- Email: [email protected]
Identifying opportunities for improved colorectal cancer screening in primary care
Preventive Medicine 39 (2004) 239 – 246 www.elsevier.com/locate/ypmed
Identifying opportunities for improved colorectal cancer screening in primary care Parita Patel, M.D., a,* Samuel N. Forjuoh, M.B., Ch.B., Dr.P.H., a Andrejs Avots-Avotins, M.D., Ph.D., b and Tushar Patel, M.B., Ch.B. b a
Department of Family & Community Medicine, Scott & White Clinic, Texas A&M University System Health Science Center College of Medicine, Temple, TX 76508, USA b Division of Gastroenterology, Department of Medicine, Scott & White Clinic, Texas A&M University System Health Science Center College of Medicine, Temple, TX, 76508, USA Available online 10 May 2004
Abstract Background. Although current recommendations advocate screening persons 50 years of age or older for colorectal cancer (CRC), actual screening practice is highly variable among primary care physicians (PCPs). Knowledge of the factors that influence whether or not screening is offered during a clinic visit is essential to develop effective screening strategies. Methods. A cross-sectional telephone survey of one in four randomly selected patients aged 50 years or older (n = 400) attending a primary care clinic within an integrated health care system in central Texas was conducted. A survey of all PCPs (n = 32) at the practice sites was also administered. Results. The visit type was an important determinant of whether CRC screening was discussed, with most discussion occurring during visits for physicals (P < 0.0001). This finding was corroborated by the physician survey. Patient age and education were also associated with a higher likelihood of having been offered CRC screening (P = 0.009 and 0.014, respectively). Patient race, gender, primary language, PCP, or clinics attended were not significantly associated with the discussion of CRC screening. Conclusions. Discussions regarding CRC screening are most likely to occur during preventive care visits. Thus, facilitating preventive visits especially for the elderly represents an opportunity to improve CRC screening rates in primary care practice. D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. Keywords: Cancer screening; Prevention; Health care practice
Introduction Colorectal cancer (CRC) is the second leading cause of cancer deaths in the United States [1]. In 2002, 148,300 new cases were diagnosed and 56,600 people died of this disease. Early detection has been shown to decrease CRC mortality [2]. As a result, many organizations such as the American Cancer Society, the National Cancer Institute, the American Abbreviations: BE, barium enema; CI, confidence intervals; CRC, Colorectal cancer screening; CS, colonoscopy; FOBT, fecal occult blood test; FS, flexible sigmoidoscopy; OR, crude odds ratio; PCP, primary care physician. * Corresponding author. Department of Family & Community Medicine, College of Medicine, Scott & White Memorial Hospital, Texas A&M University System Health Science Center, 2401 South 31st Street, Temple, TX 76508. Fax: +1-254-680-1234. E-mail address: [email protected] (P. Patel).
Gastroenterological Association, and the U.S. Preventive Services Task Force endorse routine screening for CRC [3,4]. Despite these recommendations and the compelling evidence supporting screening for CRC, most persons at average risk are not screened [5,6]. The individual and societal benefits of CRC screening cannot be realized unless screening activities are appropriately promoted and completed. The delivery of preventive services such as cancer screening by primary care practitioners depends on multiple factors such as the practice setting, payer mix, insurance, and available resources. In addition, cancer screening is dependent on individual patient and physician characteristics. Patient-related factors include age, gender, education, ethnicity, health care beliefs, and socioeconomic status, whereas physician factors may include knowledge, experience, access to specialists and tests, and perceptions towards screening. Many studies have reported these and other specific
0091-7435/$ - see front matter D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2004.03.026
240
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
individual factors as contributors to suboptimal screening for CRC [7 –13]. Although arbitrarily separated into patient, physician, and health care system-related factors, these multiple factors are highly interrelated [14]. Because these factors can dynamically influence each other, factors determining the delivery of CRC screening will reflect any limitations or constraints placed by the practice or health care delivery setting and vary within each practice setting. Although the dynamics of the patient –physician interaction have a powerful effect on screening practice and recommendations, the presence of multiple potentially confounding factors primarily related to practice setting has precluded an adequate evaluation of its role in reported community-based or national studies. Examples of extrinsic factors that can influence studies of the patient –physician interaction include the availability of local facilities and/or cognitive and technical skills to perform flexible sigmoidoscopies, the availability and access to gastroenterologists or radiologists, variations in managed care screening guidelines, and other reimbursement-related issues. CRC screening rates may be improved by increasing discussions about CRC screening between physicians and their patients. Lack of physician recommendation ranks as one of the top three reasons patients do not undergo CRC screening [8,15], and surveys have indicated that less than 50% of patients at risk are counseled by their doctors to undergo testing [1,16]. A recent study emphasized the importance of physician recommendations and called for more studies to examine factors influencing discussions of CRC screening by physicians [17].
Clearly, there is a need to better understand the barriers affecting the actual offering of CRC screening by primary care physicians (PCPs). Furthermore, the design of globally applicable interventions to increase discussion of CRC screening in primary care practice requires knowledge of specific factors that influence whether or not CRC screening is offered. However, there is a lack of information regarding these factors. To identify opportunities for improved screening, we sought to elucidate specific factors that could influence the offering of CRC screening during a patient – physician encounter. A conceptual model was developed and separate patient and physician surveys addressed potential influencing factors. The study was performed within the context of an integrated health care system to decrease the impact of extrinsic confounding health care system-related variables that could potentially influence whether or not CRC screening was offered during a patient – physician interaction. Thus, we expect that these results will be applicable and generally relevant to patient –physician interactions in primary care practice.
Methods Study design and setting A conceptual model was developed based on the work by Myers et al. [18] (Fig. 1). This model was used to identify potential variables that may influence the discussion of CRC
Fig. 1. Conceptual model on which the study was based. This model was used to identify potential variables that may influence the discussion of CRC screening in primary care practice.
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
screening in primary care practice. These variables were considered for further evaluation as determinants of whether or not screening was offered during a physician – patient encounter. The specific role of selected factors was evaluated by separate patient- and physician-directed surveys. The study was approved by the Scott & White Institutional Review Board. A cross-sectional telephone survey was administered to patients scheduled to see a primary care physician (PCP) in one of four family medicine clinics of the Scott & White Healthcare System between March and May 2002. All enrolled patients were members of the Scott & White Health Plan and had their primary insurance through this Health Maintenance Organization, sometimes in conjunction with secondary payers such as Medicare. The clinics are located in Killeen, Belton, and Temple (two clinics) in Bell County, Texas. Each of the clinics has the ability to perform flexible sigmoidoscopy (FS) and fecal occult blood testing (FOBT) within their facility. The physician-directed survey was administered to the physicians at each of the practice sites. All physicians were board certified in family practice and performed flexible sigmoidoscopies. Specific CRC screening guidelines had been made available to all physicians at the four clinics through lectures given by one of the investigators (AAA), along with distribution of printed health plan guidelines. Identical access for colonoscopy and radiological studies was available to all physicians at a single site (Scott & White Memorial Hospital). Patient survey All patients aged 50 years or older, who were scheduled to see a PCP at one of the four clinics and were members of the Scott & White Health Plan, were eligible for enrollment into the study. All eligible patients were given a card by a front desk attendant before their clinic visit inviting them to participate in a survey designed to study healthcare practices. The specific subject of the survey (CRC screening) was not mentioned on the card. Of those who indicated a willingness to participate, one in four was randomly selected to undergo a telephone interview with a research assistant who is bilingual in English and Spanish. The interview was completed within a week of the clinic visit. The telephone survey collected demographic data including gender, age, race, education, and the PCP of each individual. The participants were then asked about any discussion of CRC screening and about any prior screening for CRC. The terminology used for descriptions of the specific testing for CRC screening was adapted from the Behavioral Risk Factor Surveillance System, which is available from the Centers for Disease Control and Prevention [19].
241
about physician age, gender, experience (years in practice), as well as physicians’ beliefs regarding CRC screening and their perceptions regarding their actual CRC screening practice. The physicians at each site had not been informed of the patient survey, which was performed before the administration of the physician survey. Physicians were asked how often they discussed CRC screening with patients over the age of 50 during specific types of patient encounters and how often they specifically discussed flexible sigmoidoscopy (FS), fecal occult blood testing (FOBT), barium enema (BE), or colonoscopy (CS) with patients who were offered CRC screening. Response choices were ‘‘never,’’ ‘‘rarely,’’ ‘‘sometimes,’’ ‘‘often,’’ or ‘‘always.’’ Statistical analysis All data were collected by a research assistant and entered on a data collection sheet. Missing data were excluded from further analysis. For the patient survey, two primary outcome variables were considered: discussion of CRC screening at the most recent clinic visit and ever (overall visits). The former was measured by assessing whether the patient reported that at their most recent visit their PCP discussed CRC screening, did not discuss screening, or was undecided. Responses were dichotomized as ‘‘yes’’ or ‘‘no/did not remember’’ for analysis. The latter was measured by aggregating patient responses for three questions on the survey: whether the patient had ever discussed CRC screening with their PCP at any visit; whether the patient had ever discussed CRC screening with any other physician at any visit; and whether the patient discussed CRC screening with their PCP at their most recent visit. Responses were similarly dichotomized as ‘‘yes’’ or ‘‘no/did not remember’’ for analysis. These outcome variables were related with factors that were predicted to influence whether or not screening is offered based on our conceptual model. All other variables in the patient survey were informative questions focusing on the patient’s understanding of CRC screening, and more specifically on FS, FOBT, BE, and CS. Frequencies were determined for all variables. A chi-square measure of association ( P < 0.05) tested for levels of association between selected variables and the primary outcome variables. Crude odds ratios (ORs) and 95% confidence intervals (CIs) for having ever discussed CRC screening were computed to quantify degree of associations. Finally, multiple logistic regression modeling was used to estimate adjusted ORs and 95% CIs.
Results Patient survey response rates and patient characteristics
Physician survey A separate written survey was administered to all PCPs in the four clinics. This questionnaire collected information
Of 2,716 patients who were invited to participate in the study, 1,620 (58.1%) consented, of which 400 patients were randomly selected and contacted for a telephone
242
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
interview within 1 week of their clinic visit. The median age of the participants was 65 years (mean = 65.4 years; SD = 9.8), with the majority of patients (65.1%) being in the 55- to 74-year range, which is representative of the primary target group for CRC screening. Nearly two-thirds (63.8%) of the participants were women. The majority of visits were for acute or urgent care or follow-up or chronic care, which is reflective of the practice pattern at the survey clinic sites. CRC screening discussion More than half of the patients (n = 214; 53.5%) reported that they had discussed CRC screening with their PCP during a previous clinic visit. An additional 47 patients reported having discussed CRC screening with some other physician during a previous clinic visit. Thus, of the 400 patients, 139 (34.8%) reported never discussing CRC screening with their PCP or any other physician at any visit before the most recent visit. At their most recent clinic visit, 53 (13.3%) of the 400 surveyed patients reported to have discussed CRC screening with their PCP. This number included 44 (11.0%) who were discussing CRC screening again and only 9 (2.3%) who were discussing CRC screening for the first time with any physician. The characteristics of the 130 patients who indicated that they had never discussed CRC screening at any time including their most recent clinic visit were compared with those of the 270 patients who indicated that they had discussed CRC screening previously or during their most recent office visit (Table 1). Discussion of CRC screening was significantly associated with patient age, education, and belief in CRC screening, as well as the type of clinic visit. Increasing age was associated with a significantly lower likelihood of having discussed CRC screening. Patients with less education were significantly less likely to have discussed CRC screening than those with a higher education. Since health beliefs can influence colorectal cancer screening behavior [20], we assessed belief in CRC screening by asking patients if they felt that people of their age should be screened for colorectal cancer. Those respondents who did not believe in CRC screening were significantly less likely to have discussed it. Discussion of CRC screening was also significantly less likely during clinic visit for acute or urgent care and follow-up or chronic care. However, there was no significant association of CRC screening discussion with clinic attended (P = 0.56), patient gender (P = 0.49), race (P = 0.80), or primary language (P = 0.69). After multivariate adjustment, clinic visit type maintained the strongest association with the likelihood of having discussed CRC screening. Patient education was no longer significantly associated with having CRC screening discussion. The association of having discussed CRC screening with patient belief in CRC screening also became marginal. Compared to patient clinic visits for follow-up or chronic
Table 1 Comparison between patients who have previously discussed (n = 270) and those who have never discussed (n = 130) CRC screening Characteristic
Clinic Killeen Belton Santa Fe Northside Gender Male Female Age group < 55 years 55 – 64 years 65 – 74 years 75 years or older Race White Non-White Education groups GED or less Some college/bus/tech College degree/postgrad Primary language English Other Belief in CRC screening Yes No Visit type Acute care/urgent care Physicals Follow-up/chronic care
CRC screening CRC screening P value previously discussed never discussed n (%) n (%) 0.56 163 25 60 22
(66.0) (64.1) (74.1) (66.7)
84 (34.0) 14 (35.9) 21 (25.9) 11 (33.3)
101 (69.7) 169 (66.3)
44 (30.3) 86 (33.7)
0.49
0.009 48 95 84 43
(76.7) (72.5) (65.1) (54.4)
13 36 45 36
(21.3) (27.5) (34.9) (45.6) 0.80
230 (67.3) 40 (69.0)
112 (32.7) 18 (31.0)
105 (60.3) 84 (77.1) 78 (68.4)
69 (39.7) 25 (22.9) 36 (31.6)
266 (67.7) 4 (57.1)
127 (32.3) 3 (42.9)
252 (69.0) 18 (51.4)
113 (31.0) 17 (48.6)
106 (65.8) 65 (86.7) 99 (60.4)
55 (34.2) 10 (13.3) 65 (39.6)
0.014
0.69
0.034
< 0.0001
care, patients visiting the doctor’s office for physicals were four times as likely to have discussed CRC screening (adjusted OR = 4.09, 95% CI = 1.91 – 8.74). Patients under 55 years were three times (adjusted OR = 2.99, 95% CI = 1.31– 6.83) as likely to have discussed CRC screening as those aged 75 years or older (Table 2). Type of screening discussion and follow-up of screening recommendations Of the 270 patients who reported prior CRC screening discussion at any time, 210 (77.8%) had discussed FOBT, 158 (58.5%) had discussed FS, 115 (42.6%) had discussed BE, and 83 (30.9%) had discussed CS. Of note, FOBT was described to patients as being a test that may use a special kit at home to determine if the stool contains blood. Actual performance of discussed screening tests by patients varied widely. Of the 210 patients who reported to have discussed FOBT, only 88 (41.9%) had performed the test within the previous year. However, of those 158 patients who had discussed FS, 137 (86.7%) had performed the test within the past 5 years. Comparable statistics for BE within the past 5 years and CS within the past 10 years were 57.4% and 69.3%, respectively.
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
243
Table 2 Odds ratios and 95% confidence intervals for having ever discussed CRC screening Characteristic
Crude OR (95% CI)
Clinic Killeen Belton Santa Fe Northside Gender Male Female Age group < 55 years 55 – 64 years 65 – 74 years 75 years or older Race White Non-White Education groups GED or less Some college/bus/tech College degree/postgrad Primary language English Other Belief in CRC screening Yes No Visit type Acute care/urgent care Physicals Follow-up/chronic care
Adjusted P value
OR (95% CI)
0.56 0.97 (0.45 – 2.10) 0.89 (0.34 – 2.37) 1.43 (0.59 – 3.44) Reference
0.39 1.30 (0.57 – 2.96) 1.32 (0.47 – 3.68) 2.04 (0.80 – 5.19) Reference
0.49 1.17 (0.75 – 1.81) Reference
0.53 1.16 (0.73 – 1.86) Reference
0.01 3.09 (1.45 – 6.58) 2.21 (1.23 – 3.97) 1.56 (0.88 – 2.77) Reference
0.04 2.99 (1.31 – 6.83) 2.05 (1.09 – 3.87) 1.46 (0.80 – 2.68) Reference
0.80 0.92 (0.51 – 1.68) Reference
0.61 1.20 (0.60 – 2.39) Reference
0.02 0.70 (0.43 – 1.16) 1.55 (0.85 – 2.82) Reference
0.11 1.11 (0.64 – 1.94) 1.88 (0.99 – 4.57) Reference
0.56 1.57 (0.35 – 7.12) Reference
0.41
0.04
0.05 2.13 (0.99 – 4.57) Reference
0.001
Fig. 2. Frequency of screening test used amongst persons with prior CRC screening. The bars illustrate the number of survey participants who had previously undergone a CRC screening test. The shaded portion represents those persons who reported testing was for the purpose of cancer screening, whereas the nonshaded portion of the bar represents those persons who reported testing was for purposes other than cancer screening or for unknown reasons (FOBT = fecal occult blood testing).
had undergone FS, FOBT, CS, or BE testing, respectively (Fig. 2). The majority of participants indicated that they would repeat the screening tests. Interestingly, many indicated they would do so if recommended by their physician, highlighting the importance of the patient – physician interaction. Participants identified BE as the most likely screening test not to be repeated (8.6%, 10/115), followed by CS 3.6% (3/83) and FS 3.2% (5/158). As expected, the reluctance rate for FOBT (1.9%) was quite low (4/210). Those who would not repeat any screening test cited pain as the most common reason (13/21). Physician survey response rates and CRC screening discussion
2.12 (0.36 – 12.42) Reference
2.11 (1.05 – 4.24) Reference
1.27 (0.81 – 1.99) 4.26 (2.04 – 8.90) Reference
P value
Of the 36 physicians surveyed, 32 (88.9%) responded. The median age of the respondents was 43 years (mean 42; range: 30 –57 years). There were 25 males and 7 females. The median number of years of practice experience was 12 years (mean 13; range 1 –25 years). While all physicians work in a clinic setting with their own panel of patients, some of the physicians have significantly smaller practices due to administrative responsibilities. Two
0.001 1.22 (0.76 – 1.97) 4.09 (1.91 – 8.74) Reference
Reasons for screening Although many patients were unaware of the reason for testing, screening for cancer was identified as the reason given by 53.8%, 32.1%, 59.0%, and 29.6% of persons who
Table 3 Physician reported initiation of CRC screening discussion by encounter type (n = 32)a Type of visit
Never, n (%)
Rarely, n (%)
Sometimes, n (%)
Often, n (%)
Always, n (%)
Acute care
6 (18.8) 0 (0.0) 1 (3.2)
20 (62.5) 0 (0.0) 2 (6.5)
5 (15.6) 0 (0.0) 18 (58.1)
1 (3.1) 15 (48.9) 9 (29.0)
0 (0.0) 17 (53.1) 1 (3.2)
Physicals Chronic care a
May not add to total due to missing data.
244
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
Table 4 Physician reported preferences for CRC screening tests (n = 32) Test Barium enema
Never, Rarely, Sometimes, Often, Always, n (%) n (%) n (%) n (%) n (%)
0 (0.0) Colonoscopy 0 (0.0) Fecal occult blood testing 0 (0.0) Flexible sigmoidoscopy 0 (0.0)
10 (31.3) 3 (9.4) 0 (0.0) 0 (0.0)
11 (34.4) 10 (31.3) 3 (9.4) 2 (6.2)
5 (15.6) 10 (31.3) 10 (31.3) 10 (31.3)
6 (18.7) 9 (28.1) 19 (59.4) 20 (62.5)
respondents (6.3%) indicated that they would not screen patients for CRC, whereas 23 (71.9%) indicated they would and 7 did not respond to the question. Twenty-four respondents indicated their perception of the proportion of their patients who had been screened for CRC, with a median of 65% (mean, 60%; range 20 – 98%). Most physicians reported to have discussed CRC screening always or often during patient clinic visits for physicals, and rarely, if at all, during other types of visits (Table 3). FOBT and FS were more likely to be discussed always with patients than BE and CS (Table 4).
Discussion An understanding of the factors associated with whether or not discussion of CRC screening is initiated and offered during a patient – physician encounter may result in the development of interventions that could dramatically increase screening rates. This study identified age as an important patient factor associated with the likelihood of discussing CRC screening with their physician. The study also demonstrated that most discussion of CRC screening is done at appointments for physicals; a finding that was corroborated by a survey of the physicians at the various practice sites. No physician factors were identified to be associated with the likelihood of offering CRC screening to patients. Encouragingly, we found that CRC screening had been discussed at some point with 67.5% of patients at risk. However, the frequency of screening, particularly with regards to FOBT, was not always optimal. A major strength of this study is that several potentially confounding factors were eliminated in the setting of an integrated health care system. These factors include identical access to local resources such as flexible sigmoidoscopy, specialist resources such as gastroenterologists, radiologists, oncologists, or surgeons, uniform screening guidelines, and a managed care population. Each of these factors can influence whether or not CRC screening is offered during a patient encounter. Our findings are especially relevant to managed care systems particularly given the recent inclusion of CRC screening as a benchmark Health Plan Employer Data and Information Set measure, which will promote focused efforts to improve screening rates.
Our study demonstrated that the likelihood of a patient having discussed CRC screening decreased with increasing age. This contrasts with findings of recent studies of actual screening behavior that indicate a higher likelihood of screening of older patients.[16,21]. There are several possible reasons for these observations. We speculate that physicians may be more likely to offer and discuss CRC screening with younger, more healthy persons but less likely to consider it for their older patients who may have more comorbid conditions. Increasing awareness of the age at which CRC screening should begin may also have led to an increase in CRC screening discussions with the younger patients. Another possibility is that younger patients are more aware of the benefits of preventive health strategies and are taking a more active role in their health care by asking their physicians to discuss screening with them. In our study, 72% of the patients undergoing FOBT reported that they had this test done for reasons other than cancer screening or for unknown reasons. This may reflect a gap in patient understanding of the test and its implications. Alternatively, the physician may not have adequately communicated the reasons for the testing. This raises the possibility that although patients are willing to undergo testing recommended by their physician, many may have a poor understanding of the complex testing options available for CRC screening. Since complexity of issues surrounding testing has been identified as a potential barrier to discussion of other cancer screening tests [22 – 24], identifying methods to simplify such issues would likely increase screening. CRC screening was more likely to be discussed during routine physicals. This is likely due to the inherent focus on preventive health during physicals and the increased time available for such appointments. This assertion is supported by a study by Dunn et al. [25] in which lack of time was identified as one of the most common barriers to discussing breast and prostate cancer screening tests. Follow-up and chronic care visits are identified in our study as providing inadequate opportunities to address screening especially in a Medicare population for whom routine physical appointments are not reimbursed. Because multiple screening testing options are available for CRC screening, adequate patient counseling is required for shared decision making. Physician perceptions about the time required for an adequate discussion may influence whether or not screening is discussed during a specific visit. Unless preventive care visits are facilitated, for example, by appropriate reimbursement, innovative approaches to identify and achieve adequate patient counseling within the context of visits for purposes other than general health maintenance will be required. In contrast to other studies, gender did not appear to play a role in the actual offering of CRC screening. It has been proposed that women may be more likely to be offered CRC screening, as they are more likely to attend clinics for other
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
preventive screening. However, no such difference was identified in our study. Despite the different clinic sizes and location, the offering of CRC screening did not appear to be different in each of the clinics. This may represent the benefits of an integrated health care system. In this setting, we expect that interventions to improve screening are also likely to be similar at all practice sites. A potential limitation of this study is the reliance on patient recall. However, we expect that the impact of this will be minimal since patient surveys were conducted within a week of the clinic visit, and patients had previously provided consent to be contacted regarding our health care practices. Furthermore, patient surveys correlate well with chart audits and are preferable to physician surveys for cancer screening practices [22]. Nevertheless, inadequate patient recall regarding prior testing that may be known to the physician remains a limitation. Although our study was performed within the context of a specific managed care setting, we believe that these results are generalizable because of the exclusion of specific practice setting-related confounding factors, which was one of the primary aims of the study. In other settings, additional factors such as local referral practices, facilities, access, reimbursement, and payer mix may also influence discussions regarding CRC screening. Finally, the small numbers of patients in some groups such as primary language may have resulted in a reduced power to detect specific differences. Although CRC screening has been shown to reduce mortality, the full benefits of screening can only be achieved if the population is adequately screened. Improved screening of persons at average risk for CRC cancer in primary care is thus essential. This study investigated important variables associated with the likelihood of physicians discussing CRC screening with their patients. Age has been identified in other studies as a factor influencing the likelihood of having actually undergone screening. Since the discussion of screening is a vital determinant of patients’ decisions to eventually undergo screening tests, targeted interventions aimed at older and perhaps less educated patients in individual practices would be expected to improve screening rates. Both surveys also identified visit type as being a crucial factor. It is therefore important to promote visits for preventive health that allow adequate time to focus on discussion of CRC screening in addition to other recommended preventive issues. Providing appropriate reimbursement for such visits thus represents an opportunity to improve CRC screening in primary care practice.
Acknowledgments This study was supported by a grant award to Dr. Parita Patel from the Scott & White Memorial Hospital, Scott, Sherwood and Brindley Foundation. The assistance of Rinska Flores, Hope Gonzalez, and Rebecca Adams in
245
conducting the survey and of Susan Cromwell and Mark Riggs in analysis of the survey data is greatly appreciated.
References [1] Trends in screening for colorectal cancer—United States, 1997 and 1999. MMWR Morb Mortal Wkly Rep 2001;50(9):162 – 6. [2] Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA 2003;289(10):1288 – 96. [3] Walsh JM, Terdiman JP. Colorectal cancer screening: clinical applications. JAMA 2003;289(10):1297 – 302. [4] Winawer S, Fletcher R, Rex D, Bond J, Burt R, Ferrucci J, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale—Update based on new evidence. Gastroenterology 2003; 124(2):544 – 60. [5] Anderson LM, May DS. Has the use of cervical, breast, and colorectal cancer screening increased in the United States? Am J Public Health 1995; 85(6):840 – 2. [6] Wender RC. Barriers to screening for colorectal cancer. Gastrointest Endosc Clin N Am 2002;12(1):145 – 70. [7] Harewood GC, Wiersema MJ, Melton III LJ. A prospective, controlled assessment of factors influencing acceptance of screening colonoscopy. Am J Gastroenterol 2002;97(12):3186 – 94. [8] Rawl SM, Menon U, Champion VL, Foster JL, Skinner CS. Colorectal cancer screening beliefs. Focus groups with first-degree relatives. Cancer Pract 2000;8(1):32 – 7. [9] Kremers SP, Mesters I, Pladdet IE, van den BB, Stockbrugger RW. Participation in a sigmoidoscopic colorectal cancer screening program: a pilot study. Cancer Epidemiol Biomark Prev 2000; 9(10):1127 – 30. [10] Taylor ML, Anderson R. Colorectal cancer screening: physician attitudes and practices. WMJ 2002;101(5):39 – 43. [11] Cooper GS, Yuan Z, Veri L, Rimm AA, Stange KC. Colorectal carcinoma screening attitudes and practices among primary care physicians in counties at extremes of either high or low cancer case-fatality. Cancer 1999;86(9):1669 – 74. [12] Cooper GS, Fortinsky RH, Hapke R, Landefeld CS. Factors associated with the use of flexible sigmoidoscopy as a screening test for the detection of colorectal carcinoma by primary care physicians. Cancer 1998;82(8):1476 – 81. [13] Klabunde CN, Frame PS, Meadow A, Jones E, Nadel M, Vernon SW. A national survey of primary care physicians’ colorectal cancer screening recommendations and practices. Prev Med 2003;36(3):352 – 62. [14] Petersen GM. Barriers to preventive intervention. Gastroenterol Clin North Am 2002;31(4):1061 – 8 [viii]. [15] Weitzman ER, Zapka J, Estabrook B, Goins KV. Risk and reluctance: understanding impediments to colorectal cancer screening. Prev Med 2001;32(6):502 – 13. [16] Cokkinides VE, Chao A, Smith RA, Vernon SW, Thun MJ. Correlates of underutilization of colorectal cancer screening among U.S. adults, age 50 years and older. Prev Med 2003;36(1):85 – 91. [17] Taylor V, Lessler D, Mertens K, Tu SP, Hart A, Chan N, et al. Colorectal cancer screening among African Americans: the importance of physician recommendation. J Natl Med Assoc 2003; 95(9):806 – 12. [18] Myers RE, Turner B, Weinberg D, Hauck WW, Hyslop T, Brigham T, et al. Complete diagnostic evaluation in colorectal cancer screening: research design and baseline findings. Prev Med 2001; 33(4):249 – 60. [19] Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System Questionnaire. Atlanta (GA): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2002.
246
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
[20] Macrae FA, Hill DJ, St John DJ, Ambikapathy A, Garner JF. Predicting colon cancer screening behavior from health beliefs. Prev Med 1984;13(1):115 – 26. [21] Walsh JM, Posner SF, Perez-Stable EJ. Colon cancer screening in the ambulatory setting. Prev Med 2002;35(3):209 – 18. [22] Montano DE, Phillips WR. Cancer screening by primary care physicians: a comparison of rates obtained from physician self-report, patient survey, and chart audit. Am J Public Health. 1995;85(6): 795 – 800.
[23] Baier M, Calonge N, Cutter G, McClatchey M, Schoentgen S, Hines S, et al. Validity of self-reported colorectal cancer screening behavior. Cancer Epidemiol Biomark Prev 2000;9(2):229 – 32. [24] Gordon NP, Hiatt RA, Lampert DI. Concordance of self-reported data and medical record audit for six cancer screening procedures. J Natl Cancer Inst 1993;85(7):566 – 70. [25] Dunn AS, Shridharani KV, Lou W, Bernstein J, Horowitz CR. Physician – patient discussions of controversial cancer screening tests. Am J Prev Med 2001;20(2):130 – 4.
Identifying opportunities for improved colorectal cancer screening in primary care Parita Patel, M.D., a,* Samuel N. Forjuoh, M.B., Ch.B., Dr.P.H., a Andrejs Avots-Avotins, M.D., Ph.D., b and Tushar Patel, M.B., Ch.B. b a
Department of Family & Community Medicine, Scott & White Clinic, Texas A&M University System Health Science Center College of Medicine, Temple, TX 76508, USA b Division of Gastroenterology, Department of Medicine, Scott & White Clinic, Texas A&M University System Health Science Center College of Medicine, Temple, TX, 76508, USA Available online 10 May 2004
Abstract Background. Although current recommendations advocate screening persons 50 years of age or older for colorectal cancer (CRC), actual screening practice is highly variable among primary care physicians (PCPs). Knowledge of the factors that influence whether or not screening is offered during a clinic visit is essential to develop effective screening strategies. Methods. A cross-sectional telephone survey of one in four randomly selected patients aged 50 years or older (n = 400) attending a primary care clinic within an integrated health care system in central Texas was conducted. A survey of all PCPs (n = 32) at the practice sites was also administered. Results. The visit type was an important determinant of whether CRC screening was discussed, with most discussion occurring during visits for physicals (P < 0.0001). This finding was corroborated by the physician survey. Patient age and education were also associated with a higher likelihood of having been offered CRC screening (P = 0.009 and 0.014, respectively). Patient race, gender, primary language, PCP, or clinics attended were not significantly associated with the discussion of CRC screening. Conclusions. Discussions regarding CRC screening are most likely to occur during preventive care visits. Thus, facilitating preventive visits especially for the elderly represents an opportunity to improve CRC screening rates in primary care practice. D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. Keywords: Cancer screening; Prevention; Health care practice
Introduction Colorectal cancer (CRC) is the second leading cause of cancer deaths in the United States [1]. In 2002, 148,300 new cases were diagnosed and 56,600 people died of this disease. Early detection has been shown to decrease CRC mortality [2]. As a result, many organizations such as the American Cancer Society, the National Cancer Institute, the American Abbreviations: BE, barium enema; CI, confidence intervals; CRC, Colorectal cancer screening; CS, colonoscopy; FOBT, fecal occult blood test; FS, flexible sigmoidoscopy; OR, crude odds ratio; PCP, primary care physician. * Corresponding author. Department of Family & Community Medicine, College of Medicine, Scott & White Memorial Hospital, Texas A&M University System Health Science Center, 2401 South 31st Street, Temple, TX 76508. Fax: +1-254-680-1234. E-mail address: [email protected] (P. Patel).
Gastroenterological Association, and the U.S. Preventive Services Task Force endorse routine screening for CRC [3,4]. Despite these recommendations and the compelling evidence supporting screening for CRC, most persons at average risk are not screened [5,6]. The individual and societal benefits of CRC screening cannot be realized unless screening activities are appropriately promoted and completed. The delivery of preventive services such as cancer screening by primary care practitioners depends on multiple factors such as the practice setting, payer mix, insurance, and available resources. In addition, cancer screening is dependent on individual patient and physician characteristics. Patient-related factors include age, gender, education, ethnicity, health care beliefs, and socioeconomic status, whereas physician factors may include knowledge, experience, access to specialists and tests, and perceptions towards screening. Many studies have reported these and other specific
0091-7435/$ - see front matter D 2004 The Institute For Cancer Prevention and Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2004.03.026
240
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
individual factors as contributors to suboptimal screening for CRC [7 –13]. Although arbitrarily separated into patient, physician, and health care system-related factors, these multiple factors are highly interrelated [14]. Because these factors can dynamically influence each other, factors determining the delivery of CRC screening will reflect any limitations or constraints placed by the practice or health care delivery setting and vary within each practice setting. Although the dynamics of the patient –physician interaction have a powerful effect on screening practice and recommendations, the presence of multiple potentially confounding factors primarily related to practice setting has precluded an adequate evaluation of its role in reported community-based or national studies. Examples of extrinsic factors that can influence studies of the patient –physician interaction include the availability of local facilities and/or cognitive and technical skills to perform flexible sigmoidoscopies, the availability and access to gastroenterologists or radiologists, variations in managed care screening guidelines, and other reimbursement-related issues. CRC screening rates may be improved by increasing discussions about CRC screening between physicians and their patients. Lack of physician recommendation ranks as one of the top three reasons patients do not undergo CRC screening [8,15], and surveys have indicated that less than 50% of patients at risk are counseled by their doctors to undergo testing [1,16]. A recent study emphasized the importance of physician recommendations and called for more studies to examine factors influencing discussions of CRC screening by physicians [17].
Clearly, there is a need to better understand the barriers affecting the actual offering of CRC screening by primary care physicians (PCPs). Furthermore, the design of globally applicable interventions to increase discussion of CRC screening in primary care practice requires knowledge of specific factors that influence whether or not CRC screening is offered. However, there is a lack of information regarding these factors. To identify opportunities for improved screening, we sought to elucidate specific factors that could influence the offering of CRC screening during a patient – physician encounter. A conceptual model was developed and separate patient and physician surveys addressed potential influencing factors. The study was performed within the context of an integrated health care system to decrease the impact of extrinsic confounding health care system-related variables that could potentially influence whether or not CRC screening was offered during a patient – physician interaction. Thus, we expect that these results will be applicable and generally relevant to patient –physician interactions in primary care practice.
Methods Study design and setting A conceptual model was developed based on the work by Myers et al. [18] (Fig. 1). This model was used to identify potential variables that may influence the discussion of CRC
Fig. 1. Conceptual model on which the study was based. This model was used to identify potential variables that may influence the discussion of CRC screening in primary care practice.
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
screening in primary care practice. These variables were considered for further evaluation as determinants of whether or not screening was offered during a physician – patient encounter. The specific role of selected factors was evaluated by separate patient- and physician-directed surveys. The study was approved by the Scott & White Institutional Review Board. A cross-sectional telephone survey was administered to patients scheduled to see a primary care physician (PCP) in one of four family medicine clinics of the Scott & White Healthcare System between March and May 2002. All enrolled patients were members of the Scott & White Health Plan and had their primary insurance through this Health Maintenance Organization, sometimes in conjunction with secondary payers such as Medicare. The clinics are located in Killeen, Belton, and Temple (two clinics) in Bell County, Texas. Each of the clinics has the ability to perform flexible sigmoidoscopy (FS) and fecal occult blood testing (FOBT) within their facility. The physician-directed survey was administered to the physicians at each of the practice sites. All physicians were board certified in family practice and performed flexible sigmoidoscopies. Specific CRC screening guidelines had been made available to all physicians at the four clinics through lectures given by one of the investigators (AAA), along with distribution of printed health plan guidelines. Identical access for colonoscopy and radiological studies was available to all physicians at a single site (Scott & White Memorial Hospital). Patient survey All patients aged 50 years or older, who were scheduled to see a PCP at one of the four clinics and were members of the Scott & White Health Plan, were eligible for enrollment into the study. All eligible patients were given a card by a front desk attendant before their clinic visit inviting them to participate in a survey designed to study healthcare practices. The specific subject of the survey (CRC screening) was not mentioned on the card. Of those who indicated a willingness to participate, one in four was randomly selected to undergo a telephone interview with a research assistant who is bilingual in English and Spanish. The interview was completed within a week of the clinic visit. The telephone survey collected demographic data including gender, age, race, education, and the PCP of each individual. The participants were then asked about any discussion of CRC screening and about any prior screening for CRC. The terminology used for descriptions of the specific testing for CRC screening was adapted from the Behavioral Risk Factor Surveillance System, which is available from the Centers for Disease Control and Prevention [19].
241
about physician age, gender, experience (years in practice), as well as physicians’ beliefs regarding CRC screening and their perceptions regarding their actual CRC screening practice. The physicians at each site had not been informed of the patient survey, which was performed before the administration of the physician survey. Physicians were asked how often they discussed CRC screening with patients over the age of 50 during specific types of patient encounters and how often they specifically discussed flexible sigmoidoscopy (FS), fecal occult blood testing (FOBT), barium enema (BE), or colonoscopy (CS) with patients who were offered CRC screening. Response choices were ‘‘never,’’ ‘‘rarely,’’ ‘‘sometimes,’’ ‘‘often,’’ or ‘‘always.’’ Statistical analysis All data were collected by a research assistant and entered on a data collection sheet. Missing data were excluded from further analysis. For the patient survey, two primary outcome variables were considered: discussion of CRC screening at the most recent clinic visit and ever (overall visits). The former was measured by assessing whether the patient reported that at their most recent visit their PCP discussed CRC screening, did not discuss screening, or was undecided. Responses were dichotomized as ‘‘yes’’ or ‘‘no/did not remember’’ for analysis. The latter was measured by aggregating patient responses for three questions on the survey: whether the patient had ever discussed CRC screening with their PCP at any visit; whether the patient had ever discussed CRC screening with any other physician at any visit; and whether the patient discussed CRC screening with their PCP at their most recent visit. Responses were similarly dichotomized as ‘‘yes’’ or ‘‘no/did not remember’’ for analysis. These outcome variables were related with factors that were predicted to influence whether or not screening is offered based on our conceptual model. All other variables in the patient survey were informative questions focusing on the patient’s understanding of CRC screening, and more specifically on FS, FOBT, BE, and CS. Frequencies were determined for all variables. A chi-square measure of association ( P < 0.05) tested for levels of association between selected variables and the primary outcome variables. Crude odds ratios (ORs) and 95% confidence intervals (CIs) for having ever discussed CRC screening were computed to quantify degree of associations. Finally, multiple logistic regression modeling was used to estimate adjusted ORs and 95% CIs.
Results Patient survey response rates and patient characteristics
Physician survey A separate written survey was administered to all PCPs in the four clinics. This questionnaire collected information
Of 2,716 patients who were invited to participate in the study, 1,620 (58.1%) consented, of which 400 patients were randomly selected and contacted for a telephone
242
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
interview within 1 week of their clinic visit. The median age of the participants was 65 years (mean = 65.4 years; SD = 9.8), with the majority of patients (65.1%) being in the 55- to 74-year range, which is representative of the primary target group for CRC screening. Nearly two-thirds (63.8%) of the participants were women. The majority of visits were for acute or urgent care or follow-up or chronic care, which is reflective of the practice pattern at the survey clinic sites. CRC screening discussion More than half of the patients (n = 214; 53.5%) reported that they had discussed CRC screening with their PCP during a previous clinic visit. An additional 47 patients reported having discussed CRC screening with some other physician during a previous clinic visit. Thus, of the 400 patients, 139 (34.8%) reported never discussing CRC screening with their PCP or any other physician at any visit before the most recent visit. At their most recent clinic visit, 53 (13.3%) of the 400 surveyed patients reported to have discussed CRC screening with their PCP. This number included 44 (11.0%) who were discussing CRC screening again and only 9 (2.3%) who were discussing CRC screening for the first time with any physician. The characteristics of the 130 patients who indicated that they had never discussed CRC screening at any time including their most recent clinic visit were compared with those of the 270 patients who indicated that they had discussed CRC screening previously or during their most recent office visit (Table 1). Discussion of CRC screening was significantly associated with patient age, education, and belief in CRC screening, as well as the type of clinic visit. Increasing age was associated with a significantly lower likelihood of having discussed CRC screening. Patients with less education were significantly less likely to have discussed CRC screening than those with a higher education. Since health beliefs can influence colorectal cancer screening behavior [20], we assessed belief in CRC screening by asking patients if they felt that people of their age should be screened for colorectal cancer. Those respondents who did not believe in CRC screening were significantly less likely to have discussed it. Discussion of CRC screening was also significantly less likely during clinic visit for acute or urgent care and follow-up or chronic care. However, there was no significant association of CRC screening discussion with clinic attended (P = 0.56), patient gender (P = 0.49), race (P = 0.80), or primary language (P = 0.69). After multivariate adjustment, clinic visit type maintained the strongest association with the likelihood of having discussed CRC screening. Patient education was no longer significantly associated with having CRC screening discussion. The association of having discussed CRC screening with patient belief in CRC screening also became marginal. Compared to patient clinic visits for follow-up or chronic
Table 1 Comparison between patients who have previously discussed (n = 270) and those who have never discussed (n = 130) CRC screening Characteristic
Clinic Killeen Belton Santa Fe Northside Gender Male Female Age group < 55 years 55 – 64 years 65 – 74 years 75 years or older Race White Non-White Education groups GED or less Some college/bus/tech College degree/postgrad Primary language English Other Belief in CRC screening Yes No Visit type Acute care/urgent care Physicals Follow-up/chronic care
CRC screening CRC screening P value previously discussed never discussed n (%) n (%) 0.56 163 25 60 22
(66.0) (64.1) (74.1) (66.7)
84 (34.0) 14 (35.9) 21 (25.9) 11 (33.3)
101 (69.7) 169 (66.3)
44 (30.3) 86 (33.7)
0.49
0.009 48 95 84 43
(76.7) (72.5) (65.1) (54.4)
13 36 45 36
(21.3) (27.5) (34.9) (45.6) 0.80
230 (67.3) 40 (69.0)
112 (32.7) 18 (31.0)
105 (60.3) 84 (77.1) 78 (68.4)
69 (39.7) 25 (22.9) 36 (31.6)
266 (67.7) 4 (57.1)
127 (32.3) 3 (42.9)
252 (69.0) 18 (51.4)
113 (31.0) 17 (48.6)
106 (65.8) 65 (86.7) 99 (60.4)
55 (34.2) 10 (13.3) 65 (39.6)
0.014
0.69
0.034
< 0.0001
care, patients visiting the doctor’s office for physicals were four times as likely to have discussed CRC screening (adjusted OR = 4.09, 95% CI = 1.91 – 8.74). Patients under 55 years were three times (adjusted OR = 2.99, 95% CI = 1.31– 6.83) as likely to have discussed CRC screening as those aged 75 years or older (Table 2). Type of screening discussion and follow-up of screening recommendations Of the 270 patients who reported prior CRC screening discussion at any time, 210 (77.8%) had discussed FOBT, 158 (58.5%) had discussed FS, 115 (42.6%) had discussed BE, and 83 (30.9%) had discussed CS. Of note, FOBT was described to patients as being a test that may use a special kit at home to determine if the stool contains blood. Actual performance of discussed screening tests by patients varied widely. Of the 210 patients who reported to have discussed FOBT, only 88 (41.9%) had performed the test within the previous year. However, of those 158 patients who had discussed FS, 137 (86.7%) had performed the test within the past 5 years. Comparable statistics for BE within the past 5 years and CS within the past 10 years were 57.4% and 69.3%, respectively.
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
243
Table 2 Odds ratios and 95% confidence intervals for having ever discussed CRC screening Characteristic
Crude OR (95% CI)
Clinic Killeen Belton Santa Fe Northside Gender Male Female Age group < 55 years 55 – 64 years 65 – 74 years 75 years or older Race White Non-White Education groups GED or less Some college/bus/tech College degree/postgrad Primary language English Other Belief in CRC screening Yes No Visit type Acute care/urgent care Physicals Follow-up/chronic care
Adjusted P value
OR (95% CI)
0.56 0.97 (0.45 – 2.10) 0.89 (0.34 – 2.37) 1.43 (0.59 – 3.44) Reference
0.39 1.30 (0.57 – 2.96) 1.32 (0.47 – 3.68) 2.04 (0.80 – 5.19) Reference
0.49 1.17 (0.75 – 1.81) Reference
0.53 1.16 (0.73 – 1.86) Reference
0.01 3.09 (1.45 – 6.58) 2.21 (1.23 – 3.97) 1.56 (0.88 – 2.77) Reference
0.04 2.99 (1.31 – 6.83) 2.05 (1.09 – 3.87) 1.46 (0.80 – 2.68) Reference
0.80 0.92 (0.51 – 1.68) Reference
0.61 1.20 (0.60 – 2.39) Reference
0.02 0.70 (0.43 – 1.16) 1.55 (0.85 – 2.82) Reference
0.11 1.11 (0.64 – 1.94) 1.88 (0.99 – 4.57) Reference
0.56 1.57 (0.35 – 7.12) Reference
0.41
0.04
0.05 2.13 (0.99 – 4.57) Reference
0.001
Fig. 2. Frequency of screening test used amongst persons with prior CRC screening. The bars illustrate the number of survey participants who had previously undergone a CRC screening test. The shaded portion represents those persons who reported testing was for the purpose of cancer screening, whereas the nonshaded portion of the bar represents those persons who reported testing was for purposes other than cancer screening or for unknown reasons (FOBT = fecal occult blood testing).
had undergone FS, FOBT, CS, or BE testing, respectively (Fig. 2). The majority of participants indicated that they would repeat the screening tests. Interestingly, many indicated they would do so if recommended by their physician, highlighting the importance of the patient – physician interaction. Participants identified BE as the most likely screening test not to be repeated (8.6%, 10/115), followed by CS 3.6% (3/83) and FS 3.2% (5/158). As expected, the reluctance rate for FOBT (1.9%) was quite low (4/210). Those who would not repeat any screening test cited pain as the most common reason (13/21). Physician survey response rates and CRC screening discussion
2.12 (0.36 – 12.42) Reference
2.11 (1.05 – 4.24) Reference
1.27 (0.81 – 1.99) 4.26 (2.04 – 8.90) Reference
P value
Of the 36 physicians surveyed, 32 (88.9%) responded. The median age of the respondents was 43 years (mean 42; range: 30 –57 years). There were 25 males and 7 females. The median number of years of practice experience was 12 years (mean 13; range 1 –25 years). While all physicians work in a clinic setting with their own panel of patients, some of the physicians have significantly smaller practices due to administrative responsibilities. Two
0.001 1.22 (0.76 – 1.97) 4.09 (1.91 – 8.74) Reference
Reasons for screening Although many patients were unaware of the reason for testing, screening for cancer was identified as the reason given by 53.8%, 32.1%, 59.0%, and 29.6% of persons who
Table 3 Physician reported initiation of CRC screening discussion by encounter type (n = 32)a Type of visit
Never, n (%)
Rarely, n (%)
Sometimes, n (%)
Often, n (%)
Always, n (%)
Acute care
6 (18.8) 0 (0.0) 1 (3.2)
20 (62.5) 0 (0.0) 2 (6.5)
5 (15.6) 0 (0.0) 18 (58.1)
1 (3.1) 15 (48.9) 9 (29.0)
0 (0.0) 17 (53.1) 1 (3.2)
Physicals Chronic care a
May not add to total due to missing data.
244
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
Table 4 Physician reported preferences for CRC screening tests (n = 32) Test Barium enema
Never, Rarely, Sometimes, Often, Always, n (%) n (%) n (%) n (%) n (%)
0 (0.0) Colonoscopy 0 (0.0) Fecal occult blood testing 0 (0.0) Flexible sigmoidoscopy 0 (0.0)
10 (31.3) 3 (9.4) 0 (0.0) 0 (0.0)
11 (34.4) 10 (31.3) 3 (9.4) 2 (6.2)
5 (15.6) 10 (31.3) 10 (31.3) 10 (31.3)
6 (18.7) 9 (28.1) 19 (59.4) 20 (62.5)
respondents (6.3%) indicated that they would not screen patients for CRC, whereas 23 (71.9%) indicated they would and 7 did not respond to the question. Twenty-four respondents indicated their perception of the proportion of their patients who had been screened for CRC, with a median of 65% (mean, 60%; range 20 – 98%). Most physicians reported to have discussed CRC screening always or often during patient clinic visits for physicals, and rarely, if at all, during other types of visits (Table 3). FOBT and FS were more likely to be discussed always with patients than BE and CS (Table 4).
Discussion An understanding of the factors associated with whether or not discussion of CRC screening is initiated and offered during a patient – physician encounter may result in the development of interventions that could dramatically increase screening rates. This study identified age as an important patient factor associated with the likelihood of discussing CRC screening with their physician. The study also demonstrated that most discussion of CRC screening is done at appointments for physicals; a finding that was corroborated by a survey of the physicians at the various practice sites. No physician factors were identified to be associated with the likelihood of offering CRC screening to patients. Encouragingly, we found that CRC screening had been discussed at some point with 67.5% of patients at risk. However, the frequency of screening, particularly with regards to FOBT, was not always optimal. A major strength of this study is that several potentially confounding factors were eliminated in the setting of an integrated health care system. These factors include identical access to local resources such as flexible sigmoidoscopy, specialist resources such as gastroenterologists, radiologists, oncologists, or surgeons, uniform screening guidelines, and a managed care population. Each of these factors can influence whether or not CRC screening is offered during a patient encounter. Our findings are especially relevant to managed care systems particularly given the recent inclusion of CRC screening as a benchmark Health Plan Employer Data and Information Set measure, which will promote focused efforts to improve screening rates.
Our study demonstrated that the likelihood of a patient having discussed CRC screening decreased with increasing age. This contrasts with findings of recent studies of actual screening behavior that indicate a higher likelihood of screening of older patients.[16,21]. There are several possible reasons for these observations. We speculate that physicians may be more likely to offer and discuss CRC screening with younger, more healthy persons but less likely to consider it for their older patients who may have more comorbid conditions. Increasing awareness of the age at which CRC screening should begin may also have led to an increase in CRC screening discussions with the younger patients. Another possibility is that younger patients are more aware of the benefits of preventive health strategies and are taking a more active role in their health care by asking their physicians to discuss screening with them. In our study, 72% of the patients undergoing FOBT reported that they had this test done for reasons other than cancer screening or for unknown reasons. This may reflect a gap in patient understanding of the test and its implications. Alternatively, the physician may not have adequately communicated the reasons for the testing. This raises the possibility that although patients are willing to undergo testing recommended by their physician, many may have a poor understanding of the complex testing options available for CRC screening. Since complexity of issues surrounding testing has been identified as a potential barrier to discussion of other cancer screening tests [22 – 24], identifying methods to simplify such issues would likely increase screening. CRC screening was more likely to be discussed during routine physicals. This is likely due to the inherent focus on preventive health during physicals and the increased time available for such appointments. This assertion is supported by a study by Dunn et al. [25] in which lack of time was identified as one of the most common barriers to discussing breast and prostate cancer screening tests. Follow-up and chronic care visits are identified in our study as providing inadequate opportunities to address screening especially in a Medicare population for whom routine physical appointments are not reimbursed. Because multiple screening testing options are available for CRC screening, adequate patient counseling is required for shared decision making. Physician perceptions about the time required for an adequate discussion may influence whether or not screening is discussed during a specific visit. Unless preventive care visits are facilitated, for example, by appropriate reimbursement, innovative approaches to identify and achieve adequate patient counseling within the context of visits for purposes other than general health maintenance will be required. In contrast to other studies, gender did not appear to play a role in the actual offering of CRC screening. It has been proposed that women may be more likely to be offered CRC screening, as they are more likely to attend clinics for other
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
preventive screening. However, no such difference was identified in our study. Despite the different clinic sizes and location, the offering of CRC screening did not appear to be different in each of the clinics. This may represent the benefits of an integrated health care system. In this setting, we expect that interventions to improve screening are also likely to be similar at all practice sites. A potential limitation of this study is the reliance on patient recall. However, we expect that the impact of this will be minimal since patient surveys were conducted within a week of the clinic visit, and patients had previously provided consent to be contacted regarding our health care practices. Furthermore, patient surveys correlate well with chart audits and are preferable to physician surveys for cancer screening practices [22]. Nevertheless, inadequate patient recall regarding prior testing that may be known to the physician remains a limitation. Although our study was performed within the context of a specific managed care setting, we believe that these results are generalizable because of the exclusion of specific practice setting-related confounding factors, which was one of the primary aims of the study. In other settings, additional factors such as local referral practices, facilities, access, reimbursement, and payer mix may also influence discussions regarding CRC screening. Finally, the small numbers of patients in some groups such as primary language may have resulted in a reduced power to detect specific differences. Although CRC screening has been shown to reduce mortality, the full benefits of screening can only be achieved if the population is adequately screened. Improved screening of persons at average risk for CRC cancer in primary care is thus essential. This study investigated important variables associated with the likelihood of physicians discussing CRC screening with their patients. Age has been identified in other studies as a factor influencing the likelihood of having actually undergone screening. Since the discussion of screening is a vital determinant of patients’ decisions to eventually undergo screening tests, targeted interventions aimed at older and perhaps less educated patients in individual practices would be expected to improve screening rates. Both surveys also identified visit type as being a crucial factor. It is therefore important to promote visits for preventive health that allow adequate time to focus on discussion of CRC screening in addition to other recommended preventive issues. Providing appropriate reimbursement for such visits thus represents an opportunity to improve CRC screening in primary care practice.
Acknowledgments This study was supported by a grant award to Dr. Parita Patel from the Scott & White Memorial Hospital, Scott, Sherwood and Brindley Foundation. The assistance of Rinska Flores, Hope Gonzalez, and Rebecca Adams in
245
conducting the survey and of Susan Cromwell and Mark Riggs in analysis of the survey data is greatly appreciated.
References [1] Trends in screening for colorectal cancer—United States, 1997 and 1999. MMWR Morb Mortal Wkly Rep 2001;50(9):162 – 6. [2] Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA 2003;289(10):1288 – 96. [3] Walsh JM, Terdiman JP. Colorectal cancer screening: clinical applications. JAMA 2003;289(10):1297 – 302. [4] Winawer S, Fletcher R, Rex D, Bond J, Burt R, Ferrucci J, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale—Update based on new evidence. Gastroenterology 2003; 124(2):544 – 60. [5] Anderson LM, May DS. Has the use of cervical, breast, and colorectal cancer screening increased in the United States? Am J Public Health 1995; 85(6):840 – 2. [6] Wender RC. Barriers to screening for colorectal cancer. Gastrointest Endosc Clin N Am 2002;12(1):145 – 70. [7] Harewood GC, Wiersema MJ, Melton III LJ. A prospective, controlled assessment of factors influencing acceptance of screening colonoscopy. Am J Gastroenterol 2002;97(12):3186 – 94. [8] Rawl SM, Menon U, Champion VL, Foster JL, Skinner CS. Colorectal cancer screening beliefs. Focus groups with first-degree relatives. Cancer Pract 2000;8(1):32 – 7. [9] Kremers SP, Mesters I, Pladdet IE, van den BB, Stockbrugger RW. Participation in a sigmoidoscopic colorectal cancer screening program: a pilot study. Cancer Epidemiol Biomark Prev 2000; 9(10):1127 – 30. [10] Taylor ML, Anderson R. Colorectal cancer screening: physician attitudes and practices. WMJ 2002;101(5):39 – 43. [11] Cooper GS, Yuan Z, Veri L, Rimm AA, Stange KC. Colorectal carcinoma screening attitudes and practices among primary care physicians in counties at extremes of either high or low cancer case-fatality. Cancer 1999;86(9):1669 – 74. [12] Cooper GS, Fortinsky RH, Hapke R, Landefeld CS. Factors associated with the use of flexible sigmoidoscopy as a screening test for the detection of colorectal carcinoma by primary care physicians. Cancer 1998;82(8):1476 – 81. [13] Klabunde CN, Frame PS, Meadow A, Jones E, Nadel M, Vernon SW. A national survey of primary care physicians’ colorectal cancer screening recommendations and practices. Prev Med 2003;36(3):352 – 62. [14] Petersen GM. Barriers to preventive intervention. Gastroenterol Clin North Am 2002;31(4):1061 – 8 [viii]. [15] Weitzman ER, Zapka J, Estabrook B, Goins KV. Risk and reluctance: understanding impediments to colorectal cancer screening. Prev Med 2001;32(6):502 – 13. [16] Cokkinides VE, Chao A, Smith RA, Vernon SW, Thun MJ. Correlates of underutilization of colorectal cancer screening among U.S. adults, age 50 years and older. Prev Med 2003;36(1):85 – 91. [17] Taylor V, Lessler D, Mertens K, Tu SP, Hart A, Chan N, et al. Colorectal cancer screening among African Americans: the importance of physician recommendation. J Natl Med Assoc 2003; 95(9):806 – 12. [18] Myers RE, Turner B, Weinberg D, Hauck WW, Hyslop T, Brigham T, et al. Complete diagnostic evaluation in colorectal cancer screening: research design and baseline findings. Prev Med 2001; 33(4):249 – 60. [19] Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System Questionnaire. Atlanta (GA): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2002.
246
P. Patel et al. / Preventive Medicine 39 (2004) 239–246
[20] Macrae FA, Hill DJ, St John DJ, Ambikapathy A, Garner JF. Predicting colon cancer screening behavior from health beliefs. Prev Med 1984;13(1):115 – 26. [21] Walsh JM, Posner SF, Perez-Stable EJ. Colon cancer screening in the ambulatory setting. Prev Med 2002;35(3):209 – 18. [22] Montano DE, Phillips WR. Cancer screening by primary care physicians: a comparison of rates obtained from physician self-report, patient survey, and chart audit. Am J Public Health. 1995;85(6): 795 – 800.
[23] Baier M, Calonge N, Cutter G, McClatchey M, Schoentgen S, Hines S, et al. Validity of self-reported colorectal cancer screening behavior. Cancer Epidemiol Biomark Prev 2000;9(2):229 – 32. [24] Gordon NP, Hiatt RA, Lampert DI. Concordance of self-reported data and medical record audit for six cancer screening procedures. J Natl Cancer Inst 1993;85(7):566 – 70. [25] Dunn AS, Shridharani KV, Lou W, Bernstein J, Horowitz CR. Physician – patient discussions of controversial cancer screening tests. Am J Prev Med 2001;20(2):130 – 4.