Abstract
Introduction. Colorectal cancer is the second leading cause of cancer death in the United States. Black Americans suffer even higher incidence and death rates than the general population. Genetics and patient perceptions explain some of this difference, however, modifiable health care system factors such as lack of access to colon cancer screening also contribute. Partnering an academic health center with local community groups, we piloted a colorectal cancer screening program at a Federally Qualified Health Center (FQHC) serving predominately low socioeconomic status Black Americans. The program was designed to identify and remove barriers to screening and improve screening rates. Method. At a single center FQHC, we developed an outreach program centered around (1) patient and provider education, (2) immunochemical fecal occult blood test (iFOBT) distribution, and (3) patient navigation. We identified 402 eligible patients, of which 228 (56.7%) completed screening. Results. Our 56.7% screening rate represented a twofold increase above prepilot levels at the clinic. Nine (4%) iFOBT returned positive. Three of these nine patients completed colonoscopy. Screening rates and follow through were higher under a single navigator model. Conclusions. Our academic–community partnership provided an effective, evidence based, and sustainable model for increasing colorectal cancer screening in a high risk, low resource community.
Keywords
Introduction
For average risk individuals, the U.S. Preventive Services Task Force and American Cancer Society (ACS) recommends beginning colorectal cancer (CRC) screening at age 50, using fecal occult blood test yearly, sigmoidoscopy every 5 years, or colonoscopy every 10 years (Lin et al., 2016). Early detection through screening significantly improves the likelihood of survival. Five year survival rates are 90% for well localized, 71% for regional, and 13% for metastatic disease (National Cancer Institute, 2016). Despite these well-known benefits of screening, historically, less than half of eligible U.S. adults obtained age and risk appropriate screening (Meissner, Breen, Klabunde, & Vernon, 2006; Smith, Cokkinides, & Eyre, 2007), and CRC remains the second leading cause of cancer deaths in the United States, with an estimated 134,490 new cases and 49,190 deaths in 2016 (ACS, 2016; National Cancer Institute, 2016). From 2000 to 2013, however, overall CRC screening rates increased by 24.6% reaching 58.2% in 2013 (Sabatino, White, Thompson, & Klabunde, 2015). Subsequently, overall U.S. population projections now estimate reductions in both the annual incidence (−3.1%) and mortality (−2.8%) of CRC over the next 10 years (Joseph et al., 2012).
Unfortunately, ethnic minorities, uninsured, and lower socioeconomic status (SES) peoples are not projected to experience these same benefits (Sabatino et al., 2015). Black Americans in particular continue to suffer higher CRC incidence and mortality than the general population (LoConte et al., 2011; National Cancer Institute, 2016), and evidence suggests that the disparity in CRC screening and mortality between Black and White Americans is actually increasing (Klabunde et al., 2011; Meissner et al., 2006; Robbins, Siegel, & Jemal, 2012; Soneji, Iyer, Armstrong, & Asch, 2010). Higher rates of obesity, physical inactivity, and smoking among Black Americans are thought to account for some of this disparity (Wei et al., 2004); however, a 2012 model by Lansdorp-Vogelaar et al. (2012) estimated that “more than 40% of disparity in CRC incidence and approximately 20% of disparity in CRC mortality between blacks and whites can be explained by differences in screening uptake.” Proposed reasons for low CRC screening rates among Black Americans include patient-centered barriers, including fatalistic views of cancer, more medical comorbidities, lower perceived risk, poor knowledge of screening options, and fear of diagnosis (May, Whitman, Varlyguina, Bromley, & Spiegel, 2015; Sly, Edwards, Shelton, & Jandorf, 2013; Wong, Bloomfield, Crookes, & Jandorf, 2013), as well as economic, health care system, and provider barriers, ranging from fewer screening discussions and choices offered by physicians to lack of transportation, lack of health insurance, and inability to take time off from work (Bromley, May, Federer, Spiegel, & van Oijen, 2015; Coleman-Wallace, Baltrus, Wallace, Blumenthal, & Rust, 2013; James, Daley, & Greiner, 2011; Robinson et al., 2010; White, Vernon, Franzini, & Du, 2010).
Seeking to improve CRC screening among minority and low SES populations, the Centers for Disease Control and Prevention (CDC) conducted demonstration projects to develop evidence-based interventions (DeGroff, Boehm, Green, Holden, & Seeff, 2008; Seeff et al., 2008). From this work, the CDC identified establishing partnerships with community health programs, particularly local comprehensive cancer control programs, as an important facilitator to increase screening rates (Breslau et al., 2010; Freeman, Muth, & Kerner, 1995; Lane, Cavanagh, Messina, & Anderson, 2010). Other notable interventions included incorporating multidisciplinary community–academic partnerships in planning and implementation, providing patient and provider education, and using patient navigation (Christie et al., 2008) to enroll patients in all possible screening modalities.
Provider education programs receiving greater CDC Colorectal Cancer Control Program funding and higher quality ratings have included both group and one-on-one sessions focused on provider reminders, assessment and feedback, and quality assurance and improvement activities (Maxwell et al., 2014). Patient education programs, on the other hand, have primarily focused on small media campaigns, with preferred tools consisting of brochures, posters, inserts, and lists of questions to ask providers (Kreuter et al., 2012).
Patient navigation seeks to guide patients through potential barriers to care in order to ensure timely diagnosis and treatment (Chen et al., 2008; Hendren et al., 2011). Navigation models have used various modes of contact, including in-person, telephone, and mailers, as well as various types of navigators, from health care providers to lay community members. Across these models, patient navigation has consistently increased CRC screening rates in low SES and minority populations (Lasser et al., 2011; Percac-Lima et al., 2009). With regard to Black Americans specifically, Myers et al. (2014) observed a significantly increased odds of completing CRC screening at 6 months, odds ratio (OR) 2.1 (95% confidence interval [CI] [1.5, 2.9]) and 12 months, OR 1.7 (95% CI [1.2, 2.3]) among patients randomly assigned to receive tailored navigation compared to standard of care. Patients receiving navigation have also reported improved emotional support, assistance with information needs and problem solving, logistical coordination of care and a high rate of satisfaction with care (Carroll et al., 2010; Treml, Conlon, Wegner, Baliker, & Remington, 2009).
Fecal occult blood testing (FOBT) is a guideline accepted option for CRC screening (National Cancer Institute, 2016). Immunochemical tests of blood in the stool, also known as iFOBT or FIT, are superior to standard FOBT in their ability to discriminate for lower gastrointestinal bleeding, have fewer medication and dietary interactions, and often require only one test. iFOBT also provides an excellent alternative for patients who are unable to undergo colonoscopy due to work restrictions, problems with bowel preparation (Peipins, Soman, Berkowitz, & White, 2012), intolerance of sedation, inability to afford out-of-pocket costs, or distrust (Robinson et al., 2011). iFOBT may also be particularly helpful in increasing screening rates within the Black American community, for although Hawley et al. (2008) observed a greater preference for colonoscopy among Black Americans, multiple subsequent randomized trials observed that Black Americans are significantly more likely to complete stool-based screening modalities (Gupta et al., 2013; Inadomi et al., 2012; Myers et al., 2014).
The Milwaukee Regional Cancer Care Network, a forum of more than 40 community stakeholders, including health care providers, public health practitioners, researchers, and cancer control advocates, identified improving colorectal screening rates as a priority for the Milwaukee area. In 2008, project partners created a community advisory board (CAB) to develop a proposal for a pilot screening and navigation program at Progressive Community Health Centers (PCHC).
Our project sought to increase CRC screening rates at a Federally Qualified Healthcare Center (FQHC) by using an academic–community health center partnership to implement CDC evidence-based interventions of education, iFOBT, and patient navigators. The project sought to screen at least 200 asymptomatic, average risk Black American men and women between 50 and 75 years of age.
This article adds to the literature by providing specific and comprehensive information on a strategic approach to improve CRC screening in a racial minority population by promoting linkages between academic and community health researchers, and partnering this research with education, training, and tools to improve health promotion among providers, patients, and patient navigators.
Method
Partners
This project was done in collaboration with the University of Wisconsin Carbone Cancer Center (UWCCC), the University of Wisconsin School of Medicine and Public Health (UWSMPH), PCHC, the Milwaukee Regional Cancer Care Network, and the Center for Urban Population Health. The UWCCC is the only National Cancer Institute-designated comprehensive cancer center in Wisconsin and is the home for the Wisconsin Comprehensive Cancer Control Program (Treml et al., 2009). UWSMPH is the largest medical school in Wisconsin. The Center for Urban Population Health is a partnership between UWSMPH, Aurora Health Care, and the University of Wisconsin–Milwaukee’s College of Health Sciences.
Pilot Site
PCHC is a FQHC consisting of two clinics that provide primary care services, health education classes, and oral health care. The clinic serves a low-income population, with 59% on Medicaid and 26% uninsured. In 2008, it served 5,501 patients, 79% of whom were Black Americans.
Intervention Development
Once PCHC was identified as the pilot site, the Milwaukee Regional Cancer Care Network created a CAB to oversee and coordinate implementation of the project. The CAB was composed of representatives from all five health systems in Milwaukee: the PCHC pilot site, ACS Sankofa Project, National Cancer Institute’s Cancer Information Service North Central Region, Wisconsin Primary Health Care Association, Wisconsin Comprehensive Cancer Control Program, and the Center for Urban and Population Health, UWCCC, a CRC survivor, and a pilot site patient (Figure 1). In the 6 months leading up to project implementation, the CAB met monthly to review literature, CRC screening resources, new promotional materials, pilot site screening data, and strategies to remove known screening barriers. Integrating available guidelines with staff interviews from all five local health systems, the CAB developed a patient navigation screening protocol tailored to meet pilot site resources, staff, and workflows. Protocol development also emphasized sustainability and generalizability to other local health systems. The resulting intervention included four main components: (1) community and patient education, (2) provider training and education, (3) expansion of improved screening technology, and (4) patient navigation (Supplemental Figure 1).
Composition of the Milwaukee (Wisconsin) Regional Cancer Care Network
Patient and community educational materials included clinic-based education (educational boards and looped video display), displays at community events (e.g., health fairs), two billboards, four bus stop advertisements, and articles contributed to a local newspaper (Supplemental Figure 2). Messaging for these materials was based on the “Stages of Change” theoretical framework, a framework identified and recommended by the National Colorectal Cancer Roundtable (Sarfaty, 2008). The concept is that when health messages are appropriately matched with a patient’s stage of change (precontemplative, contemplative, preparation, action, maintenance, and relapse), patients are more likely to progress to the next stage and change their behavior. A specific example of how we employed stages of change framework in our physician messaging campaign is presented in Figure 2.
Physician Messaging Based on Stages of Change Theoretical Framework
Provider training and education consisted of two 1-hour long sessions with clinic nursing staff and physicians. The initial session covered current CRC screening guidelines and their application to PCHC patients, implementation protocol, iFOBT patient education, and how to handle and process samples. A second session was held after the clinic converted to an electronic medical record (EMR) system, at which time all the previous learning objectives were reviewed, including how the pilot protocol would now be integrated with the new EMR.
We used two different iFOBTs during the course of our project: a simpler one-step test at the beginning and a three-step test at the end, due to the one-step test being recalled. The flow of iFOBT services is presented in Figure 3. The CAB also developed and tested a screening kit instruction brochure for patients who wanted to take the kits home.
Colorectal Cancer Screening Pilot Design
The initial navigator was a medical assistant hired from among the clinic staff. The navigator maintained some clinical duties but was prioritized to be available for patient consultation when patients chose to participate in CRC screening. Using chart review, the navigator also identified individuals who were out of date on their screening and called them to schedule screening appointments. The CAB further adapted a patient screening checklist and follow-up call script for use by the navigator (Sarfaty, 2008). The navigator was charged with following up with patients (via phone and in person during clinic visits) about whether the patient was able to complete the iFOBT screen and identify and address any barriers to completion. If the patient screened positive, the navigator assisted the patient in obtaining a colonoscopy. Several months into the project, the navigator resigned. In response, the pilot site leadership team decided to transition from a single navigator to a navigation team consisting of multiple nurses from the clinic staff.
Study Sample
We enrolled Black American patients 50 to 75 years of age at average colon cancer risk, defined as having no family history of colon cancer, and no known inflammatory bowel disease or prior personal history of colon cancer. Patients were enrolled on a continual basis from April 2010 to December 2011, being identified during clinic visits and were immediately connected with a navigator at the conclusion of their clinic appointment. Patients were excluded for having a positive personal or family history of colon cancer, active symptoms of colon cancer, or were already up to date with screening. The project screened 402 patients for CRC (Table 1).
Summary Results of Colorectal Cancer Screening Program, April 2010 to December 2011
NOTE. EMR = electronic medical record implementation; iFOBT = immunochemical fecal occult blood test.
Implementation
Navigators provided qualifying patients with education, an iFOBT kit, and an instructional DVD with demonstrations, and assessed patients for potential barriers to completing screening. Patients were instructed to return completed kits to PCHC within 1 week. Navigators contacted patients who did not return kits within this timeframe via telephone and addressed further barriers to completion. PCHC nurses followed up iFOBT results within 1 week of completion, either by phone or during the next clinic visit, whichever occurred first. In partnership with the ACS, participants received $10 gift cards on completion of screening. Uninsured patients with positive iFOBT results were referred for colonoscopy at no cost by a local medical center. Navigators facilitated referral and follow-up processes to ensure completion (Figure 3). Based on prior published studies, we anticipated a 3% positive screening rate (Inadomi et al., 2012).
Funding and Institutional Review Board
University of Wisconsin Health Sciences Institutional Review Board deemed this quality improvement project exempt from human subjects research review.
Results
Education
Ninety-two percent of clinic staff (including medical assistants, nursing staff, nurse practitioners, and physicians) participated in the first education session at the beginning of the screening program.
Screening
Prior to the pilot, the most common CRC screening methods were either a one-time in-office guaiac FOBT via digital rectal examination or referral to colonoscopy. Free or reduced-rate colonoscopy at a local medical center was available for those who lacked insurance coverage for the procedure. Under this baseline screening system, the overall referral rate for CRC screening was approximately 30%, and not all who were referred went on to complete a colonoscopy.
Demographics of the clinic population at the time of this project were 79% Black, 70% under the federal poverty level, 59% Medicaid, and 26% uninsured. The clinic served a total of 5,501 patients and had 12,339 medical encounters per year. During the course of the project, 402 patients were identified who met the criteria for screening. Under our single navigator and one step iFOBT model, 43 of 52 patients (82.7%) completed iFOBT screening, while 185 of 335 (55.7%) completed screening under our team navigator and multistep iFOBT program (Table 1). The overall rate of 56.7% represented an approximate twofold increase in CRC screening compared to the baseline rate (30%–57%).
Among those completing iFOBT, nine patients (3.9%) screened positive, consistent with prior prevalence studies. From this positive group, four of nine patients (44%) went on to obtain colonoscopy. Prior research suggested that roughly one third would complete their colonoscopy (Inadomi et al., 2012).
Navigation
Patient navigators and nursing staff contacted all 402 eligible patients (100%) for referral to iFOBT. Furthermore, they contacted all 228 patients (100%) with the results of their iFOBT screening. All nine patients who had a positive screen were referred to colonoscopy. Four patients were scheduled for colonoscopy. No patients were diagnosed with CRCs or adenomatous polyps during the course of the pilot.
Rates of successful completion decreased after shifting from a single navigator (a medical assistant who focused on CRC screening) to a pool of nurses who functioned as navigators in addition to their usual role as a clinic nurse (Table 1).
Completion rate also fell after the introduction of an EMR, even though the total number of iFOBT kits offered and completed increased during that time. The change to an EMR and from a single navigator to a pool of nurses happened at roughly the same time and were unanticipated changes when the project was conceived and initiated.
Discussion
Using education, iFOBT, and patient navigation, our community–academic partnership successfully piloted a CRC screening program at an urban FQHC in Milwaukee, Wisconsin, and resulted in a substantially increased CRC screening rate among low-income Black Americans. This intervention effectively doubled the CRC screening rate and has been a sustainable change for these two FQHC clinics. Based on this substantial increase in CRC screening, PCHC has now prioritized iFOBT as their preferred method of CRC screening for patients who are unable or unwilling to obtain a colonoscopy. Colonoscopy is still available and recommended for patients who prefer and are able to undergo colonoscopy. This approach aligns with studies observing that navigation is most effective when tailored to a patient’s screening preference (Myers et al., 2014).
As a result of the pilot project’s success, PCHC secured long-term reimbursement for iFOBT, enabling the clinic to continue the screening program beyond the pilot period and demonstrating the sustainability of this project.
This project also demonstrated the relative success of a single medical assistant patient navigator over a pool of nurses. Based on this observation, we propose that patient navigation is more effective when the navigator role is clearly defined, assigned, and held accountable for improving CRC screening rates. The clinic has since reverted to a dedicated single individual navigator focused on cancer and chronic disease navigation funded out of clinical revenue.
Limitations
We did not collect participant demographic data. Our cumulative enrollment process yielded a perfect 100% initial contact and enrollment rate, eliminating any possibility of nonrepresentative recruitment. Our method of enrolling patients, however, namely, in-person contact at the conclusion of a routine clinic appointment, may have selected for the most resilient, well-educated, and motivated patients, with overall better insurance status and fewer work or transportation barriers given the simple fact that they were all able to make their scheduled clinic appointments. If this is true, it could bias our results toward higher screening rates.
Because this was not an observational study of comparative samples, however, there was by definition no differential recruitment, thus any bias in our findings would not be an issue of selection bias threatening the validity of a statistically significant observation, but rather speaks to the limitations of the generalizability of our findings. Our findings hold for patients who attend their clinic appointments and receive initial in-person contact by a patient navigator. For patients who fail to attend clinic appointments, however, our intervention would be ineffective. Methods incorporating navigation through mailers and telephone outreach (Lasser et al., 2011; Myers et al., 2014; Percac-Lima et al., 2009) would be more appropriate for these patients, however, even within these studies, the numbers of patients excluded due to an inability to contact them through mail or telephone were substantial, ranging from 23% to 61%.
We encountered several unanticipated challenges during the implementation of this project, but they are not necessarily limitations of study design as much as important real-world lessons to be learned and may help guide successful future interventions. Our navigation model changed mid-project, from a single dedicated medical assistant to a team of nurses. Because nurses were also responsible for processing the iFOBTs, there may have been an inherent conflict of interest in asking them to also serve as patient navigators, since greater navigation would directly increase their workload of processing tests. Switching to the more laborious three-step iFOBT may have further compounded this potential conflict of interest, negatively impacting patient navigation and, in turn, CRC screening rates. The three-step iFOBT was also more difficult for patients and likely decreased their motivation and ability to complete screening. Lastly, transitioning to an EMR during the course of the project diverted staff efforts from CRC screening to learning the EMR system. Collectively, these unforeseeable setbacks likely decreased our screening rate. As such, we believe it may be possible to achieve even higher screening rates at this FQHC with continued use of a single navigator, and full transition to EMR.
Because several changes occurred at the same time (e.g., change in navigation model and iFOBT test), it is difficult to assess the impact of any one specific factor. Furthermore, no qualitative data were collected. Obtaining perspectives from participants, providers, and the community regarding what worked, and what did not work, would have been helpful to the development of further projects.
Our intervention was not designed to generate new qualitative or quantitative descriptions of potential barriers to screening or test such observations under formal study hypotheses. Rather, our intervention was designed to use education and patient navigation to identify and remove already known barriers from prior literature (Bromley et al., 2015; Coleman-Wallace et al., 2013; James et al., 2011; May et al., 2015; Robinson et al., 2010; Sly et al., 2013; White et al., 2010; Wong et al., 2013). For example, our provider education sessions addressed the known barriers of poor provider awareness and fewer discussions and options provided to patients. Our community education intervention addressed lack of patient knowledge about perceived risk and screening options, and patient navigation addressed financial and logistical barriers.
Strengths
Our project is unique in that it is one of a small number of studies assessing patient navigation in conjunction with iFOBT instead of colonoscopy. Studying iFOBT is particularly important since this testing modality appears to be more available, accessible, and acceptable, among low-income, uninsured, and Black Americans. This project adds to the literature by showing that with academic–community support, a screening program comprised of dedicated patient navigation, simple one-step iFOBTs, and provider/patient education about proper CRC screening guidelines, can be implemented to reduce cancer screening disparities among low-income minority populations. It also adds to the generalizability of breast and cervical cancer studies showing similar improvements in screening rates among minority and low-income patients after incorporating patient navigation and education (Glick, Clarke, Blanchard, & Whitaker, 2012; Marshall et al., 2016).
Future Research
Lasser et al. (2011) discuss the need for further research on the optimal frequency and duration of primary care–based patient navigation, particularly with respect to sustainability and long-term efficacy. Given the relative success of our single dedicated navigator model, it may be that with a single full-time specialized navigator, interactions could be shorter, cover more topics (e.g., breast, cervical, and lung cancer screenings) while still maintaining efficacy and not impeding the workflow of other providers.
While our 56.7% iFOBT screening rate represents a significant improvement in CRC screening, unlike colonoscopy, iFOBT does not offer the additional diagnostic and therapeutic benefits of definitively detecting and removing precancerous polyps or confirming cancer. Because of our intervention, nine patients received actionable information about a positive iFOBT test result. Only four of them, however, continued on to obtain colonoscopy. Additional research must address whether patient-navigated increases in CRC iFOBT screening maintain their effectiveness in terms of getting patients followed through to colonoscopy, and in turn, are truly effective at reducing cancer incidence and mortality in Black Americans, other minorities, uninsured, and low-SES communities.
In conclusion, our pilot intervention demonstrates of the power of partnering invested groups in targeting cancer disparities. Key lessons learned for future implementation include the importance of dividing responsibilities among partners so that each is able to provide assets to the group without any one becoming over-burdened, tailoring a program to meet facility capacity, resources, and workflows, while also being flexible to changes in staffing, lab, and technologies, and matching educational messages and screening choices with patient and population preferences. This academic–community partnership approach adds to the literature by providing specific and comprehensive information on how to improve screening disparities in an urban, low-income Black American community by employing a multifaceted strategy of iFOBT, patient and provider education, and patient navigation.
Footnotes
This project was funded by the University of Wisconsin Carbone Cancer Center Support Grant P30 CA014520 and the Wisconsin Partnership Program of the University of Wisconsin School of Medicine and Public Health.
Supplementary Material
Supplemental Figures 1 and 2 are available with the article online at 
.