Implementation of culturally targeted patient navigation system for screening colonoscopy in a direct referral system

Trends in CRC screening, incidence and mortality

Overall CRC screening rates have improved in recent years due to aggressive attempts to motivate screening and expanded insurance coverage. Data from the 2008 National Health Interview survey found that the prevalence of structural CRC screening (colonoscopy and flexible sigmoidoscopy) among US adults aged 50–75 years increased from 51.9% in 2005 to 62.9% in 2008 [8]. Fecal occult blood tests (FOBTs) also remain a viable option for those who prefer a non-invasive procedure (or in resource-poor areas), though they lack the benefit of simultaneous screening and treatment [9–11].

Despite the increased availability and efficacy of CRC screening, disparities exist regarding screening rates and ultimately CRC incidence and mortality. Doubeni et al. [12] used the Medicare Current Beneficiary Survey of Medicare enrollees aged 65–80 years to analyze CRC screening rates for African Americans and non-Hispanic whites between 2000 and 2005. Despite overall improvement in screening rates, disparities persisted over the 5-year period. Screening rates among non-Hispanic whites in 2000, 2003 and 2005 were 49.0%, 52.8% and 56.6%, respectively, whereas among African Americans the rates were 41.0%, 49.6% and 52%.

These differences have significant consequences: African Americans have the highest CRC incidence and mortality of any racial/ethnic group in USA [13–15]. The American Cancer Society [13] reports that African Americans have a 20% higher incidence rate and a 45% higher mortality rate than non-Hispanic whites. In fact, although CRC mortality has declined by 39% among non-Hispanic whites since 1960 and mortality among African Americans has ‘increased’ by 28% [16]. Higher mortality rates among African Americans may be largely explained by the later stage at which CRC is diagnosed [17], emphasizing the importance of improving adherence to screening guidelines in this population.

Determinants of screening adherence

To improve screening rates overall and eliminate racial disparities, it is necessary to understand the multifactorial barriers to screening adherence. System-level barriers such as insurance, health care policies and economic constraints often hinder low-income African American patients’ participation in cancer prevention services [18]. Personal and sociocultural barriers also affect adherence: lack of education or low health literacy, medical mistrust, fatalism and fear of the procedure [19–22].

Shifts in health care delivery policy have begun to address system-level barriers. One major shift in colonoscopy among low-income groups was a public health policy mandating coverage of cancer screening costs by Medicare and Medicaid [23]. Effective mid-2003, New York State Medicare and Medicaid reimbursement policies were modified to include coverage of SC for CRC. Another important shift was the establishment of Direct Endoscopic Referral Systems (DERS), whereby primary care physicians (PCPs) refer average-risk patients directly for SC, avoiding the inconvenience and delay of an interim office consult with a gastroenterologist [24]. The effectiveness of these developments can be enhanced through interpersonal interventions that directly target personal and sociocultural barriers.

Use of patient navigation to improve adherence to screening

A patient navigator is a specially trained person within the healthcare setting who helps a patient move through the system to obtain medical care [25]. Historically, most PN programs have assisted cancer patients in obtaining follow-up of suspicious findings and treatment [26–28]. Recently, the concept of PN has expanded to screening, focusing primarily on breast and cervical cancer screening [29–32]. Phillips et al. [33] found that PN significantly improved screening mammography adherence in a randomized sample of ∼4000 inner-city dwelling women with a baseline adherence rate of 78%. After a 9-month intervention, mammography adherence was 87% in the intervention group and 76% in the control group (P < 0.002).

Colonoscopy is a more complex screening test than mammography and cervical PAP smears and therefore may require additional interventions to increase adherence. Our group was among the first to introduce PN to facilitate completion of colonoscopy, specifically among minority primary care patients. In 2003, we established DERS at Mount Sinai Hospital and hired our first PN in April 2004. Mount Sinai’s first PN was a bilingual Latina health educator trained in issues related to DERS, CRC screening, bowel preparation, scheduling, transport and supportive counseling. The implementation of this system resulted in an improvement in adherence to physician-recommended SC from 40% to 66% over a 3-year period [6]. While we were encouraged by the success of our first navigation program, we noticed that Latinos were 1.67 times more likely to complete SC than African Americans (P = 0.013). Because our first PN was Latina, we hypothesized that the ethnicity of the PN might affect the screening outcome.

Culturally targeted patient navigation

Targeted interventions are developed specifically for the demographic, behavioral or psychosocial characteristics shared by members of a particular population subgroup [34]. In addition to focusing on overcoming system barriers, CTPN incorporates discussion of personal and cultural barriers into the navigation paradigm. Previous studies have reported that culturally targeted interventions are perceived as more credible [35], result in greater retention of knowledge over time [36] and are more likely to result in increases in targeted health-promoting behaviors than non-targeted PN [37].

Inclusion of peers as patient navigators

Research confirms the benefits of using peer educators and lay health advisors in healthcare interventions. In cancer prevention, community health workers have increased screening knowledge [38], mammography and breast self-examination rates [39, 40]. Peer navigators have improved diagnostic follow-up after mammography among minority woman [41] and SC adherence among patients identified as unlikely to follow through with their first scheduled screening [42].

Recruitment

In this institutional review board-approved trial, African American patients referred for SC by their PCPs were recruited during a scheduled, routine (non-acute) visit at Mount Sinai’s primary care clinic. PCPs were educated about eligibility criteria for the study and referred eligible patients. PCPs explained the study to each referred patient. Interested participants were introduced to a research assistant after the appointment to discuss the study and give informed consent. PCPs ordered the SC using an electronic medical record in which the criteria for DERS were delineated.

Eligibility criteria included African American patients >50 years of age, without active gastrointestinal symptoms, significant comorbidities that would necessitate a medical consult prior to undergoing a colonoscopy (e.g. chronic obstructive pulmonary disease, severe heart disease) or a history of inflammatory bowel disease or CRC. Patients must not have undergone colonoscopy for at least 5 years or have been up to date with other forms of CRC screening (e.g. FOBT, flexible sigmoidoscopy) and they must have telephone service. Referrals were reviewed by nurses in the Division of Gastroenterology via electronic medical record review after recruitment to confirm eligibility and evaluate for contraindications to colonoscopy or sedation. Subsequently, the Project Coordinator randomized participants to each study arm (Pro- versus Peer-PN). Randomization was created by our statistician using SPSS 19.0 to fulfill our cells in a 2:3 ratio (Pro:Peer), as determined a priori for maximum power.

Recruitment and characteristics of patient navigators

Five Peer-PNs were successfully recruited as lay navigators through fliers posted in the endoscopy suite and primary care clinic of Mount Sinai School of Medicine and through a previous study. African Americans >50 years of age who had recently undergone SC were considered eligible to apply. There were no education or employment requirements. Of the five peer navigators, three were under the age of 65 years, four were female and at least three had completed 12 or more years of education. All but one had previous telephone work experience. Peer-PNs were compensated on an hourly basis.

By comparison, four African American Pro-PNs were successfully recruited though Internet employment postings. Applicants were required to hold a bachelor’s degree or higher and have experience working with minority communities and conducting research. There was no eligibility requirement pertaining to colonoscopy completion. Of the four Pro-PNs, all were females under the age of 35 years. Three held master’s degrees. Pro-PNs’ compensation was based on annual salaries of Mount Sinai School of Medicine. Salaries for all navigators were supported by a grant from the National Cancer Institute.

Training culturally targeted patient navigators

CTPN training incorporated four approaches to cultural targeting: peripheral, linguistic, evidential and sociocultural [7]. The peripheral approach involved a member of a particular group presenting material in a way that was most likely to appeal to the group’s other members. For example, all navigators and participants were African American. The linguistic approach focused on using colloquial language and limiting medical jargon to target low health literacy and limited general comprehension of CRC and colonoscopy. Navigators were trained to provide information in layperson terms. ‘CRC’ was referred to as ‘colon cancer’ and ‘physician’ as ‘doctor’. CTPNs were provided with a glossary of alternate definitions and explanations for relevant medical terms.

The evidential approach imparted the importance of CRC screening for African Americans. CTPNs were trained to discuss statistics such as incidence and mortality rates of CRC among African Americans versus other groups, lower awareness of and participation in CRC screening in African American communities, and the potential benefit to African American communities if CRC screening were to become more prevalent. The sociocultural approach addressed health-related issues in the context of the broader social and cultural values and concerns of African Americans. CTPNs were trained to discuss a range of barriers including concerns about safety, fear of cancer diagnosis and treatment, racial and gender concordance between patient and provider, physicians’ motives and medical mistrust. Studies by our group and other community work have shown that discomfort with the sexual connotation of colonoscopy, fear of pain and embarrassment are common barriers to compliance [18, 22, 43]. Although these particular barriers exist across racial/ethnic groups, there is evidence that they are more widespread among African Americans [44–47]. CTPNs were thus trained to emphasize that the gastroenterologist performing the colonoscopy would make every effort to preserve participants’ privacy and dignity in order to reduce embarrassment and emotional discomfort. Discussions were framed to address possible concerns about being treated disrespectfully and insensitively.

The CEDIP model, developed by the research team and based largely on motivational interviewing techniques [48–51], is a semi-structured directive and anticipatory approach to PN implemented to enhance cultural competency through mastery of five tasks, forming the acronym: clarify, empathize, disclose, inform and plan. During CEDIP training, navigators explored and addressed barriers to colonoscopy, responded to a range of participant responses and were trained to systematically address financial, structural, psychological and sociocultural barriers to SC. Because they had no personal experience with colonoscopy, Pro-PNs were trained using a nearly identical model, CEEP (clarify, empathize, educate and plan).

Pro-PNs completed training over a 2-week period while Peer-PN training was extended over 3 months; the two groups were mainly trained separately. Training consisted of three sessions reviewing CRC topics and navigator responsibilities, a session with a gastroenterologist to provide a clinical perspective, telephone communications training and a series of one-on-one structured role plays simulating a navigation encounter. Periodic assessments were administered to evaluate PN knowledge. At the beginning of training, the Peer-PNs had significantly lower knowledge levels than the Pro-PNs. By the last assessment there was still a difference but it was no longer significant. See Shelton et al. [52] for details about the training process and navigator characteristics.

Peer versus professional navigators and protocol

The principal difference between Peer- and Pro-PNs was that Peer-PNs modeled coping with concerns about the exam (e.g. anxiety, fear and/or discomfort) by discussing strategies that helped them get through their own exams. This disclosure about colonoscopy from a ‘similar other’ gave patients the opportunity to hear and benefit from information relevant to their own CRC screening expectations and experience. Other than this key distinction, navigators followed an identical protocol. Based on our previous work, navigation included three phone calls. Call attempts were made at various times of the day and days of the week, including the weekend. Call logs were examined regularly to determine when to consider a person not reachable. At this point a letter was sent to the participant, asking them to contact us. During the initial call, the PN made a colonoscopy appointment, answered basic questions and provided information about the preparation and the procedure. Following the first call, the PN mailed written instructions for bowel preparation and a reminder postcard with the patient’s colonoscopy date. The PN subsequently contacted each participant 2 weeks and 3 days prior to their procedure to remind them of their appointment, confirm receipt of mailed information, review bowel preparation instructions, assess transportation needs and make arrangements if necessary, and provide education and support. PNs offered a detailed explanation of what to expect on the day of the procedure from time of registration until discharge.

Data analysis and outcomes measured

Demographic information was collected during an interview (Time-1 Assessment) with a research assistant at the time of recruitment. Socio-demographic characteristics were compared between study participants in the Peer-PN group versus those in the Pro-PN group. Analyses were conducted using SPSS 19.0 software. Equality of proportions for categorical variables was compared using a chi-square test. Equality of means was tested using one-way analyses of variance. Fidelity monitoring was conducted with 10% of audio-recorded telephone calls to ensure adherence to the content of the scripts and equivalent performance between the Peer- and Pro-PNs. The mean number of phone calls attempted and mean number of minutes spent per navigator per participant were measured to detect differences in navigation delivery.

We report three outcomes of PN success as suggested by the National Patient Navigator Leadership Summit of the ACS [53]. Our main outcomes were completion of colonoscopy via chart audit, patient satisfaction with the program and patients’ level of trust in the navigator. Trust in the navigator was measured in an interview with a research assistant after the 2-week navigation call (Time-2 Assessment) in terms of personal trust in the navigator and trust in the information presented by the navigator (message and source credibility). Patients indicated their level of agreement with six statements about personal trust on a 5-point scale. Message and source credibility was measured by indicating agreement with eight statements on a 4-point scale (see Fig. 2 for both scales). Patient satisfaction was assessed in a post-colonoscopy interview with a research assistant (Time-3 Assessment) through two questions: one on a 5-point scale and one on a 4-point scale (Table I).

Study participants

In our preliminary analysis of the first 24 months of this ongoing study (May 2008–May 2010), we enrolled 288 African American patients, 240 (83.3%) of whom were medically cleared and randomized [134 in the Peer-PN arm (55.8%) and 106 in the Pro-PN arm (44.2%)] (Fig. 1). Recruitment was excellent with 91.4% of eligible primary care patients agreeing to enroll in the study. A total of 48 participants (16.7%) were excluded from randomization due to medical ineligibility (n = 45; 15.6%) and personal reasons (n = 3; 1.0%); thus, 240 participants were randomized. However, after randomization, an additional 48 participants (16.7%) were excluded from analysis because they were not able to receive navigation services. Reasons for exclusion included the following: unable to be reached (n = 32; 66.7%), became ineligible due to medical conditions (n = 14; 29.1%), a pending colonoscopy (n = 1; 2.1%) and deceased (n = 1; 2.1%). Thus, 192 patients (80.0%) (116 in the Peer-PN arm and 76 in the Pro-PN arm) successfully received navigation services although not all subsequently completed a colonoscopy.

Baseline socio-demographic characteristics of study participants are shown in Table I. Women represented 70.0% of the study population. The mean age was 59 years; 66.7% of participants had less than or equal to a high-school education. The majority (78.2%) were unmarried and 49.5% of participants had annual household incomes <$10 000. Approximately half (50.2%) were covered by Medicaid while 29.1% were covered by Medicare and 20.7% had private health insurance. The Peer-PN and Pro-PN groups were comparable, with no statistically significant differences in socio-demographic characteristics.

Mean calls attempted and minutes spent per participant

Data showed no significant difference between the mean number of calls attempted per completer between Peer-PNs (8.4) and Pro-PNs (10.1) (P = 0.072). Similarly, the difference between the amount of time spent on the phone with each completer was not significant (51.6 min for Peer-PNs; 52.1 min for Pro-PNs) (P = 0.900).

Colonoscopy screening rates, patient satisfaction and trust

In our sample, 74.7% of all patients completed SC. In the Peer-PN group, 71.3% completed as did 80.0% in the Pro-PN group. There was no statistically significant difference in SC adherence between the two groups (P = 0.178). 83.8% of participants completed the Time-3 Assessment. Overall, they rated navigation very well: 85.1% of participants in the Peer-PN group rated it ‘Very Good’ or ‘Excellent’, versus 94.1% of participants in the Pro-PN group (P = 0.072). Ninety-five percent of all participants would recommend a navigator to family or friends; 85.9% of participants completed the Time-2 Assessment. Personal and informational trust in navigators was high in both groups. As Fig. 2 shows, statements in the ‘Trust in Navigator’ scale were negative comments about the navigator starting with, ‘I doubt that the Navigator really cares about me as a person’. Strongly disagreeing (score = 1) indicated the highest level of trust. Total scores ranged from 6 (highest level of trust) to 30 (lowest level of trust). Peer-PNs on average received a score of 10.6 and Pro-PNs received a score of 10.9 (P = 0.564). With a high score of 32 indicating the highest level of message and source credibility, Peer- and Pro-PNs on average both received a score of 30.2 (P = 0.971). There were no statistically significant differences in either measure of patient satisfaction or either measure of trust in the navigator. No participants experienced unintended effects or harms.