Targeted interventions by patient characteristics to improve fecal immunochemical test completion could reduce disparities in colorectal cancer screening and improve overall compliance with screening recommendations.
Objectives: Fecal immunochemical tests (FITs) can efficiently screen for colorectal cancer (CRC), but little is known on the timing to their completion. We investigate the time to return of a FIT following an order and describe patient characteristics associated with FIT return.
Study Design: Retrospective cohort study.
Methods: We identified 63,478 members of Kaiser Permanente Washington, aged 50 to 74 years, who received a FIT order from 2011 through 2012. Patient characteristics were ascertained through administrative and electronic health record data sources. We compared time from FIT order to return by patient characteristics using Kaplan-Meier and Cox regression methods.
Results: About half (53.7%) of members completed a FIT. Median time from order to return was 13 days (mean, 44.5 days; interquartile range, 6-42 days). There was higher completion of FITs among Asian patients (hazard ratio [HR], 1.43; 95% CI, 1.38-1.48), black patients (HR, 1.13; 95% CI, 1.08-1.19), and Hispanic patients (HR, 1.10; 95% CI, 1.04-1.16) compared with white patients; among patients with recent CRC testing (vs no testing in past 2 years; HR, 1.90; 95% CI, 1.86-1.95); and among patients with Medicare insurance (vs commercial; HR, 1.30; 95% CI, 1.24-1.37). Factors associated with decreased FIT completion included younger age (50-54 years vs 70-74 years; HR, 0.87; 95% CI, 0.82-0.92), obesity (vs normal body mass index; HR, 0.88; 95% CI, 0.86-0.91), and higher Charlson Comorbidity Index score (≥3 vs 0; HR, 0.82; 95% CI, 0.79-0.87).
Conclusions: Time to return of FIT varies by patient characteristics. We observed greater FIT completion among people of color, suggesting that racial disparities in CRC may not be due to patient completion of the test after an order is received.
Am J Manag Care. 2019;25(4):174-180Takeaway Points
The US Preventive Services Task Force (USPSTF) recommends several colorectal cancer (CRC) screening strategies for average-risk adults aged 50 to 75 years.1 These screening options received the highest grade by the USPSTF, reflecting the strength of evidence that shows high certainty of substantial benefit from CRC screening.2-4
One option is an annual high-sensitivity guaiac-based fecal occult blood test (gFOBT); another option is an annual fecal immunochemical test (FIT).1 FITs are consistently a preferred screening tool among patients at average risk for CRC.5 Compared with colonoscopy, FIT use is less invasive, is less expensive, and carries less risk of adverse events, giving it the potential to translate into increased CRC screening participation by underscreened groups and a lower overall cost to screening programs.6-9 As stool-based screening tests, FITs are an improvement over previous stool tests due to a higher sensitivity in detection of CRC and reduced burden on patients (ie, FIT requires only 1 stool test, whereas gFOBT requires 3).10-17 However, no previous research has investigated the overall completion of FIT orders in a community-based setting and assessed the time to completion for this modality.
The process of cancer screening completion consists of several steps, from identification of the eligible population to completion of the test and follow-up of testing results.18 Understanding where failures in the screening process occur (ie, uptake, longitudinal adherence, follow-up of positive tests) is critical for improving the process. Previous research has highlighted the deficits in overall compliance with CRC screening recommendations without identifying failures at specific steps in the screening process.19 In a previous study among members of Kaiser Permanente Washington (KPWA), we noted the prominent use of stool tests (either gFOBT or FIT) among first-time screenees, with 72% of study participants using a stool-based test for their first CRC screening after age 50 years.19 Further, the retrospective cohort showed that important patient characteristics, including higher body mass index (BMI) and female gender, were associated with lower uptake of any CRC screening tests when considering all screening methods combined. Other studies have shown disparities in general CRC screening for all modalities in socioeconomically disadvantaged populations.20 Such findings encompass 2 critical aspects in the cancer care continuum: (1) the administration of screening options to patients by clinicians as part of regular primary care, resulting in CRC screening order; and (2) the follow-through of screening on the part of the patient once given a screening order.21
Here we focus on understanding factors associated with time to completion of FIT orders, as this step is required for FIT use to be effective. Building on previous findings regarding the acceptability and performance of FIT use,22 we characterized patterns of return of a FIT following clinician order and identified patient characteristics associated with lower rates of FIT screening completion within an integrated healthcare delivery system.
Study Population and Setting
The Population-based Research to Optimize the Screening Process (PROSPR) consortium was created by the National Cancer Institute to allow multiple study sites to coordinate research for the improvement of screening practices for breast, cervical, and colorectal cancers in community settings.23 One aim of the consortium is to understand screening processes and potential failures and successes of the screening process to improve overall patient health. The collection of data throughout the screening process can further inform comparative effectiveness research for CRC screening.21 KPWA, a mixed-model health insurance and care delivery system in Washington state, is a member of the PROSPR consortium. KPWA data were extracted in 2016 from the Virtual Data Warehouse (VDW), which houses patient information in separate content areas: enrollment/demographics, utilization, laboratory, pharmacy, census, tumor registry, vital signs, and social history (tobacco, alcohol, sexual activity, etc).24
The study protocol received institutional review board approval through the KPWA Human Subjects Division for waiver of consent to enroll participants, link study data, and perform statistical analyses.
Identification of FIT Orders and Test Return
CRC screening guidelines at KPWA follow USPSTF recommendations as previously described.1 In 2011, KPWA clinicians replaced offering of the 3-sample SENSA gFOBT with the 1-sample FIT because of the improved diagnostic characteristics of the test, and they began to offer and distribute the FIT at routine clinical visits. Routine reminders are not part of the usual clinical workflow. Our cohort was defined as patients who received an order for a FIT between January 1, 2011, and December 31, 2012. From clinical laboratory data, we identified the date of FIT receipt with Current Procedural Terminology codes (ie, 82270, 82271, 82272, 82273, 82274) and Healthcare Common Procedure Coding System codes (ie, G01017, G0328, G0394).25,26
We excluded FIT orders due to (1) standing future orders (ie, automatic orders for future receipt of FIT) (n = 104); (2) orders that had a valid reason for being cancelled (see eAppendix [available at ajmc.com]), such as a new order for colonoscopy or flexible sigmoidoscopy, as those individuals are unlikely to complete a FIT in addition to other CRC screening (n = 2358); (3) inpatient orders (n = 44); and (4) orders included in the treatment arm of an ongoing randomized controlled trial (n = 81).27 Among the remaining orders, we selected the first order per patient as the index order for the start of follow-up. Our study was focused on average-risk adults; hence, orders from patients were excluded if they had a prior diagnosis of CRC (n = 66); full or partial colectomy, ileostomy, or proctectomy (n = 79); Crohn disease (n = 457); or ulcerative colitis (n = 496).
Patient characteristics were selected based on identified risk factors for CRC.28 We obtained patient characteristics through administrative and clinical patient records. Age at the time of the FIT order was calculated based on patient date of birth. The calculated variable of BMI, based on the most recent measurements within the past year of FIT order, was extracted by the VDW from vital signs recorded at the corresponding clinical visit. We used the Charlson Comorbidity Index (CCI) score as of the date of FIT order, as well as status for specific comorbidities of interest.29
Self-reported race/ethnicity was grouped as non-Hispanic white (reference group), non-Hispanic black, Hispanic (all races), non-Hispanic Asian, non-Hispanic multiracial, and non-Hispanic unknown/missing race. Gender was binary as male (reference group) and female. Age at FIT order was divided into 5-year categories (50-54, 55-59, 60-64, 65-69, and 70-74 years), with the oldest as the reference category. Insurance status was grouped according to Medicare, commercial (reference group), and other payment options (including private pay, self-pay, and Medicaid). The most recent BMI measurement (within the past year) at the time of the first FIT order was categorized as underweight (below 18.5 kg/m2), normal (18.5-24.9 kg/m2) (reference category), overweight (25.0-29.9 kg/m2), and obese (30.0 kg/m2 and above).30 The CCI score31 was calculated with the lowest score (0) as the reference group and compared with groups of higher scores (1, 2, and 3 or more). Additional analyses were performed for 3 comorbidities with the highest prevalence in the sample population: diabetes, chronic pulmonary disease, and history of myocardial infarction. Receipt of any CRC testing (eg, FIT, colonoscopy) within the 2 years prior to the index date was recorded as a binary variable.
We evaluated the association of demographic characteristics, CCI score, and medical conditions with the return of a FIT following clinician order. We performed time-to-event analysis from the date recorded of first FIT order until the date recorded for a received FIT. Individuals were censored for death, for disenrollment from KPWA, and at 365 days from the date of the first FIT order. We used 365 days as the end point because FITs are recommended as an annual screening tool; thus, patients could expect another order at the time of their annual visit. Time to completion was described using nonparametric Kaplan-Meier estimates, and differences between groups within variables were initially estimated by log-rank tests, with graphical display through 50 days from first FIT order. A univariate analysis was used to describe the mean, median, and interquartile range of time to completion of the FIT among those with a returned order. Among all patients in our population, Cox proportional hazard models were used to estimate unadjusted and adjusted hazard ratios (HRs) and 95% CIs comparing the time from order to receipt of FIT by patient-level factors with the variable-specific referent category. The adjusted model included all variables, without the specific comorbidities included in the CCI. Indicator variables were used for all variables in the models, which included an indicator for missing data where applicable. To estimate the adjusted HR for specific comorbidities, CCI score was removed from the model and replaced by each of the comorbidities in separate models. We used Breslow’s method for ties.32
Completion of FIT order, rather than incompletion, was the outcome of interest in this analysis to maintain consistency of interpretation with the results of the time-to-event analysis. Therefore, an HR greater than 1 indicates better FIT completion, a clinically positive outcome, and conversely, an HR less than 1 indicates a higher risk of the adverse outcome (ie, incomplete FIT).
All analyses were performed using SAS version 9.4 (SAS Institute; Cary, North Carolina). Additional analyses of each model were conducted in which categorical variables for age, BMI, and CCI scores were replaced with the corresponding continuous variable and tested for trends.
We identified 63,478 men and women aged 50 to 74 years who were members of KPWA in 2011-2012 and received a FIT order for inclusion in our final study population. Among those with a FIT order, about half (54%) of the sample population returned a FIT to the laboratory within 1 year of the order date. In our sample population, nearly 60% of FIT order recipients were female and approximately 75% were white (Table 1). There was an even distribution of subjects in each age group. Among those with a recorded BMI (71.0%), nearly 40% were obese. Approximately 25% of our patients had a CCI score of 1 or greater.
The proportion of patients who returned a completed FIT varied by patient race/ethnicity, age, and BMI (Table 1). Among patients who completed a FIT, the median return time of the FIT was 13 days. There was little variation in the median time to completion by patient characteristics (Table 2), with a few exceptions. On average, members of Asian race/ethnicity completed a FIT 4 days earlier than white members. Older adults (70-74 years) returned a FIT 4 days earlier than those in the youngest categories (50-54 and 55-59 years). The substantial difference between the mean and median across groups, along with the difference between the median and third quartile, suggests a heavily right-skewed distribution of time to return (Table 2). The Kaplan-Meier curves display the time-to-completion experience comparing member groups, namely age groups, race/ethnicity, BMI, and CCI score (Figure).
In multivariate adjusted models (Table 3), patient factors statistically significantly associated with decrease in FIT completion included being a woman, younger age compared with oldest age (70-74 years), being overweight, CCI score (≥1), and, more specifically, diagnosis with diabetes, chronic pulmonary disease, and myocardial infarction. Compared with white race/ethnicity, Asian, black, and Hispanic race/ethnicity were statistically significantly associated with improvements in FIT completion (Asian: adjusted HR, 1.43; 95% CI, 1.38-1.48; black: adjusted HR, 1.13; 95% CI, 1.07-1.19; Hispanic: adjusted HR, 1.10; 95% CI, 1.04-1.16).
A sensitivity analysis in which time was censored at 6 months, rather than 1 year, showed near-identical results, with very small attenuations of HRs (data not shown).
In our study, adults with the intention to return a FIT kit did so within 2 to 3 weeks from order date; otherwise, the FIT kit was more likely to not be returned. These results provide important evidence on the timing of return, which has not been previously reported and can support outreach interventions to improve FIT return. Further, we identify some differences by patient characteristics that indicate which patients might need additional support for stool-based CRC screening.
Our observation that women who receive a FIT order are less likely to return a completed FIT compared with men is consistent with previous study findings, suggesting gender-specific factors that influence completion of CRC screening, such as prior breast cancer screening.33 Women have been shown to be less likely to participate in CRC screening, possibly due to the additional need to participate in screening programs for breast and cervical cancers.24,34
We saw a consistent linear trend with improved rates of FIT return with increasing age. Interventions aimed at improving initiation of CRC screening through referrals, tracking of patient outcomes, and mitigating patient barriers have been shown to be effective at increasing CRC screening among adults younger than 65 years and may help to attenuate the observed differences in FIT completion by age.35 Being insured by Medicare remained significantly associated with return in the adjusted model accounting for age, which is appropriate with increasing risk of CRC.
Across all variables, missing data were associated with slow return and overall incompletion of FITs. The attenuated HR in the adjusted models compared with that in the unadjusted models was likely a result of correlated missingness across variables. Although there are a variety of reasons for which members could have missing capture of data variables, the most probable case is that it reflects overall healthcare engagement. Thus, the observation that those patients who are missing data for a given variable less commonly returned their FIT or did so over more time is not surprising. For cases in which missing data are a reflection of lower healthcare engagement, additional support to understand and overcome barriers for such patients may prove to be challenging but could result in the greatest return on investment.
Higher BMI has been associated with less engagement with healthy behaviors, particularly in women,36 and it is associated with reduced participation in CRC screening compared with patients with normal BMI.35,37 CRC is particularly relevant to overweight and obese adults, as increased weight is a strong risk factor for CRC.38 In a subanalysis using a Cox proportional hazards model, we did see a marginally significant interaction (P = .04) between gender and obesity on time to FIT completion (results not shown). The presence of comorbidities overall was associated with a slight reduction in completion of a FIT. We did not specifically evaluate interactions between patient characteristics, but important subgroups, such as overweight women with diabetes, could be potential target populations for screening follow-up.
Our findings regarding differences by race/ethnicity contrast with those of previous studies, which have suggested lower rates of CRC screening among people of color and lower completion of FITs, specifically for black and Hispanic individuals.39,40 Burnett-Hartman et al evaluated in stratified analyses the entire PROSPR network, indicating disparities in completion by race39; our results, however, support results in stratified analysis suggesting that Asian/Pacific Islanders are the earliest both to complete a FIT and to complete CRC screening overall. Results by other racial groups are similar to the stratified analysis but not for the pooled analysis reflective of larger samples in other systems. In our population, non-Hispanic black patients were more likely to complete FIT screening than non-Hispanic white patients. This is contrary to findings of previous studies, which have stated that a considerable proportion of the disparities in overall CRC survival between black and white patients might be attributable to differences in screening.41 However, our results are consistent with those of a recent randomized clinical trial that found that nonwhite participants were more likely to adhere to gFOBT than white participants, whereas white participants adhered more often to colonoscopy.42 As use of colonoscopy is generally more common than gFOBT/FIT for CRC screening in fee-for-service systems, this could explain part of the overall disparity in CRC screening adherence among minority populations.
It is important to note that completing a FIT is only 1 step in the CRC screening continuum, and a positive FIT requires a follow-up colonoscopy, which might incur out-of-pocket costs, even for insured patients.43 The potential for disparities in such follow-up merits further investigation. A recent study conducted at KPWA using mailed FITs and support over the phone was shown to double the number of adults who were currently compliant with CRC screening recommendations.27 Mailed outreach was effective for improving rates of CRC screening among underserved populations and had a markedly higher effect on screening with FITs compared with invitation for colonoscopy.44 Results from another study suggested that improved CRC screening will most likely be achieved through optimizing the time during current primary care visits rather than through outreach to encourage patients to attend primary care visits.42,45
Strengths and Limitations
Our study has several strengths in methodology, including a sufficient sample size to evaluate patient characteristics, data systems to capture return of FIT kits with a contemporary sample, and ability to ascertain patient covariates as confounders in the analysis. However, there are some limitations because of our study population that may limit the generalizability of our findings. It is important to recognize that ours is an insured population. Our results might not reflect the experience of uninsured or underinsured adults who may not have access to consistent healthcare. As a screening test, a FIT is a covered service under the Affordable Care Act; thus, included members should not have experienced any direct expense associated with completing a FIT during the study period.1,46 Despite federally mandated coverage, nonparticipation in CRC screening has been shown to be associated with concern for out-of-pocket costs among insured people with low socioeconomic status or in racial minority populations.43 However, for those who are uninsured, FIT use may be the most economically feasible method of receiving CRC screening. We primarily investigated patient characteristics for which previous research has asserted disparities in screening compliance. The VDW is limited in the ability to explore the reasons behind the observed patterns, but this research provides an important groundwork for future design for interventions to improve overall completion.
To our knowledge, this is the first study to assess screening completion based on return of FITs following clinician order. This research is important to reach the Healthy People 2020 goal of 70.5% completion of CRC screening and is a valuable contribution to the existing knowledge on CRC screening to reduce underuse.37 We have demonstrated that among adults eligible for CRC screening, the majority of those who complete the test do so within 2 weeks of the order. Of note, we did observe higher rates of FIT completion by nonwhite race/ethnicity. Targeted interventions, beyond mailed kits, and clinic workflows to improve return of FITs should be investigated as potential means to increase overall return rates and address disparities by patient characteristics such as obesity and age.
This study’s contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute, the National Institutes of Health, or the Agency for Healthcare Research and Quality.Author Affiliations: Kaiser Permanente Washington Health Research Institute (CBH, JC, KJW), Seattle, WA; Department of Epidemiology, University of Washington (CBH, AIP, AH, JC), Seattle, WA; Fred Hutchinson Cancer Research Center (CBH, AIP), Seattle, WA.
Source of Funding: Research in this publication was supported by the National Cancer Institute of the National Institutes of Health under award number U54 CA163261 and award number T32 CA094880 (“Cancer Prevention Training: Epidemiology, Nutrition, Genetics & Survivorship”) to Mr Haas.
Author Disclosures: Drs Chubak and Wernli are employees of Kaiser Permanente Washington, have received several grants on colorectal cancer, and present on colorectal cancer research at meetings and conferences. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (CBH, AH, KJW); acquisition of data (JC, KJW); analysis and interpretation of data (CBH, AIP, JC, KJW); drafting of the manuscript (CBH, AIP, KJW); critical revision of the manuscript for important intellectual content (CBH, AIP, AH, JC, KJW); statistical analysis (CBH); provision of patients or study materials (KJW); obtaining funding (JC, KJW); administrative, technical, or logistic support (KJW); and supervision (AH, KJW).
Address Correspondence to: Cameron B. Haas, MPH, Kaiser Permanente Washington Health Research Institute, 1730 Minor Ave, Ste 1600, Seattle, WA 98101. Email: email@example.com.REFERENCES
1. Bibbins-Domingo K, Grossman DC, Curry SJ, et al; US Preventive Services Task Force. Screening for colorectal cancer: US Preventive Services Task Force recommendation statement [erratum in JAMA. 2016;316(5):545. doi: 10.1001/jama.2016.9943] [erratum in JAMA. 2017;317(21):2239. doi: 10.1001/jama.2017.5918]. JAMA. 2016;315(23):2564-2575. doi: 10.1001/jama.2016.5989.
2. Church TR. Screening for colorectal cancer—which strategy is the best? J Natl Cancer Inst. 2011;103(17):1282-1283. doi: 10.1093/jnci/djr300.
3. Knudsen AB, Zauber AG, Rutter CM, et al. Estimation of benefits, burden, and harms of colorectal cancer screening strategies: modeling study for the US Preventive Services Task Force. JAMA. 2016;315(23):2595-2609. doi: 10.1001/jama.2016.6828.
4. Levin B, Lieberman DA, McFarland B, et al; American Cancer Society Colorectal Cancer Advisory Group; US Multi-Society Task Force; American College of Radiology Colon Cancer Committee. Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology. 2008;134(5):1570-1595. doi: 10.1053/j.gastro.2008.02.002.
5. Levin TR, Jamieson L, Burley DA, Reyes J, Oehrli M, Caldwell C. Organized colorectal cancer screening in integrated health care systems. Epidemiol Rev. 2011;33:101-110. doi: 10.1093/epirev/mxr007.
6. Allison JE, Fraser CG, Halloran SP, Young GP. Comparing fecal immunochemical tests: improved standardization is needed. Gastroenterology. 2012;142(3):422-424. doi: 10.1053/j.gastro.2012.01.015.
7. Haug U, Brenner H. A simulation model for colorectal cancer screening: potential of stool tests with various performance characteristics compared with screening colonoscopy. Cancer Epidemiol Biomarkers Prev. 2005;14(2):422-428. doi: 10.1158/1055-9965.EPI-04-0411.
8. Haug U, Brenner H. New stool tests for colorectal cancer screening: a systematic review focusing on performance characteristics and practicalness. Int J Cancer. 2005;117(2):169-176. doi: 10.1002/ijc.21016.
9. 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 [erratum in JAMA. 2016;316(5):545. doi: 10.1001/jama.2016.9941] [erratum in JAMA. 2016;316(13):1412. doi: 10.1001/jama.2016.13819]. JAMA. 2016;315(23):2576-2594. doi: 10.1001/jama.2016.3332.
10. Vasilyev S, Smirnova E, Popov D, et al. A new-generation fecal immunochemical test (FIT) is superior to quaiac-based test in detecting colorectal neoplasia among colonoscopy referral patients. Anticancer Res. 2015;35(5):2873-2880.
11. Dancourt V, Lejeune C, Lepage C, Gailliard MC, Meny B, Faivre J. Immunochemical faecal occult blood tests are superior to guaiac-based tests for the detection of colorectal neoplasms. Eur J Cancer. 2008;44(15):2254-2258. doi: 10.1016/j.ejca.2008.06.041.
12. Hol L, van Leerdam ME, van Ballegooijen M, et al. Screening for colorectal cancer: randomised trial comparing guaiac-based and immunochemical faecal occult blood testing and flexible sigmoidoscopy. Gut. 2010;59(1):62-68. doi: 10.1136/gut.2009.177089.
13. Huang Y, Li Q, Ge W, Cai S, Zhang S, Zheng S. Predictive power of quantitative and qualitative fecal immunochemical tests for hemoglobin in population screening for colorectal neoplasm. Eur J Cancer Prev. 2014;23(1):27-34. doi: 10.1097/CEJ.0b013e328364f229.
14. Imperiale TF. Noninvasive screening tests for colorectal cancer. Dig Dis. 2012;30(suppl 2):16-26. doi: 10.1159/000341884.
15. Imperiale TF, Ransohoff DF, Itzkowitz SH, Turnbull BA, Ross ME; Colorectal Cancer Study Group. Fecal DNA versus fecal occult blood for colorectal-cancer screening in an average-risk population. N Engl J Med. 2004;351(26):2704-2714. doi: 10.1056/NEJMoa033403.
16. Lejeune C, Le Gleut K, Cottet V, et al. The cost-effectiveness of immunochemical tests for colorectal cancer screening. Dig Liver Dis. 2014;46(1):76-81. doi: 10.1016/j.dld.2013.07.018.
17. Young GP, Fraser CG, Halloran SP, Cole S. Guaiac based faecal occult blood testing for colorectal cancer screening: an obsolete strategy? Gut. 2012;61(7):959-960. doi: 10.1136/gutjnl-2011-301810.
18. Anhang Price R, Zapka J, Edwards H, Taplin SH. Organizational factors and the cancer screening process. J Natl Cancer Inst Monogr. 2010;2010(40):38-57. doi: 10.1093/jncimonographs/lgq008.
19. Wernli KJ, Hubbard RA, Johnson E, et al. Patterns of colorectal cancer screening uptake in newly eligible men and women. Cancer Epidemiol Biomarkers Prev. 2014;23(7):1230-1237. doi: 10.1158/1055-9965.EPI-13-1360.
20. Smith RA, Manassaram-Baptiste D, Brooks D, et al. Cancer screening in the United States, 2015: a review of current American cancer society guidelines and current issues in cancer screening. CA Cancer J Clin. 2015;65(1):30-54. doi: 10.3322/caac.21261.
21. Tiro JA, Kamineni A, Levin TR, et al. The colorectal cancer screening process in community settings: a conceptual model for the Population-Based Research Optimizing Screening through Personalized Regimens consortium. Cancer Epidemiol Biomarkers Prev. 2014;23(7):1147-1158. doi: 10.1158/1055-9965.EPI-13-1217.
22. Chubak J, Bogart A, Fuller S, Laing SS, Green BB. Uptake and positive predictive value of fecal occult blood tests: a randomized controlled trial. Prev Med. 2013;57(5):671-678. doi: 10.1016/j.ypmed.2013.08.032.
23. Population-based Research to Optimize the Screening Process (PROSPR). National Cancer Institute website. healthcaredelivery.cancer.gov/prospr. Updated January 10, 2019. Accessed February 13, 2018.
24. Ross TR, Ng D, Brown JS, et al. The HMO Research Network Virtual Data Warehouse: a public data model to support collaboration. EGEMS (Wash DC). 2014;2(1):1049. doi: 10.13063/2327-9214.1049.
25. Beebe M. CPT 2009: Current Procedural Terminology. Chicago, IL: American Medical Association; 2008.
26. HCPCS 2009: Medicare’s National Level II Codes. Chicago, IL: American Medical Association; 2008.
27. Green BB, Wang CY, Anderson ML, et al. An automated intervention with stepped increases in support to increase uptake of colorectal cancer screening: a randomized trial. Ann Intern Med. 2013;158(5, pt 1):301-311. doi: 10.7326/0003-4819-158-5-201303050-00002.
28. Hornbrook MC, Hart G, Ellis JL, et al. Building a virtual cancer research organization. J Natl Cancer Inst Monogr. 2005;(35):12-25. doi: 10.1093/jncimonographs/lgi033.
29. Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol. 2000;53(12):1258-1267. doi: 10.1016/S0895-4356(00)00256-0.
30. BMI classification. World Health Organization website. apps.who.int/bmi/index.jsp?introPage=intro_3.html. Published May 16, 2016. Accessed February 13, 2018.
31. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.
32. Breslow N. Covariance analysis of censored survival data. Biometrics. 1974;30(1):89-99. doi: 10.2307/2529620.
33. Farraye FA, Wong M, Hurwitz S, et al. Barriers to endoscopic colorectal cancer screening: are women different from men? Am J Gastroenterol. 2004;99(2):341-349.
34. Friedemann-Sánchez G, Griffin JM, Partin MR. Gender differences in colorectal cancer screening barriers and information needs. Health Expect. 2007;10(2):148-160. doi: 10.1111/j.1369-7625.2006.00430.x.
35. Fedewa SA, Cullati S, Bouchardy C, et al. Colorectal cancer screening in Switzerland: cross-sectional trends (2007-2012) in socioeconomic disparities. PLoS One. 2015;10(7):e0131205. doi: 10.1371/journal.pone.0131205.
36. Persky S, de Heer HD, McBride CM, Reid RJ. The role of weight, race, and health care experiences in care use among young men and women. Obesity (Silver Spring). 2014;22(4):1194-1200. doi: 10.1002/oby.20677.
37. Fedewa SA, Ma J, Sauer AG, et al. How many individuals will need to be screened to increase colorectal cancer screening prevalence to 80% by 2018? Cancer. 2015;121(23):4258-4265. doi: 10.1002/cncr.29659.
38. Johnson CM, Wei C, Ensor JE, et al. Meta-analyses of colorectal cancer risk factors. Cancer Causes Control. 2013;24(6):1207-1222. doi: 10.1007/s10552-013-0201-5.
39. Burnett-Hartman AN, Mehta SJ, Zheng Y, et al; PROSPR Consortium. Racial/ethnic disparities in colorectal cancer screening across healthcare systems. Am J Prev Med. 2016;51(4):e107-e115. doi: 10.1016/j.amepre.2016.02.025.
40. Fedewa SA, Corley DA, Jensen CD, et al. Colorectal cancer screening initiation after age 50 years in an organized program. Am J Prev Med. 2017;53(3):335-344. doi: 10.1016/j.amepre.2017.02.018.
41. Lansdorp-Vogelaar I, Kuntz KM, Knudsen AB, van Ballegooijen M, Zauber AG, Jemal A. Contribution of screening and survival differences to racial disparities in colorectal cancer rates. Cancer Epidemiol Biomarkers Prev. 2012;21(5):728-736. doi: 10.1158/1055-9965.EPI-12-0023.
42. Inadomi JM, Vijan S, Janz NK, et al. Adherence to colorectal cancer screening: a randomized clinical trial of competing strategies. Arch Intern Med. 2012;172(7):575-582. doi: 10.1001/archinternmed.2012.332.
43. Green BB, Bogart A, Chubak J, et al. Nonparticipation in a population-based trial to increase colorectal cancer screening. Am J Prev Med. 2012;42(4):390-397. doi: 10.1016/j.amepre.2011.11.014.
44. Gupta S, Halm EA, Rockey DC, et al. Comparative effectiveness of fecal immunochemical test outreach, colonoscopy outreach, and usual care for boosting colorectal cancer screening among the underserved: a randomized clinical trial. JAMA Intern Med. 2013;173(18):1725-1732. doi: 10.1001/jamainternmed.2013.9294.
45. Fenton JJ, Reid RJ, Baldwin LM, Elmore JG, Buist DS, Franks P. Influence of primary care use on population delivery of colorectal cancer screening. Cancer Epidemiol Biomarkers Prev. 2009;18(2):640-645. doi: 10.1158/1055-9965.EPI-08-0765.
46. Seiler N, Malcarney MB, Horton K, Dafflitto S. Coverage of clinical preventive services under the Affordable Care Act: from law to access. Public Health Rep. 2014;129(6):526-532. doi: 10.1177/003335491412900611.