20.7, 26.0) for sigmoidoscopy; and 19% (95% CI=15.9, 23.1) for colonoscopy. In both surveys, 80 years was the most commonly mentioned stopping age.
Fewer than 1% of physicians reported that they do not routinely recommend CRC screening with any of the six modalities included in the survey (Table 3). The most frequent practice is to recommend two modalities (56%), with FOBT and colonoscopy the most commonly recommended tests (50%). The second most-frequent practice is to recommend three modalities (20%); FOBT, sigmoidoscopy, and colonoscopy are the most-often recommended tests in this practice model (14%). Seventeen percent of physicians recommend a single modality, with colonoscopy the most commonly recommended test (15%). Only 1% of physicians recommend FOBT and no other modalities. Approximately 7% of physicians recommend four or more modalities: FOBT, sigmoidoscopy, colonoscopy, and DCBE or CT colonography or fecal DNA testing.
Of the physicians who recommend sigmoidoscopy, colonoscopy, or both (98%), the great majority refer their patients to another provider for the procedure. Only 4% reported that they perform sigmoidoscopy; 3% perform colonoscopy, and 0.5% indicated that they perform both procedures. The colonoscopy results are very similar to those of the 1999–2000 national survey, but the sigmoidoscopy results differ markedly: in 1999–2000, 29% of U.S. physicians reported that they performed sigmoidoscopy.14
Nearly two thirds of physicians (61%) reported that their practice had implemented guidelines for CRC screening (Table 4). Paper charts (56%) were the predominant type of medical-record system used in physicians' practices; 18% had full electronic medical records (EMRs), 10% had partial EMRs, and 16% were in transition from paper charts to EMRs. Reminder systems were relatively uncommon: 30% of physicians indicated that their practice has a mechanism such as a flag in patients' charts, computer prompt, or flow sheet to remind the physician or other members of the care team that a patient is due for CRC screening; 15% reported that their practice uses mechanisms such as the U.S. mail, telephone calls, e-mail, or personalized web pages to remind patients that they are due for CRC screening. Few physicians (12%) indicated that they receive reports of CRC screening rates for their patient panel.
This study used data from a nationally representative survey of physicians conducted in 2006–2007 and compared them to selected results from an earlier national survey14 to show that physicians' CRC screening recommendations and practices have changed substantially over a 7-year period. Colonoscopy is now the most frequently recommended test; in 1999–2000, it was FOBT. Moreover, despite national guidelines that list multiple test options for CRC screening and that promote the precept of patient choice in deciding on an acceptable screening strategy,18 this study documents declines in the percentages of physicians who routinely recommend FOBT, sigmoidoscopy, and DCBE, with the most dramatic decline for sigmoidoscopy. The patterns of and trends in physician recommendation reported in this study parallel those for CRC test use as documented in the general population and among Medicare beneficiaries.19–21
Another notable finding is that, consistent with physicians' reduced propensity to recommend FOBT, sigmoidoscopy, and DCBE, considerably fewer in 2006–2007 than in 1999–2000 perceived these tests to be very effective in reducing CRC mortality, even though no new evidence has been published in the last 7 years indicating that the tests have less effectiveness than previously established. In contrast, the percentage of physicians who believe that colonoscopy is very effective in reducing CRC mortality increased by nearly ten percentage points between 1999–2000 and 2006–2007.
Although national guidelines in place at the time of this study16,22,23 specified four different tests as acceptable CRC screening options (i.e., FOBT, sigmoidoscopy, colonoscopy, and DCBE) and recommended that physicians present those test options to patients in an informed/shared decision-making process, study results indicate that in practice most physicians are applying the guidelines selectively. More than 50% routinely recommend two tests (and no others), while 17% recommend only one test. Fewer than 10% routinely recommend all four test modalities. These findings are noteworthy in light of research showing that patients have distinct preferences for CRC screening tests24–28 and that, in practices where only one CRC screening test is offered, many patients do not follow through to obtain screening because of concerns about the test.29 Moreover, a growing body of work indicates that patient preferences for CRC screening tests vary according to the information provided about costs and procedure risks as well as by income level and race/ethnicity.30–32
Another practice change documented by this study is a dramatic decline in the percentage of physicians who perform sigmoidoscopy. In 1999–2000, 29% of U.S. physicians reported that they performed this procedure in their practices.14 The present study shows that only 4% do so. This is a large decline over a relatively short time period. This finding provides important context for interpreting data from national surveys of the general population and the Medicare program, which show substantially reduced sigmoidoscopy use since 2000.19–21 Highly publicized editorials and a 2000 American College of Gastroenterology guideline advocating colonoscopy as the preferred CRC screening strategy may have motivated many physicians to stop performing sigmoidoscopies.33,34 Few studies have examined physicians' specific reasons for performing or not performing this procedure. Low procedure volumes, inadequate reimbursement, and lack of time and support staff have been documented as barriers to physicians' provision of sigmoidoscopy in their practices.14,35,36
Survey results from 1999–2000 showed that many U.S. physicians had adopted nonstandard CRC screening practices such as recommending screening at earlier starting ages or shorter intervals—or both—for asymptomatic, average-risk patients.14 The present study indicates modest improvements in the percentages of physicians who identify guideline-consistent starting ages and test intervals for the CRC screening modalities that they routinely recommend. Nevertheless, many physicians continue to recommend that screening begin at age <50 years or be repeated at too-frequent intervals. This is particularly true for colonoscopy, with 43% of physicians who recommend it indicating that they believe patients should undergo the procedure at intervals more frequent than once every 10 years. The propensity of endoscopists to recommend repeat colonoscopy at too-frequent intervals also has been documented.37,38 The overuse of colonoscopy has implications for patients' exposure to procedure risks as well as for national capacity to achieve higher levels of CRC screening and follow-up among age-eligible adults. Future work will examine factors associated with physicians who report nonguideline-consistent CRC screening practices.
Another important study finding is that many physicians' practices lack office systems to support CRC screening. Slightly more than 60% reported that their practice had implemented CRC screening guidelines, and 28% indicated that their office uses a full or partial EMR. Fewer than one third indicated that their practice has a reminder system to prompt the physician about CRC screening, while fewer than a quarter use reminders to prompt patients. Few physicians receive reports on CRC screening rates for their patient panel. The importance of office systems to increasing CRC screening rates in practice is well established.12,13;39–41 In particular, the great potential of EMRs in promoting appropriate screening within primary care practices has been noted.42 Given the heavy reliance of physicians on referral to other providers for endoscopy and the limited office systems to support CRC screening recommendations/activities documented by this study, it is not surprising that gaps remain between high levels of physician recommendation and actual CRC screening rates.5
A limitation of this study is that it is based on physicians' reports of their recommendations and practices; self-reported data were not validated with other data sources such as medical records or claims. Medical-record and claims data, however, may not be entirely accurate sources of information about physicians' CRC screening recommendations and practices; under-ascertainment of screening practices in these data sources has been documented.43,44 Another limitation is that, for physicians who recommended more than one test modality, the precise patterns of use of the multiple modalities (e.g., FOBT or colonoscopy versus FOBT and colonoscopy) are unknown. This is because the survey items ascertaining physicians' recommendations did not ask about test combinations or preferences for one test over others.
Study strengths include the use of a nationally representative survey with a large sample size and high response rate. Moreover, surveys such as this one can provide more detailed information about physicians' recommendations and practices than is typically available in medical-record or claims data.
Raising CRC screening rates remains a public health challenge. Modeling has shown that CRC mortality in the U.S. could be reduced by 50% within the next decade, largely through increased screening uptake.45 Yet national surveys of the general population continue to show suboptimal screening use, along with lack of awareness of the need for screening and lack of recommendation from a doctor to obtain it. These continue to be noted as key reasons why many age-eligible adults are not screened.5 At a time when many physicians offer patients only one or two of the screening options (e.g., colonoscopy and FOBT), one updated guideline has expanded the menu of recommended test options to include CT colonography and fecal DNA testing2; another states that patients may no longer need screening when they become aged 75 or 85 years.3 How physicians might operationalize these new guidelines in practice, particularly the more extensive test menu and suggested stopping ages, requires future study.
Funding support for this study was provided by the National Cancer Institute (contract number N02-PC-51308); the Agency for Healthcare Research and Quality (inter-agency agreement numbers Y3-PC-5019-01 and Y3-PC-5019-02); and the CDC (inter-agency agreement number Y3-PC-6017-01).
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the National Cancer Institute, the Agency for Healthcare Research and Quality, or the CDC.
No financial disclosures were reported by the authors of this paper.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.