Uptake of Colorectal Cancer Screening by Physicians Is Associated With Greater Uptake by Their Patients

Journal Pre-proof Uptake of Colorectal Cancer Screening by Physicians is Associated With Greater Uptake by Their Patients Owen Litwin, Jessica M. Sontrop, Eric McArthur, Jill Tinmouth, Linda Rabeneck, Christopher Vinden, Manish M. Sood, Nancy N. Baxter, Peter Tanuseputro, Blayne Welk, Amit X. Garg PII: DOI: Reference:

S0016-5085(19)41480-7 https://doi.org/10.1053/j.gastro.2019.10.027 YGAST 62972

To appear in: Gastroenterology Accepted Date: 10 October 2019 Please cite this article as: Litwin O, Sontrop JM, McArthur E, Tinmouth J, Rabeneck L, Vinden C, Sood MM, Baxter NN, Tanuseputro P, Welk B, Garg AX, Uptake of Colorectal Cancer Screening by Physicians is Associated With Greater Uptake by Their Patients, Gastroenterology (2019), doi: https:// doi.org/10.1053/j.gastro.2019.10.027. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 by the AGA Institute

Title: Uptake of Colorectal Cancer Screening by Physicians is Associated With Greater Uptake

by Their Patients Short title: Colorectal cancer screening in physicians Authors: Owen Litwin,1 Jessica M. Sontrop,1,2 Eric McArthur,2 Jill Tinmouth,2,3,4,5,6 Linda Rabeneck,2,3,4,5 Christopher Vinden,2,7 Manish M. Sood,2,8 Nancy N. Baxter,2,3,4,5,9 Peter Tanuseputro,2,8 Blayne Welk,1,2,10 Amit X. Garg1,2,11 Author affiliations: 1. Department of Epidemiology and Biostatistics, Western University, London, ON, Canada. 2. Institute for Clinical Evaluative Sciences, ON, Canada. 3. Cancer Care Ontario, Toronto, ON, Canada. 4. Department of Medicine, University of Toronto, Toronto, ON, Canada. 5. Public Health Sciences, Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada. 6. Department of Medicine, Sunnybrook Health Sciences Centre, Toronto, ON, Canada. 7. Division of General Surgery, Department of Surgery, Western University, London, ON, Canada. 8. Ottawa Hospital Research Institute, University of Ottawa, Ottawa, ON, Canada. 9. Department of Surgery, Li Ka Shing Knowledge Institute, St. Michael's Hospital, University of Toronto, Toronto, ON, Canada. 10. Division of Urology, Department of Surgery, Western University, London, ON, Canada. 11. Department of Medicine, Western University, London, ON, Canada. Funding/Support: This study was supported by the Institute for Clinical Evaluative Sciences (ICES) Western site. ICES is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Core funding for ICES Western is provided by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine and Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI). This project was conducted with members of the provincial ICES Kidney, Dialysis, and Transplantation Research Program, which receives programmatic grant funding from the Canadian Institutes of Health Research. The opinions, results and conclusions are those of the authors and are independent from the funding sources. Parts of this material are based on data and/or information compiled and provided by the Canadian Institute for Health Information (CIHI). However, the analyses, conclusions, opinions, and statements expressed in the material are those of the author(s), and not necessarily those of CIHI. No endorsement by ICES, AMOSO, SSMD, LHRI, or the MOHLTC is intended or should be inferred. Owen Litwin received funding from an Ontario Graduate Scholarship and a Canadian Institutes of Health Research Master’s award. Amit Garg is supported by the Dr. Adam Linton Chair in Kidney Health Analytics, and the Clinician Investigator Salary Award from the Canadian Institutes of Health Research. Manish M Sood receives support from the Jindal Research Chair for the Prevention of Kidney Disease. Dr. Tinmouth is the recipient of a Canadian Institutes of Health Research Embedded Clinician Researcher Award. Abbreviations used in this paper: ICES, Institute for Clinical Evaluative Sciences; OHIP, Ontario Health Insurance Plan.

1

Corresponding Author: Amit Garg, Institute for Clinical Evaluative Sciences, Western Site, Room ELL-218, Westminster Tower, London Health Sciences Centre, Victoria Hospital, 800 Commissioners Road East, London, Ontario N6A 5W9 ([email protected]). Conflicts of interest: All authors have completed the ICMJE uniform disclosure form at http://www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years, no other relationships or activities that could appear to have influenced the submitted work. Author Contributions: Owen Litwin and Jessica Sontrop drafted the manuscript. Owen Litwin and Eric McArthur conducted the statistical analysis. All authors contributed to the study concept and design, data interpretation, and revising the manuscript. Amit Garg provided administrative and financial support for the study. Acknowledgments: The College of Physicians and Surgeons of Ontario provided a list of all practicing physicians in Ontario, and ICES coordinated the encryption and linkage of this physician dataset to administrative healthcare databases housed at ICES.

Word count (abstract): 252 Word count (manuscript, excluding abstract): 5270

2

ABSTRACT

Background & Aims: Physicians’ own screening practices might affect screening in their patients. We conducted a population-based study to evaluate whether family physicians who underwent colorectal cancer testing were more likely to have patients who underwent colorectal cancer testing. Methods: We collected demographic and healthcare information on residents of Ontario, Canada from administrative databases; the sample was restricted to individuals at average risk of colorectal cancer who were 52–74 years old as of April 21, 2016. We obtained a list of all registered physicians in the province; physicians (n=11,434) were matched with non-physicians (n=45,736) on age, sex, and residential location. Uptake of colorectal tests was defined by a record of a fecal occult blood test in the past 2 years, flexible sigmoidoscopy in the past 5 years, or colonoscopy in the past 10 years. Patients were assigned to family physicians based on billing claim frequency, and then the association between colorectal testing in family physicians and their patients was examined using a modified Poisson regression model. Results: Uptake of colorectal tests by physicians and non-physicians (median age 60; 71% men) was 67.9% (95% CI, 67.0%–68.7%) and 66.6% (95% CI, 66.2%–67.1%), respectively. Physicians were less likely than non-physicians to undergo fecal occult blood testing and were more likely to undergo colonoscopy; prevalence ratios were 0.44 (95% CI, 0.42–0.47) and 1.24 (95% CI, 1.22–1.26), respectively. Uptake of colorectal tests by family physicians was associated with greater uptake by their patients (adjusted prevalence ratio, 1.10; 95% CI, 1.08– 1.12). Conclusions: Approximately one third of physicians and non-physicians are overdue for colorectal cancer screening. Patients are more likely to be tested if their family physician has been tested. There is an opportunity for physicians to increase their participation in colorectal cancer screening, which could in turn motivate their patients to undergo screening. KEY WORDS: colon cancer, early detection, physician health, health promotion

INTRODUCTION Colorectal cancer is the fourth most common cancer in high-income countries and the second most common cause of cancer deaths.1 Screening for colorectal cancer can reduce mortality2–5 and is costeffective in average-risk populations.6,7 While recommendations for screening average-risk individuals vary by country, most guidelines recommend a fecal occult blood test every 1-2 years, and some

3

recommend flexible sigmoidoscopy every five years or colonoscopy every ten years (recommendations from six organizations are summarized in Supplementary Table 1).2,8–14 In contrast to screening programs for cervical cancer, which achieve 60% to 80% participation,15–17 screening rates for colorectal cancer remain suboptimal, ranging from 50% to 65% for average-risk adults in the U.S and Canada.15,18–21 Data from the U.S. suggest that increasing colorectal cancer screening to 80% could prevent an estimated 200,000 deaths within 20 years.19 While many patient and health-system factors influence participation in colorectal cancer screening,22–25 physicians play an essential role—patients are 13 times more likely to be screened when physicians discuss screening options and recommend screening.22 Further, a physician’s own decision to be screened may influence screening in their patients.26 While most physicians believe that colorectal cancer screening is beneficial, data on physicians’ own screening comes largely from self-reported surveys,27,28 which may be limited by social desirability and recall biases. We conducted a population-based study using administrative healthcare data, and examined uptake of colorectal tests as a proxy for colorectal cancer screening in physicians and matched non-physicians. We further examined whether patients were more likely to be tested if their family physician was tested.

METHODS Study Design, Setting, and Data Sources We conducted a population-based study using administrative healthcare data in Ontario, Canada where all residents have universal access to healthcare, including tests for colorectal cancer screening. In 2016, Ontario’s population was 13.4 million, and 5.3 million were between the ages of 50 and 74.29 Demographic and healthcare information for Ontario residents is recorded in secure databases held at the Institute for Clinical Evaluative Sciences (ICES). We conducted this study according to a prespecified

4

protocol approved by the Health Sciences Research Ethics Board in London, Ontario, Canada. This report adheres to guidelines for observational studies conducted using routinely collected data (Supplementary Table 2).30

Datasets from six databases were linked using unique, encoded identifiers and analyzed at ICES. Additional information on these databases is provided in Supplementary Table 3. The Ontario Registered Persons Database was used to ascertain demographic and vital status information. Health characteristics were obtained from the Canadian Institute for Health Information–Discharge Abstract Database, the Ontario Cancer Registry, and the Ontario Health Insurance Plan (OHIP) Database. Data on fecal occult blood tests, flexible sigmoidoscopy, and colonoscopy were obtained using physician-claims codes from the OHIP Database. Physician-specific variables were ascertained from the ICES-derived Physician Database (IPDB). Variable sources and definitions are detailed in Supplementary Table 4. Data were near-complete for all variables in this study (missing data was <2%; described in table footnotes). Approximately 95% of Ontario physicians operate under the fee-for-service payment structure of the Ontario Health Insurance Plan, and the sensitivity and positive predictive value of procedure codes such as those used for colonoscopy have been shown to be high.31,32

Participants The study population was drawn from the Ontario Registered Persons Database. Since screening for colorectal cancer in average-risk individuals is recommended between the ages of 50 and 74 years (Supplementary Table 1),2,8 the sample was restricted to individuals aged 52–74 as of April 21, 2016 (the index date) to ensure that individuals had at least two years in the age-eligible screening window in accordance with guidelines. To be conservative, individuals with no healthcare records between 2009 and 2016 were excluded because they may have emigrated from the province. To approximate a cohort at average risk for colorectal cancer, we excluded higher risk individuals with evidence of the following

5

procedures and/or diagnoses any time before April 21, 2016: invasive colorectal lesions or anal cancer, inflammatory bowel disease, large bowel or colorectal resection, or colectomy (fee codes are provided in Supplementary Table 5). All physicians and matched non-physicians A list of all registered physicians as of April 21, 2016 was obtained from the College of Physicians and Surgeons of Ontario, and we were able to link 29,802 of the 37,125 physicians on this list (80.3%) to records in the Ontario Registered Persons Database (based on name and birth date). After exclusions (applied to physicians and non-physicians; shown in Figure 1), each physician was matched to four nonphysicians using a greedy-match algorithm without replacement on age (±2 years), sex, and forward sortation area (the first three digits of the residential postal code [average population size 8000]).

Family physicians and their patients As shown in Figure 2, family physicians were selected from the cohort of all eligible physicians based on specialty information in the IPDB. We then determined which individuals from the cohort of studyeligible non-physicians were patients of the selected family physicians by reviewing all billing claims of the family physicians during the 3-year period before April 21, 2016; the family physician with the most submitted claims for a particular patient was defined as that patient’s family physician. Uptake of colorectal tests Uptake of colorectal tests was defined by a record of any of the following as of April 21, 2016 (i) a fecal occult blood test in the past two years, (ii) a flexible sigmoidoscopy in the past five years, or (iii) a colonoscopy in the past 10 years (if multiple tests were completed, only one was counted for the overall composite outcome). This definition aligns with guidelines from the 2001 Canadian Task Force on Preventive Health Care,8 which were in effect during the study period (Supplementary Table 1), and with

6

prior analyses of physicians and the general Ontario population.21,27,33 The database codes used to capture these tests are provided in Supplementary Table 6. Although we attempted to approximate a cohort at average risk for colorectal cancer, because our administrative data sources lack information on family history and symptoms, we were unable to exclude all high-risk individuals including those with a family history of colon cancer. As well, it was not possible to discern whether a test was done solely for screening purposes or for another reason; colonoscopy in particular may be performed in response to patient symptoms or unexplained iron-deficiency anemia, or for diagnostic testing or as a treatment for some other condition.

Statistical Analysis Baseline characteristics of physicians and non-physicians are reported as of April 21, 2016, and differences between groups are expressed as standardized differences (differences greater than 10% are considered meaningfully different).34 Uptake of colorectal tests was compared between groups using prevalence differences and prevalence ratios. Prevalence differences and their 95% confidence intervals (CI) were derived from binomial regression models with an identity link function; prevalence ratios and their 95% CIs were derived from modified Poisson regression models. Generalized estimating equations were used to account for the correlation structure within matched sets. An interaction term was included in each model to test for differences in prevalence ratios for men and women, and by age group (52–59, 60–69, and 70–74 years) for each screening test, because previous studies have shown that colorectal cancer screening is higher in older age-eligible adults21,35 and in women vs men.35,36 Predictors of colorectal test uptake in physicians were examined using a multivariable regression model including the following prespecified covariates (as of April 21, 2016): age (52–59, 60–69, and 70–74 years),21,35 sex, 35,36 urban vs rural residence (population <10000]),35 neighbourhood-income quintile,36

7

medical specialty (11 categories),27 country of medical school graduation (Canada vs elsewhere),36 and the number of family physician visits in the 5 years before the index date (0, 1–2, 3–4, >5 visits).37 The association between uptake of colorectal tests in family physicians and uptake in their patients was examined using a modified Poisson regression model using generalized estimating equations to account for the correlation structure within family physicians. Models were adjusted for the following prespecified patient-level variables; age,21 sex,35,36 rurality,35 neighborhood-income quintile,36 and the number of family physician visits in the 5 years before the index date (categories are defined above and in Table 5). In a post-hoc exploratory analysis, we examined whether there was effect modification based on physician practice model (family health organizations; family health teams; comprehensive care models). Two-tailed p-values less than 0.05 were interpreted as statistically significant. All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, North Carolina). RESULTS A total of 11,447 physicians and 352,4725 non-physicians met the study’s inclusion criteria, and 11,434 of the 11,447 physicians (99.9%) were matched (1:4) to 45,736 non-physicians (Figure 1). Characteristics of matched physicians and non-physicians are shown in Table 1. Physicians had fewer comorbidities than non-physicians and fewer registered healthcare visits in the 5 years before the index date. Characteristics of unmatched non-physicians are shown in Supplementary Table 7 (to comply with privacy requirements at ICES, characteristics of unmatched physicians are not shown). As shown in Table 2, 70.2% of physicians graduated from a Canadian medical school and 46.3% were family physicians.

Uptake of colorectal tests in physicians and matched non-physicians Uptake of colorectal tests was 67.9% (95% CI, 67.0 to 68.7%) in physicians and 66.6% (95% CI, 66.2 to 67.1%) in matched non-physicians (prevalence difference, 1.2% [95% CI, 0.3 to 2.2%]; prevalence ratio, 1.02 [95% CI, 1.00 to 1.03]; p=0.01) (Table 3). Similar results were seen in the unmatched cohort

8

(Supplementary Table 8). Physicians were less likely than non-physicians to undergo fecal occult blood testing (prevalence ratio, 0.44 [95% CI, 0.42 to 0.47]) and were more likely to undergo colonoscopy (prevalence ratio, 1.24 [95% CI, 1.22 to 1.26], respectively. A significant interaction was found for age (Supplementary Table 9), but not sex (Supplementary Table 10). In the youngest age group (52–59), uptake of colorectal tests was higher in physicians than nonphysicians (prevalence difference, 4.7%; 95% CI, 3.3 to 6.1%), but in the oldest age group (70 to 74), uptake of colorectal tests was lower in physicians than non-physicians (prevalence difference, -4.4%; 95% CI, -7.4 to -1.5%); interaction p-value <0.001. Further, physicians were less likely to undergo fecal occult blood testing than non-physicians in all age groups, although this effect was more pronounced in older individuals (p=0.0047, Supplementary Table 13). Physicians were more likely than non-physicians to undergo colonoscopy in all age groups, but this effect was less pronounced in older individuals (p<0.001, Supplementary Table 13). There was no effect modification by age for flexible sigmoidoscopy.

Predictors of colorectal test uptake in physicians Uptake of colorectal tests was significantly higher in physicians living in urban settings (vs rural); in those living in the highest neighborhood-income quintile (vs the middle quintile); in anesthesiologists, radiologists, gastroenterologists, and general surgeons (vs family physicians); in graduates of Canadian medical schools (vs non-Canadian); and in those with at least one visit to a family physician in the 5 years before the index date (Table 4).

Association between family physician and patient colorectal test completion As shown in Table 5, patients were more likely to complete a colorectal test if their family physician had done so themselves; adjusted prevalence ratio, 1.10 (95% CI, 1.08 to 1.12) (Table 5). As well, patients were more likely to have the same specific colorectal test as their family physician—patients were more likely to complete a fecal occult blood test if their family physician completed this test (prevalence ratio,

9

1.27 (95% CI, 1.21 to 1.33), and the same pattern was seen for colonoscopy (prevalence ratio, 1.22; 95% CI, 1.20 to 1.25). Significant effect modification by physician practice model is shown in Supplementary Table 14. A more detailed breakdown of the different screening tests in family physicians and their patients is shown in Supplementary Table 12. Characteristics of patients of family physicians who did and did not complete a colorectal test are shown in Supplementary Table 11 (all standardized differences were less than 10%).

DISCUSSION Approximately one third of physicians and non-physicians in this screen-eligible cohort were overdue for colorectal cancer screening. Physicians were more likely than non-physicians to undergo colonoscopy and were less likely to undergo fecal occult blood testing. Uptake of colorectal tests in family physicians was associated with greater testing in their patients, and patients were more likely to have the same type of test as their physician.

In this Canadian study, colorectal test completion was below 70%, which is consistent with U.S. Morbidity and Mortality Weekly Report data and remains below the National Colorectal Council Roundtable goal of 80%.20,38 Despite having both universal health care and a programmatic screening program, our testing rates were lower than many US states.39 To our knowledge, no study in North America or Europe has examined uptake of colorectal tests in physicians using non-self-reported data, although an Israeli study similarly found that patients were more likely to complete colorectal testing if their family physician had done so themselves.26 While uptake was similar in physicians and nonphysicians, there was a significant interaction by age, where uptake was higher in physicians than nonphysicians in the youngest age group (52–59 years), and lower in physicians than non-physicians in the oldest age group (70–74 years). We also found significant effect modification of age for fecal occult blood testing and colonoscopy, but not for flexible sigmoidoscopy.

10

Despite Canadian guidelines recommending the fecal occult blood test as a first-line screening test,2,8,9 only 12% of physicians completed this test compared with 60% who completed a colonoscopy, which is consistent with previous reports.20,27,33 While it is possible that some physicians completed a fecal occult blood test on their own without submitting a billing claim for service payment, the effect of such unregistered testing on our estimates is likely minimal. Rates of fecal occult blood testing were low across all physician specialties, including certain types of physicians who were less likely to have access to kits in their office stock. As well, most would still submit the test for analysis and interpretation (captured in our coding algorithm), and a positive result would likely prompt a large bowel endoscopy (also captured in our coding algorithm). Physicians’ continued preference for colonoscopy may reflect a belief that this test has superior sensitivity and efficacy,33,40 despite the fact that only fecal occult blood testing and flexible sigmoidoscopy have been shown to reduce colorectal cancer mortality in randomized controlled trials.5,41–43 Other perceived advantages of colonoscopy may be its infrequent testing interval (10 years) and ability to remove pre-cancerous polyps during the procedure.25,43 Physicians’ familiarity with the healthcare system may also enable easier access to this test compared with non-physicians. In Canada, this preference for colonoscopy may have implications for healthcare system costs and colonoscopy wait times. Given our findings, further education of both physicians and the general public about the efficacy of fecal occult blood testing may be warranted. Uptake of colorectal tests in family physicians was associated with greater uptake in their patients, and patients were more likely to receive the same test their physician received. Effect modification by physician practice type was also observed in this cohort. A direct correlation between physicians’ and patients’ preventive health behaviours (i.e. the healthy doctor–healthy patient effect) has been demonstrated previously, including for vaccination, cancer screening, exercise, and smoking

11

cessation.26,44–46 Physicians are uniquely positioned to influence preventive health behaviours in their patients, and intuitively, physicians who practice healthy behaviours may be more effective at counseling their patients to do the same. If a physician feels comfortable to share they were screened, they may be more credible and motivating to their patients.47,48 Conversely, many physicians report difficulty counseling patients about behaviors they do not practice themselves.46,49 Given the clear benefits of colorectal cancer screening,19 programs that promote greater screening in physicians warrant consideration.

Our study has several strengths. We did not rely on self-reported information. Our results should have good external validity because we had access to data on all physicians in Ontario by linking administrative databases with a list of active members of the College of Physicians and Surgeons of Ontario. The results of this study should be interpreted in light of its limitations. In particular, the use of administrative healthcare data meant we were unable to exclude all high-risk individuals including those with a family history of colon cancer. As well, we were unable to discern whether a test was done for screening or for other purposes including diagnostic testing. Although individuals who completed tests for non-screening indications would still be considered up-to-date with colon cancer screening, the proportion tested will overestimate the true screening participation rate for an asymptomatic averagerisk population, which ranges from 50% to 65% in the U.S and Canada.15,18–21 Nonetheless, this does little to alter our main finding that about one third of physicians and non-physicians in this screen-eligible cohort were overdue for colorectal cancer screening. CONCLUSIONS In summary, we found that uptake of colorectal tests was similar in physicians and non-physicians, although physicians were more likely to undergo invasive testing. Uptake in family physicians was

12

associated with greater uptake in their patients. Our results highlight the opportunity for greater screening participation in physicians, who may in turn positively influence screening in patients.

13

FIGURE LEGENDS Figure 1. Sample flow: a

The sample of 12,204 physicians (drawn from 29,802 physicians registered with the College of

Physicians and Surgeons of Ontario as of April 21, 2016 and successfully linked to databases at the Institute for Clinical Evaluative Sciences) includes those aged 52–74 years as of April 21, 2016 with valid identifiers (i.e. a valid patient identifier, date of birth, sex, medical school graduation date, and at least one healthcare encounter between 2009 and 2016 [residents do not need to inform Ontario’s Ministry of Health and Long-Term Care upon emigrating from the province, and this criterion is used a proxy to exclude those who have emigrated]). b

The sample of 3,797,566 non-physicians (drawn from the Ontario Registered Persons Database)

includes all residents of Ontario aged 52–74 years as of April 21, 2016 with valid identifiers (i.e. a valid patient identifier, valid data on date of birth and sex, and at least one healthcare encounter between 2009 and 2016 [residents do not need to inform Ontario’s Ministry of Health and Long-Term Care upon emigrating from the province, and this criterion is used a proxy to exclude those who have emigrated]). c

The age restriction was applied to ensure that all persons in our study had screening tests assessed for

at least two years before the study entry date (screening is recommended to begin at age 50, so individuals who were age 52 on April 21, 2016 would be assessed for screening starting from their 50th year). d

These individuals are more likely to be symptomatic and undergoing long-term surveillance for

colorectal disease. Figure 2. Cohort build for the analysis of family physicians and their patients:

14

a

As described in Figure 1, this sample included all residents of Ontario aged 52–74 years as of April 21,

2016 with a valid patient identifier, valid data on date of birth and sex, and at least one healthcare encounter between 2009 and 2016 (residents do not need to inform Ontario’s Ministry of Health and Long-Term Care upon emigrating from the province, and this criterion is used a proxy to exclude those who have emigrated). Physicians were identified by linkage with the College of Physicians and Surgeons of Ontario (we excluded <5 whose medical school graduation occurred after April 21, 2016). The age restriction was applied to ensure that all persons in our study had screening compliance assessed for at least two years before the study entry date (screening is recommended to begin at age 50, so individuals who were age 52 on April 21, 2016 would be assessed for screening from age 50–52). b

From Figure 1: no evidence of any of the following before April 21, 2016: invasive colorectal lesions or

anal cancer, inflammatory bowel disease, or colorectal procedures (large bowel/rectal resections; colectomy). c

To link patients with their family physicians, we reviewed all billing claims of family physicians during

the 3-year period before April 21, 2016; the physician with the most submitted claims for a particular patient was defined as that patient’s family physician.

15

REFERENCES 1. Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J cancer 2015;136(5):E359-86. 2. Canadian Task Force on Preventive Care. Recommendations on screening for colorectal cancer in primary care. CMAJ 2016;188(5):340–8. 3. Lin JS, Piper MA, Perdue LA, et al. Screening for Colorectal Cancer: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA 2016;315(23):2576–94. 4. Tinmouth J, Vella ET, Baxter NN, et al. Colorectal Cancer Screening in Average Risk Populations: Evidence Summary. Can J Gastroenterol Hepatol 2016;2016:2878149. 5. Lauby-Secretan B, Vilahur N, Bianchini F, Guha N, Straif K. The IARC Perspective on Colorectal Cancer Screening. N Engl J Med 2018;378(18):1734–40. 6. Heitman SJ, Hilsden RJ, Au F, Dowden S, Manns BJ. Colorectal Cancer Screening for Average-Risk North Americans: An Economic Evaluation. PLoS Med 2010;7(11):e1000370. 7. Telford JJ, Levy AR, Sambrook JC, Zou D, Enns RA. The cost-effectiveness of screening for colorectal cancer. CMAJ 2010;182(12):1307–13. 8. Canadian Task Force on Preventive Health Care. Colorectal cancer screening. Recommendation statement from the Canadian Task Force on Preventive Health Care. CMAJ 2001;165(2):206–8. 9. Cancer Care Ontario. ColonCancerCheck (CCC) Screening Recommendations Summary—April 2016 [Internet]. 2016;Available from: https://www.cancercareontario.ca/sites/ccocancercare/files/assets/CCCScreeningRecommendati ons.pdf 10. U.S. Preventive Services Task Force. Screening for Colorectal Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med 2008;149(9):627. 11. US Preventive Services Task Force. Screening for Colorectal Cancer US Preventive Services Task Force Recommendation Statement. JAMA 2016;315(23):2564–75. 12. National Health Service. Bowel cancer screening [Internet]. 2015;Available from: https://www.nhs.uk/conditions/bowel-cancer-screening/ 13. Australian Government Department of Health. National Bowel Cancer Screening Program [Internet]. 2015;Available from: http://www.health.gov.au/internet/screening/publishing.nsf/Content/bowel-screening-1 14. National Screening Service. Bowel Screening [Internet]. 2012;Available from: http://www.screeningservice.ie/bowel-screening.html 15. Cancer Quality Council of Ontario. Cancer Screening: Breast, Cervical and Colorectal [Internet]. 2016 [cited 2016 Mar 17];Available from: http://www.csqi.on.ca/comparisons/cancer_screening/ 16. Centers for Disease Prevention and Control (CDC). Cancer Screening in the United States. MMWR 2012;61(3):41–5. 17. Major D, Armstrong D, Bryant H, et al. Recent trends in breast, cervical, and colorectal cancer screening test utilization in Canada, using self-reported data from 2008 and 2012. Curr Oncol 2015;22(4):297–302. 18. Singh H, Bernstein CN, Samadder JN, Ahmed R. Screening rates for colorectal cancer in Canada: a cross-sectional study. C Open 2015;3(2):149-157. 19. Meester RGS, Doubeni CA, Zauber AG, et al. Public health impact of achieving 80% colorectal cancer screening rates in the United States by 2018. Cancer 2015;121(13):2281–5. 20. Centers for Disease Control and Prevention (CDC). Vital signs: colorectal cancer screening test use--United States, 2012. MMWR Morb Mortal Wkly Rep 2013;62(44):881–8. 21. Cancer Care Ontario. Ontario Cancer Screening Performance Report 2016 [Internet]. 2016 [cited 2017 Jun 1];Available from: https://www.cancercareontario.ca/sites/ccocancercare/files/assets/CCOCancerScreeningPerfor

16

22. 23. 24.

25. 26. 27. 28. 29.

30.

31.

32. 33. 34. 35.

36. 37. 38.

39. 40. 41.

manceReport.pdf Mandelson MT, Curry SJ, Anderson LA, Nadel MR, Lee NC, Rutter CM LA. Colorectal cancer screening participation by older women. Am J Prev Med 19(3):149-54. Carcaise-Edinboro P, Bradley C. Influence of patient-provider communication on colorectal cancer screening. Med Care 2008;46(7):738-45. Rat C, Pogu C, Le Donné D, et al. Effect of Physician Notification Regarding Nonadherence to Colorectal Cancer Screening on Patient Participation in Fecal Immunochemical Test Cancer Screening. JAMA 2017;318(9):816. Pignone M, Miller DP. Using Outreach to Improve Colorectal Cancer Screening. JAMA 2017;318(9):799–800. Frank E, Dresner Y, Shani M, Vinker S. The association between physicians’ and patients’ preventive health practices. CMAJ 2013;185(8):649–53. Raza M, Bernstein CN, Ilnyckyj A. Canadian physicians’ choices for their own colon cancer screening. Can J Gastroenterol 2006;20(4):281–4. Frank E, Brogan DJ, Mokdad AH, Simoes EJ, Kahn HS GR. Health-Related Behaviors of Women Physicians vs Other Women in the United States. Arch Intern Med 1998;158(4):342–8. Statistics Canada. Ontario Census Profile, 2016 Census [Internet]. 2016 [cited 2017 Jun 1];Available from: http://www12.statcan.gc.ca/census-recensement/2016/dppd/prof/details/Page.cfm?Lang=E&Geo1=PR&Code1=35&Geo2=&Code2=&Data=Count&SearchT ext=Ontario&SearchType=Begins&SearchPR=01&B1=All&GeoLevel=PR&GeoCode=35 Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Med 2015;12(10):e1001885. Tinmouth J, Baxter NN, Rabeneck L, Paszat L, Sutradhar R, Saskin R. Validation of 5 key colonoscopy-related data elements from Ontario health administrative databases compared to the clinical record: A cross-sectional study. C Open 2018;6(3):E330-338. Pan J, Xin L, Ma YF, Hu LH LZ. Colonoscopy Reduces Colorectal Cancer Incidence and Mortality in Patients With Non-Malignant Findings: A Meta-Analysis. Am J Gastroenterol 2016;111(3):355-65. Zettler M, Mollon B, Da Silva V, Howe B, Speechley M, Vinden C. Family physicians’ choices of and opinions on colorectal cancer screening modalities. Can Fam Physician 2010;56(9). Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009;28(25):3083–3107. Honein-AbouHaidar GN, Baxter NN, Moineddin R, Urbach DR, Rabeneck L, Bierman AS. Trends and inequities in colorectal cancer screening participation in Ontario, Canada, 2005-2011. Cancer Epidemiol 2013;37(6):946–56. Lofters A, Ng R, Lobb R. Primary care physician characteristics associated with cancer screening: a retrospective cohort study in Ontario, Canada. Cancer Med 2015;4(2):212–23. Zarychanski R, Chen Y, Bernstein CN, Hébert PC. Frequency of colorectal cancer screening and the impact of family physicians on screening behaviour. CMAJ 2007;177(6):593–7. National Colorectal Cancer Roundtable. Working toward the shared goal of 80% screened for colorectal cancer by 2018 [Internet]. [cited 2019 Jul 1];Available from: https://nccrt.org/whatwe-do/80-percent-by-2018/ Joseph DA, King JB, Richards TB, Thomas CC, Richardson LC. Use of Colorectal Cancer Screening Tests by State. Prev Chronic Dis 2018;15:170535. Lieberman D. Screening for Colorectal Cancer in Individuals at Average Risk. JAMA Intern Med 2014;174(1):10. Lin JS, Piper MA, Perdue LA, Rutter C, Webber EM, O’Connor E, Smith N, Whitlock EP. Screening for Colorectal Cancer: A Systematic Review for the U.S. Preventive Services Task Force. Evidence

17

42. 43. 44. 45. 46. 47. 48. 49.

Synthesis No. 135. AHRQ Publication No. 14-05203-EF-1. Rockville. Holme Ø, Løberg M, Kalager M, et al. Effect of flexible sigmoidoscopy screening on colorectal cancer incidence and mortality: a randomized clinical trial. JAMA 2014;312(6):606-615. Redberg RF. Fecal Blood Testing or Colonoscopy. JAMA Intern Med 2016;176(8):1071. Frank E, Rothenberg R, Lewis C, Belodoff BF. Correlates of physicians’ prevention-related practices. Findings from the Women Physicians’ Health Study. Arch Fam Med 2000;9(4):359–67. Frank E, Segura C, Shen H, Oberg E. Predictors of Canadian physicians’ prevention counseling practices. Can J public Heal 2010;101(5):390–5. Oberg E, Frank E. Physicians’ health practices strongly influence patient health practices. J R Coll Physicians Edinb 2009;39(4):290–1. Frank E. Physician Health and Patient Care. JAMA J Am Med Assoc 2004;291(5):637–637. Frank E, Breyan J, Elon L. Physician disclosure of healthy personal behaviors improves credibility and ability to motivate. Arch Fam Med 2000;9(3):287–90. Vickers KS, Kircher KJ, Smith MD, Petersen LR, Rasmussen NH. Health behavior counseling in primary care: provider-reported rate and confidence. Fam Med 39(10):730–5.

Author names in bold designate shared co-first authorship.

18

19

Table 1. Characteristics of matched physicians and non-physiciansa,b,c,d Physicians (N = 11434) Age, years Median (IQR) 60 (56 to 65) 52–59 5181 (45.3) 60–69 5084 (44.5) 70–74 1169 (10.2) Sex Men 8056 (70.5) Women 3378 (29.5) f Urban vs rural residence Urban 10649 (93.2) Rural 785 (6.9) g Neighborhood income quintile 1 (lowest) 827 (7.2) 2 655 (5.7) 3 1077 (9.4) 4 1904 (16.7) 5 (highest) 6971 (61.0) h No. visits to a family physician Median (IQR) 3 (1 to 7) 0 2268 (19.8) 1–2 2715 (23.7) 3–4 2143 (18.7) >5 4308 (37.7) h No. visits to the emergency department Median (IQR) 0 (0 to 1) 0 8234 (72.0) 1–2 2692 (23.5) >3 370 (3.2) h No. hospital admissions Median (IQR) 0 (0 to 0) 0 10350 (90.5) >1 1084 (9.5) h Comorbidities Hypertension 2906 (25.4) Liver disease 217 (1.9) Diabetes 1151 (10.1) i Charlson comorbidity score Median (IQR) 0 (0 to 0) 0 11085 (96.9) >1 349 (3.1) th

Non-physicians (N = 45736)

Standardized e difference (%)

60 (56 to 65) 20726 (45.3) 20333 (44.5) 4677 (10.2)

0 0 0 0

32224 (70.5) 13512 (29.5)

0 0

42466 (92.8) 3270 (7.1)

1 1

3308 (7.2) 2620 (5.7) 4308 (9.4) 7616 (16.7) 27884 (61.0)

0 0 0 0 0

9 (4 to 15) 2993 (6.5) 4385 (9.6) 5053 (11.0) 33305 (72.8)

. 40 30 18 4

0 (0 to 1) 28811 (63.0) 12786 (28.0) 2669 (5.8)

. 19 5 9

0 (0 to 0) 40872 (89.4) 4864 (10.6)

. 4 4

17323 (37.9) 1731 (3.8) 7985 (17.5)

27 11 22

0 (0 to 0) 43585 (95.3) 2151 (4.7)

. 9 9

th

Abbreviations: IQR, Interquartile range (25 –75 percentile). a

All values are reported as No. (%) unless otherwise specified.

b

Physicians registered with the College of Physicians and Surgeons of Ontario as of April 21, 2016.

c

Physicians were matched to non-physicians using a 1:4 ratio based on age, sex and the first three digits of residential postal code (i.e. Forward Sortation Area [average population size 8000]), which was used to match on neighborhood characteristics and income.

d

Characteristics as of April 21, 2016 (the index date) unless otherwise specified.

1

e

Standardized differences are less sensitive to sample size than traditional hypothesis tests. They provide a measure of the difference between group means relative to the pooled standard deviation. A value greater than 10% is considered a meaningful difference.

f

Urban was defined as living in a municipality with a population >10000; missing data was imputed as urban for <0.5% of physicians and non-physicians. g

Quintiles of neighborhood income, adjusted for household size. Missing data (<0.5% of physicians and non-physicians) was imputed with the mode. h i

Assessed in the 5-year period before the index date (April 21, 2016).

Calculated using 5 years of hospitalization data preceding the index date. “No hospital admissions” received a score of 0.

2

Table 2. Additional characteristics in physiciansa No. physicians (N=11434) Years since medical school graduation Median (IQR) Years since first billing claim Median (IQR)

35 (30 to 40) 29 (24 to 34)

Medical school Canadian Non-Canadian c Specialty Anesthesiology d Radiology e Emergency medicine Gastroenterology Family medicine f Internal medicine g Oncology h Other Pathology i Pediatrics Psychiatry General Surgery j Other Surgery

8028 (70.2) 3406 (29.8) 482 (4.2) 404 (3.5) 310 (2.7) 114 (1.0) 5299 (46.3) 1389 (12.1) 173 (1.5) 70 (0.6) 281 (2.5) 542 (4.7) 1007 (8.8) 238 (2.1) 1138 (9.9) th

th

Abbreviations: IQR, Interquartile range (25 –75 percentile). a

All values are reported as No. (%) unless otherwise specified.

b

Missing values for country of medical school graduation were imputed as Canadian for 1% of physicians.

c

Missing values for specialty were imputed as family physician for 1% of physicians.

d

Included diagnostic radiologists and nuclear medicine specialists.

e

Included physicians whose specialty was classified in the ICES Physician Database as family physicians/emergency medicine.

f

Included general internal medicine and all internal medicine subspecialties, aside from gastroenterology and oncology, which are presented separately. g

Included medical oncologists, radiation oncologists, and gynecologic oncologists.

h

Included fellows, occupational medicine specialists, and community medicine/public health physicians.

i

Included general pediatrics, all pediatric medical and surgical subspecialties, and neonatal/perinatal medicine.

j

Included all surgical subspecialties (other than general surgery), including obstetrics and gynecology.

3

Table 3. Uptake of colorectal tests in matched physicians and non-physiciansa,b,c No. tested

Total no. individuals

FOBT, flexible sigmoidoscopy or colonoscopy Physicians 7762 11434 Non-physicians 30480 45736 FOBT Physicians 1326 11434

Proportion tested (prevalence) %, (95% CI)

Prevalence difference %, (95% CI)

67.9 (67.0 to 68.7) 66.6 (66.2 to 67.1)

1.2 (0.3 to 2.2) reference

1.02 (1.00 to 1.03) 1.00 [reference]

.01

11.6 (11.0 to 12.2)

-14.6 (-15.3 to -13.9)

0.44 (0.42 to 0.47)

<.001

a,d

Prevalence ratio (95% CI)

P Value

Non-physicians Flexible sigmoidoscopy Physicians Non-physicians Colonoscopy Physicians

11993

45736

26.2 (25.8 to 26.6)

reference

1.00 [reference]

132 621

11434 45736

1.2 (1.0 to 1.4) 1.4 (1.3 to 1.5)

-0.2 (-0.4 to 0.0) reference

0.85 (0.71 to 1.02) 1.00 [reference]

.09

6872

11434

60.1 (59.2 to 61.0)

11.5 (10.5 to 12.5)

1.24 (1.22 to 1.26)

<.001

Non-physicians

22213

45736

48.6 (48.1 to 49.0)

reference

1.00 [reference]

Abbreviations: CI, confidence interval; FOBT, fecal occult blood test. a

Physicians were matched to non-physicians using a 1:4 ratio based on age, sex and the first three digits of residential postal code (i.e. Forward Sortation Area [average population size 8000]), which was used to match on neighborhood characteristics and income. b

Uptake was defined by a record of the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. c

Individuals who received multiple tests during the screening period were counted only once for the composite outcome of fecal occult blood test, flexible sigmoidoscopy, or colonoscopy. d

Prevalence ratios were derived from modified Poisson regression models using generalized estimating equations to account for the correlation structure within match sets.

4

Table 4. Predictors of colorectal test uptake in physicians (N = 11 434)a,b Characteristics

No. physicians tested (prevalence) (%)

Age category 52–59 years 60–69 years 70–74 years Sex Men Women Rural vs urban residence Rural d Urban e Neighborhood income quintile 1 (lowest) 2 3 4 5 (highest) Specialty Anesthesiology Radiology Emergency medicine Gastroenterology Family medicine Internal medicine Oncology Other Pathology Pediatrics Psychiatry General surgery Other surgery Medical school

Adjusted prevalence ratio (95% CI)

c

P Value

3446 (66) 3520 (69) 804 (69)

0.96 (0.92 to 1.00) 1.00 (0.96 to 1.04) 1.00 [reference]

.06 .94

5398 (67) 2372 (70)

1.01 (0.98 to 1.03) 1.00 [reference]

.60

508 (65) 7262 (68)

0.94 (0.89 to 0.99) 1.00 [reference]

.01

553 (66) 422 (64) 695 (65) 1264 (66) 4836 (69)

1.04 (0.98 to 1.11) 0.99 (0.93 to 1.06) 1.00 [reference] 1.03 (0.98 to 1.08) 1.07 (1.02 to 1.12)

.22 .82

364 (76) 294 (73) 209 (67) 82 (72) 3562 (67) 896 (65) 119 (69) 49 (70) 191 (68) 355 (66) 722 (72) 172 (72) 755 (66)

1.17 (1.11 to 1.23) 1.09 (1.03 to 1.16) 1.01 (0.93 to 1.09) 1.14 (1.02 to 1.28) 1.00 [reference] 1.01 (0.97 to 1.05) 1.08 (0.98 to 1.19) 1.02 (0.88 to 1.19) 1.04 (0.96 to 1.13) 0.99 (0.93 to 1.05) 1.02 (0.98 to 1.07) 1.15 (1.07 to 1.24) 1.03 (0.98 to 1.07)

.30 .003 <.001 .01 .80 .02 .74 .10 .78 .30 .68 .30 <.001 .27

5

Characteristics Canadian Non–Canadian f No. family physician visits 0 1–2 3–4 >5

No. physicians tested (prevalence) (%)

Adjusted prevalence ratio (95% CI)

5722 (70) 2048 (62)

1.17 (1.14 to 1.21) 1.00 [reference]

1059 (47) 1765 (65) 1577 (74) 3369 (78)

1.00 [reference] 1.39 (1.32 to 1.47) 1.59 (1.51 to 1.67) 1.71 (1.63 to 1.79)

c

P Value <.001

<.001 <.001 <.001

a

Physicians registered with the College of Physicians and Surgeons of Ontario as of April 21, 2016.

b

Characteristics as of April 21, 2016 (the index date) unless otherwise specified.

c

A multivariable modified Poisson regression model was used and included all of the above characteristics.

d

Urban was defined as living in a municipality with a population >10 000.

e

Quintiles of neighborhood income, adjusted for household size.

f

Assessed in the 5-year period before the index date.

6

Table 5. Uptake of colorectal tests in family physicians and their patientsa,b Patients (n=1701907) No. patients tested

Total no. patients

Proportion Tested (%) (prevalence)

Yes (n=3337 [67.2%])

779396

1134909

68.7

No (n=1630 [32.8%])

352397

566998

62.2

Yes (n=594 [12.0%])

64748

175324

36.9

No (n=4373 [88.0%])

445263

1526583

29.2

Yes (n=2896 [58.3%])

493344

1000967

49.3

No (n=2071 [41.7%])

279162

700940

39.8

Family physicians (n=4967)

Prevalence ratio (95% CI) c

Unadjusted

c,d

Adjusted

b

FOBT, flexible sigmoidoscopy or colonoscopy

1.11 (1.09 to 1.12) 1.00 [reference]

1.10 (1.08 to 1.12) 1.00 [reference]

1.27 (1.21 to 1.33) 1.00 [reference]

1.27 (1.21 to 1.33) 1.00 [reference]

1.24 (1.21 to 1.27) 1.00 [reference]

1.22 (1.20 to 1.25) 1.00 [reference]

FOBT

Colonoscopy

Abbreviations: CI, confidence interval; FOBT, fecal occult blood test. a

All family physicians and patients met the study’s eligibility criteria (Figure 2); to identify patients’ family physicians, we reviewed the billing claims of family physicians in the physician cohort (Figure 1/2) during the 3-year period before April 21, 2016; the physician with the most submitted claims for a particular patient was defined as that patient’s family physician. b

Individuals with a record of any of the following as of April 21, 2016 were considered up-to-date: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. Individuals who received multiple tests during the screening period were counted only once for the composite outcome of fecal occult blood testing, flexible sigmoidoscopy, or colonoscopy. c

Data were analyzed using a modified Poisson regression model using generalized estimating equations to account for the correlation structure within family physicians.

d

In order to account for potential confounders, we adjusted for several prespecified patient-level variables: age (52–59, 60–69 and 70–74 years), sex, urban residence (urban [population >10000] vs rural [population <10000]), neighborhood-income quintile, and the number of family physician visits (0, 1–2, 3–4, >5 visits) in the 5 years before the index date (April 21, 2016).

7

Supplementary Table 1. Recommendations for colorectal cancer screening in average-risk individuals. Fecal occult blood testing Annual or biennial

Flexible sigmoidoscopy

Colonoscopy

Computed tomographic colonography

At periodic health examinations

No recommendation*

Not recommended

Country

Organization

Year

Age

2001

50 to 74

Canada

Canadian Task Force on 1,2 Preventive Health Care

2016

50 to 74

Biennial

Every 10 years

Not recommended

Not recommended

Cancer Care Ontario3

2016

50 to 74

Biennial

Every 10 years

Not recommended

Not recommended

United States

U.S. Preventive Services 4,5 Task Force

2008

50 to 75

Annual

Every 5 years

Every 10 years

No recommendation*

2016

50 to 75

Annual

Every 5 years

Every 10 years

Every 5 years

England

National Health Service

2015

60 to 74

Biennial

One time after age 55

Not recommended

Not recommended

Australia

National Bowel Cancer 7 Screening Program

2015

50 to 74

Biennial

Not recommended

Not recommended

Not recommended

Ireland

National Screening Service8

2012

60 to 69

Biennial

Not recommended

Not recommended

Not recommended

6

*Insufficient evidence to recommend the inclusion or exclusion of this screening modality for colorectal cancer. 1. Canadian Task Force on Preventive Health Care. Colorectal cancer screening. Recommendation statement from the Canadian Task Force on Preventive Health Care. CMAJ. 2001;165:206-208. 2. Canadian Task Force on Preventive Care. Recommendations on screening for colorectal cancer in primary care. CMAJ. 2016;188:340-348. 3. Cancer Care Ontario. ColonCancerCheck (CCC) Screening Recommendations Summary—April 2016. www.cancercareontario.ca/sites/ccocancercare/files/assets/CCCScreeningRecommendations.pdf. 4. U.S. Preventive Services Task Force. Screening for Colorectal Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. 2008;149:627. 5. US Preventive Services Task Force. Screening for Colorectal Cancer US Preventive Services Task Force Recommendation Statement. JAMA. 2016;315:2564-2575. 6. National Health Service. Bowel cancer screening. 2015. www.nhs.uk/conditions/bowel-cancer-screening/. 7. Australian Government Department of Health. National Bowel Cancer Screening Program. 2015. www.health.gov.au/internet/screening/publishing.nsf/Content/bowel-screening-1. 8. National Screening Service. Bowel Screening. 2012. www.screeningservice.ie/bowel-screening.htm

1

Supplementary Table 2. STROBE and RECORD checklist.

Title and abstract

Item No. 1

Introduction Background/rationale

2

Objectives

3

Methods Study design

4

Setting

5

2

STROBE items

RECORD items

Reported

(a) Indicate the study's design with a commonly used term in the title or the abstract. (b) Provide in the abstract an informative and balanced summary of what was done and what was found.

(1.1) The type of data used should be specified in the title or abstract. When possible, the name of the databases used should be included. (1.2) If applicable, the geographic region and time frame within which the study took place should be reported in the title or abstract. (1.3) If linkage between databases was conducted for the study, this should be clearly stated in the title or abstract.

Title and abstract

Explain the scientific background and rationale for the investigation being reported. State specific objectives, including any prespecified hypotheses.

Introduction

Present key elements of study design early in the paper.
 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection.

Study design

Introduction

Study design

Participants

6

(a) Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up.
 (b) For matched studies, give matching criteria and number of exposed and unexposed.

Variables

7

Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable.

Data sources/measurement

8

For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe

3

(6.1) The methods of study population selection (such as codes or algorithms used to identify subjects) should be listed in detail. If this is not possible, an explanation should be provided. (6.2) Any validation studies of the codes or algorithms used to select the population should be referenced. If validation was conducted for this study and not published elsewhere, detailed methods and results should be provided. (6.3) If the study involved linkage of databases, consider use of a flow diagram or other graphical display to demonstrate the data linkage process, including the number of individuals with linked data at each stage. (7.1) A complete list of codes and algorithms used to classify exposures, outcomes, confounders, and effect modifiers should be provided. If these cannot be reported, an explanation should be provided.

Study population

Data Sources, Appendix

Data sources

Bias

9

Study size

10

Quantitative variables

11

Statistical methods

12

Data access and cleaning methods

4

comparability of assessment methods if there is more than one group. Describe any efforts to address potential sources of bias. Explain how the study size was arrived at. Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why. (a) Describe all statistical methods, including those used to control for confounding. (b) Describe any methods used to examine subgroups and interactions. (c) Explain how missing data were addressed.
 (d) If applicable, explain how loss to follow-up was addressed. (e) Describe any sensitivity analyses. N/A

Methods, Discussion Not applicable; use of existing health records Other measures, Results

Statistical analysis, Results

(12.1) Authors should describe the extent to which the investigators had access to the database population used to create the study population. (12.2) Authors should provide information on the data cleaning methods used in the study.

Data sources, Results

Linkage

Results Participants

13

Descriptive data

14

Outcome Data

15

5

N/A

(12.3) State whether the study included personlevel, institutional-level, or other data linkage across two or more databases. The methods of linkage and methods of linkage quality evaluation should be provided.

Methods

(a) Report numbers of individuals at each stage of study--e.g. numbers potentially eligible, examined for eligibility, confirmed eligible, included 13 in the study, completing follow-up, and analyzed. (b) Give reasons for nonparticipation at each stage.
 (c) Consider use of a flow diagram. (a) Give characteristics of study participants (e.g. demographic, clinical, social) and information on exposures and potential confounders.
 (b) Indicate number of participants with missing data for each variable of interest.
 (c) Summarize follow-up time (e.g. average and total amount).
 Report numbers of outcome events or summary measures over

(13.1) Describe in detail the selection of the persons included in the study (i.e., study population selection), including filtering based on data quality, data availability, and linkage. The selection of included persons can be described in the text and/or by means of the study flow diagram

Study population, Results

Results

Results

Main results

16

Other analyses

17

Key results

18

Limitations

19

Interpretation

20

6

time.
 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (e.g. 95% confidence interval). Make clear which confounders were adjusted for and why they were included (b) Report category boundaries when continuous variables were categorized. (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period. Report other analyses done (e.g. analyses of subgroups and interactions, and sensitivity analyses).
 Summarize key results with reference to study objectives. Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and magnitude of any potential bias.

Give a cautious overall

Results

Results, Appendix

Results, Discussion (19.1) Discuss the implications of using data that were not created or collected to answer the specific research question(s). Include discussion of misclassification bias, unmeasured confounding, missing data, and changing eligibility over time, as they pertain to the study being reported.

Discussion

Discussion

Generalizability

21

Other information Funding

22

Accessibility of protocol, raw data, and programming code

7

interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence. Discuss the generalizability (external validity) of the study results.

Discussion

Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based.

N/A

Funding/Support; Financial disclosure

(22.1) Authors should provide information on how to access any supplemental information such as the study protocol, raw data, or programming code.

N/A

Supplementary Table 3. Data sources used in this study. Database abbreviation

Database name

Description

CIHI-DAD

Canadian Institute for Health Information – Discharge Abstract Database

CIHI-DAD contains summarized hospital discharge information for individuals receiving inpatient care in non-mental health designated beds. Each record contains patient identifiers, date of admission and discharge, clinical details of the hospitalization, including diagnoses made and procedures received.

IPDB

Institute for Clinical Evaluative Sciences Physician Database

The IPDB contains information on physician demographics, specialty (functional and certified), practice (e.g. billings, workload, types of services provided), and location. The IPDB contains information from the Corporate Provider Database, the Ontario Physician Human Resource Data Centre database, and the Ontario Health Insurance Plan database of physician billings.

NACRS

National Ambulatory Care Reporting System

NACRS contains information on patient visits to emergency care. Each record contains the patient identifier, date of registration, and clinical details of the visit.

OHIP

Ontario Health Insurance Plan Database

The OHIP database contains Ontario physician claims from inpatient, outpatient, and longterm care settings. Each record identifies the physician, the patient, the diagnosis responsible for the claim (which follows the ICD-9 coding scheme), the service provided, and the date on which the service was provided.

OCR

Ontario Cancer Registry

This registry tracks contains information on all incident cancer cases diagnosed in Ontario since 1964.

RPDB

Registered Persons Database

The RPDB contains demographic information such as age, sex, health insurance eligibility and death information for anyone who has received Ontario health care coverage. It contains postal code information that is linkable to other geographic data such as dissemination areas, which can be linked census-derived, neighborhood-level data.

8

Supplementary Table 4. Source and description of baseline characteristics. b

a

Variable

Source

Classification/Description

Age Sex Neighborhood income quintile

RPDB RPDB Census 2006 Census 2006 OHIP, IPDB NACRS CIHI-DAD ODB CIHIDAD NACRS OHIP

52–59, 60–69, and 70–74 years. Male; female.

Residential status Family physician visits Emergency department visits Hospital admissions Hypertension

Liver disease Diabetes

ODB OHIP

Charlson comorbidity score

CIHI-DAD

a

Databases linked at the individual-level. Defined in Table S3.

b

9

Quintiles of neighborhood income (1 = lowest income to 5 = highest income) Urban; Rural (urban = living in a municipality with a population >10,000). 0, 1-2, 3-4, >5 visits in the 5-year period before the index date. 0, 1-2, >3 visits in the 5-year period before the index date 0; >1 ICD10: "I10", "I11", "I12", "I13", "I15" OHIP dxcode: "401", "402", "403" ICD10: "B18", "B19", "R160", "R162", "B942", "Z225", "K7" OHIP dxcode: "571", "573", "070" ICD10: "E10", "E11", "E13", "E14" OHIP feecode: "Q040", "K029", "K030", "K045", "K046" OHIP dxcode: "250" 0; >1 (calculated from all DAD records in the 5-year period before the index date).

Supplementary Table 5. Codes used to define cohort exclusions. Variable Invasive colorectal lesion

Anal cancer Inflammatory bowel disease

Colorectal procedure

10

Definition ICD9: "1530", "1531", "1532", "1533", "1534", "1536", "1537", "1538", "1539", "1540", "1541" ICD10: "C180", "C182", "C183", "C184", "C185", "C186", "C187", "C188", "C189", "C19", "C20" OHIP dxcode: "153", "154" ICD9: "1542", "1543", "1548" ICD10: "C21" ICD9: "555" ICD10: "K50", "K51" OHIP dxcode: "555", "556" CCP: "5751", "5752", "5753", "5754", "5755", "5756", "576", "6024", "603", "6031", "6039", "604", "605", "6055" CCI: “1NM87DA”, “1NM87DE”, “1NM87DF”, “1NM87DN”, “1NM87DX”, “1NM87DY”, “1NM87GB”, “1NM87LA”, “1NM87PN”, “1NM87RD”, “1NM87RE”, “1NM87RN”, “1NM87TF”, “1NM87TG”, “1NM87WJ”, "1NM89", "1NM91", “1NQ87CA”, “1NQ87DA”, “1NQ87DE”, “1NQ87DF”, “1NQ87DX”, “1NQ87LA”, “1NQ87PB”, “1NQ87PF”, “1NQ87PN”, “1NQ87RD”, “1NQ87TF”, "1NQ89", "1NQ90" OHIP feecode: "S166", "S167", "S168", "S169", "S170", "S171", "S172", "S213", "S214", "S217", "S188", "Z753", "Z754", "Z755", "Z761", "Z784", "Z785", "E688", "S154"

Supplementary Table 6. Codes used to define colorectal tests. Outcome Fecal occult blood test

Dates OHIP: 1991 to 3 months ago

Flexible sigmoidoscopy

OHIP: 1991 to 3 months ago

Colonoscopy

OHIP: 1991 to 3 months ago

Codes used in study OHIP feecode: Any one of i) to v) i) L179, ii) Q152, iii) L181, iv) G004, v) Q133. OHIP feecode: Any one of i) to xi) and ABSENCE of the following 3 codes on the date the code below is present: E741 or E747 or E705. i) Z580 ii) Z555 iii) Z491 iv) Z492 v) Z493 vi) Z494 vii) Z495 viii) Z496 ix) Z497 x) Z498 xi) Z499 OHIP feecode: Any one of i) to x) i) Z555 and (E741 or E747 or E705) ii) Z491 and (E741 or E747 or E705) iii) Z492 and (E741 or E747 or E705) iv) Z493 and (E741 or E747 or E705) v) Z494 and (E741 or E747 or E705) vi) Z495 and (E741 or E747 or E705) vii) Z496 and (E741 or E747 or E705) viii) Z497 and (E741 or E747 or E705) ix) Z498 and (E741 or E747 or E705) x) Z499 and (E741 or E747 or E705)

Abbreviations: OHIP, Ontario Health Insurance Plan.

11

Supplementary Table 7. Characteristics of unmatched non-physicians.

a

Non-physicians n=3524725 Age, years Median (IQR) 61 (56–67) 52–59 1571549 (44.6) 60–69 1473353 (41.8) 70–74 479823 (13.6) Sex Men 1715032 (48.7) Women 1809693 (51.3) b Urban vs. rural residence Urban 3051069 (86.6) Rural 473656 (13.4) c Neighborhood income quintile 1 (lowest) 614197 (17.4) 2 678175 (19.2) 3 712738 (20.3) 4 753603 (21.4) 5 (highest) 766012 (21.7) d No. visits to a family doctor Median (IQR) 10 (5-18) 0 213149 (6.0) 1–2 284847 (8.1) 3–4 323495 (9.2) >5 2703234 (76.7) d No. visits to the emergency department Median (IQR) 0 (0-1) 0 2051776 (58.2) 1–2 1053370 (29.9) >3 254239 (7.2) d No. hospital admissions Median (IQR) 0 (0-0) 0 3117963 (88.5) >1 406762 (11.5) e Comorbidities Hypertension 1487224 (42.2) Liver disease 139867 (4.0) Diabetes 747084 (21.2) Charlson comorbidity scoref Median (IQR) 0 (0–0) 0 3336806 (94.7) >1 187919 (5.3) th

th

Abbreviations: IQR, Interquartile range (25 –75 percentile). a

All values are reported as No. (%) unless otherwise specified.

b

Urban was defined as living in a municipality with a population >10000; missing data (<0.1%) was imputed as urban.

c

Quintiles of neighborhood income, adjusted for household size. Missing data (<0.5%) was imputed with ‘3’.

d

Assessed in the five-year period before the index date (April 21, 2016).

e

Comorbidities were assessed using administrative codes in the 5-year period before the index date. Calculated using 5 years of hospitalization data preceding the index date. “No hospital admissions” received a score of 0.

f

12

Supplementary Table 8. Uptake of colorectal tests in unmatched physicians and non-physicians. Percentage tested (prevalence) (95% CI) FOBT or flexible sigmoidoscopy or colonoscopy

Prevalence c ratio (95% CI)

a,b

No. tested

Total no.

7770 2229720

11447 3524725

67.9% (67.0–68.7) 63.3% (63.2–63.3)

1.07 (1.05–1.10) 1.00 [reference]

<0.001

Physicians 1326 Non-physicians 1007898 Flexible sigmoidoscopy

11447 3524725

11.6% (11.0–12.2) 28.6% (28.6–28.6)

0.41 (0.38–0.43) 1.00 [reference]

<0.001

Physicians Non-physicians Colonoscopy

132 37574

11447 3524725

1.2% (1.0–1.4) 1.1% (1.1–1.1)

1.08 (0.91–1.28) 1.00 [reference]

0.37

Physicians Non-physicians

6880 1508102

11447 3524725

60.1% (59.2–61.0) 42.8% (42.7–42.8)

1.40 (1.37–1.44) 1.00 [reference]

<0.001

Physicians Non-physicians FOBT

P Value

Abbreviations: CI; confidence interval; FOBT, fecal occult blood testing. a

Uptake was defined by a record of the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. b

Individuals who received multiple tests during the screening period were counted only once for the composite outcome of fecal occult blood testing, flexible sigmoidoscopy, or colonoscopy. c

Prevalence ratios were derived from modified Poisson regression models.

13

Supplementary Table 9. Uptake of colorectal tests in matched physicians and non-physicians: Subgroup analysis by age. Physicians

Non-physicians

No. tested/ total no. (%)

No. tested/ total no. (%)

52 to 59

3444/5181 (66.5)

12803/20726 (61.8)

Prevalence difference (%) (95% CI) 4.7 (3.3 to 6.1)

60 to 69 70 to 74

3517/5084 (69.2) 801/1169 (68.5)

14269/20333 (70.2) 3408/4677 (72.9)

-1.0 (-2.4 to 0.4) -4.4 (-7.4 to -1.5)

Age group

c

Prevalence ratio (95% CI)

a,b

Interaction d P value

1.08 (1.05–1.10) 0.99 (0.97–1.01) 0.94 (0.90–0.98)

<0.001

Abbreviations: CI; confidence interval. a

Uptake was defined by a record of any the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. b

Individuals who received multiple tests during the screening period were counted only once for the composite outcome of fecal occult blood testing, flexible sigmoidoscopy, or colonoscopy.

c

Prevalence ratios were derived from modified Poisson regression models using generalized estimating equations to account for the correlation structure within matched sets.

d

The P value was calculated using an interaction term in the modified Poisson regression model.

14

Supplementary Table 10. Uptake of colorectal tests in matched physicians and non-physicians: Subgroup analysis by sex.

Sex Men Women

Physicians

Non-physicians

No. tested/ total no. (%)

No. tested/ total no. (%)

5391/8056 (66.9) 2371/3378 (70.2)

21268/32224 (66.0) 9212/13512 (68.2)

Prevalence difference (%) (95% CI) 0.9 (-0.2 to 2.1) 2.0 (0.3 to 3.7)

Prevalence ratio c (95% CI)

Interaction d P value

1.01 (1.00–1.03) 1.03 (1.00–1.06)

0.31

a,b

Abbreviations: CI; confidence interval. a

Uptake was defined by a record of any the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. b

Individuals who received multiple tests during the screening period were counted only once for the composite outcome of fecal occult blood testing, flexible sigmoidoscopy, or colonoscopy.

c

Prevalence ratios were derived from modified Poisson regression models using generalized estimating equations to account for the correlation structure within matched sets.

d

The P value was calculated using an interaction term in the modified Poisson regression model.

15

Supplementary Table 11. Characteristics of patients of tested versus non-tested family physicians. Testing status of patient’s family physician b Tested Not tested n=1134909 n=566998 Age, years Median (IQR) 61 (56-67) 52–59 491096 (43.3) 60–69 484141 (42.7) 70–74 159672 (14.1) Sex Men 549374 (48.4) Women 585535 (51.6) d Urban vs rural residence Urban 997469 (87.9) Rural 137440 (12.1) e Neighborhood income quintile 1 (lowest) 187163 (16.5) 2 214864 (18.9) 3 228045 (20.1) 4 244149 (21.5) 5 (highest) 260688 (23.0) f No. visits to a family doctor Median (IQR) 11 (6-18) 0 19204 (1.7) 1–2 69462 (6.1) 3–4 102955 (9.1) >5 943288 (83.1) f No. visits to the emergency department Median (IQR) 0 (0-1) 0 653627 (57.6) 1–2 348460 (30.7) >3 132822 (11.7) f No. hospital admissions Median (IQR) 0 (0-0) 0 998105 (87.9) >1 136804 (12.1) f Comorbidities Hypertension 504244 (44.4) Liver disease 44717 (3.9) Diabetes 249171 (22.0) g Charlson comorbidity score Median (IQR) 0 (0-0) 0 1072123 (94.5) >1 62786 (5.5) th

th

a,b

Standardized c difference (%)

61 (56-67) 249582 (44.0) 238842 (42.1) 78574 (13.9)

1 0 0

277875 (49.0) 289123 (51.0)

1 1

489083 (86.3) 77915 (13.7)

5 5

105653 (18.6) 114592 (20.2) 116308 (20.5) 118973 (21.0) 111472 (19.7)

6 3 1 1 8

12 (6-20) 11825 (2.1) 36626 (6.5) 49484 (8.7) 469063 (82.7)

3 2 1 1

0 (0-1) 318231 (56.1) 175508 (31.0) 73259 (12.9)

3 1 4

0 (0-0) 497287 (87.7) 69711 (12.3)

1 1

258946 (45.7) 23695 (4.2) 132470 (23.4)

3 2 3

0 (0-0) 534480 (94.3) 32518 (5.7)

1 1

Abbreviations: IQR, Interquartile range (25 –75 percentile). a All values are reported as No. (%) unless otherwise specified. b Physicians were categorized as tested if they had a record of any the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years. c Standardized differences are less sensitive to sample size than traditional hypothesis tests. They provide a measure of the difference between group means relative to the pooled standard deviation. A value >10% is considered meaningful. d Urban=living in a municipality with a population >10000; missing data (<0.5%) was imputed as urban. e Quintiles of neighborhood income, adjusted for household size. Missing data (<0.5%) was imputed with the mode. f Assessed in the 5-year period before the index date (April 21, 2016). g Calculated using 5 years of hospitalization data preceding the index date. “No hospital admissions” received a score of 0.

16

Supplementary Table 12. Uptake of colorectal tests in family physicians and their patients. Patients (n=1701907) a

Not tested

FOBT only

Colonoscopy only

FOBT and a colonoscopy

n=570114 (33.5%)

n=353577 (20.8%)

n=621782 (36.5%)

n=156434 (9.2%)

566 998 (33.3)

214601 (37.9%)

120636 (21.3%)

182247 (32.1%)

49514 (8.7%)

FOBT only (n=433 [8.7%])

132754 (7.8)

43883 (33.1%)

39517 (29.8%)

38191 (28.8%)

11163 (8.4%)

a

959585 (56.4)

298772 (31.1%)

183853 (19.2%)

385700 (40.2%)

91260 (9.5%)

42570 (2.5)

12858 (30.2%)

9571 (22.5%)

15644 (36.8%)

4497 (10.6%)

Patient’s family physician (n=4967) Not tested (n=1630 [32.8%])

Colonoscopy only (n=2743 [55.2%]) FOBT and colonoscopy (n=161 [3.2%]) Abbreviations: FOBT, fecal occult blood test. a

Includes flexible sigmoidoscopy.

17

Total no. patients (%)

Supplementary Table 13. Uptake of colorectal cancer tests in matched physicians and non-physicians: Subgroup analysis by type of test and agea No. tested

Total no.

Percentage tested (prevalence)

Prevalence ratiob (95% CI)

P Value

Physicians Non-physicians 60–69 years

617 5154

5181 20726

11.9% 24.9%

0.48 (0.44–0.52) 1.00 [reference]

0.0047

Physicians Non-physicians 70–74 years

563 20333

5084 5488

11.1% 27.0%

0.41 (0.38–0.44) 1.00 [reference]

Physicians 146 Non-physicians 1351 Flexible sigmoidoscopy Age group 52–59 years

1169 4677

12.5% 28.9%

0.43 (0.37–0.50) 1.00 [reference]

FOBT Age group 52–59 years

Physicians Non-physicians 60–69 years

60 218

5181 20726

1.2% 1.1%

1.10 (0.83–1.46) 1.00 [reference]

Physicians Non-physicians 70–74 years

51 312

5084 20333

1.0% 1.5%

0.65 (0.49–0.88) 1.00 [reference]

Physicians Non-physicians Colonoscopy

21 91

1169 4677

1.8% 1.9%

0.92 (0.57–1.47) 1.00 [reference]

3000 8976

5181 20726

57.9% 43.3%

1.34 (1.30–1.37) 1.00 [reference]

3164

5084

62.2%

1.18 (1.15–1.21)

Non-physicians 70–74 years

10713

20333

52.7%

1.00 [reference]

Physicians Non-physicians

708 2524

1169 4677

60.6% 54.0%

1.12 (1.06–1.18) 1.00 [reference]

0.14

Age group 52–59 years Physicians Non-physicians 60–69 years Physicians

<0.0001

Abbreviations: CI; confidence interval; FOBT, fecal occult blood testing. a

Uptake was defined by a record of the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years.

b

Prevalence ratios were derived from modified Poisson regression models.

18

Supplementary Table 14. Uptake of colorectal cancer tests in family physicians and their patients: Subgroup analysis by physician practice modela Total no.

Percentage tested (prevalence)

Prevalence ratiob (95% CI)

FP screened 209682 FP not screened 119721 Family health network

304968 190018

68.8% 63.0%

1.09 (1.06–1.12) 1.00 [reference]

FP screened 10002 FP not screened 6345 Family health organization

14967 9345

66.8% 67.9%

0.98 (0.92–1.05) 1.00 [reference]

FP screened 478923 FP not screened 166362 Comprehensive care model

680740 258165

70.4% 64.4%

1.09 (1.07–1.11) 1.00 [reference]

FP screened FP not screened Other/missing

25479 18796

37462 31691

68.0% 59.3%

1.14 (1.06–1.23) 1.00 [reference]

FP screened FP not screened

55310 41173

96772 77779

57.2% 52.9%

1.07 (1.02–1.12) 1.00 [reference]

No. tested

P Value

Family health group <0.0001

Abbreviations: CI; confidence interval; FOBT, fecal occult blood testing; FP, family physician. a

Uptake was defined by a record of the following as of April 21, 2016: (i) a fecal occult blood test in the past two years, (ii) flexible sigmoidoscopy in the past five years, or (iii) colonoscopy in the past 10 years.

b

Prevalence ratios were derived from modified Poisson regression models.

19

What you need to know: Background and context: A physician’s own decision to be screened for colorectal cancer might affect participation in screening by their patients. New findings: Approximately one-third of physicians and non-physicians are overdue for colorectal cancer screening. Patients are more likely to be tested if their family physician has been tested. Limitations: Associations are only described in this observational study without knowing whether they are causal. Impact: There is an opportunity to increase colorectal cancer screening among physicians, which could motivate their patients to undergo screening. Short summary: Approximately one-third of physicians and non-physicians are overdue for colorectal cancer screening. Patients are more likely to complete colorectal cancer tests if their family physician has done so themselves.