Interventions to Increase Colorectal Cancer Screening Among African Americans: A Systematic Review

Colorectal cancer (CRC) continues to affect African Americans disproportionately. Despite medical advances and widely accepted screening recommendations, African Americans are less likely to get the appropriate screening and consequently are more likely to die from colorectal cancer than their white counterparts. Appropriate communication between the patient and the provider and increased patient education may be a part of the solution to this discouraging problem. Increasing provider education and cultural awareness may increase colorectal cancer screening among African Americans. The objective of this paper to perform a systematic review of the published literature to assess the effectiveness of interventions aimed to increase colorectal cancer screening among African Americans. Seven online databases were systematically searched for article published between Jan. 2000 and April 2013 using subject terms taken from the Medical Subject Headings (MeSH), the list of standardized descriptors used by the National Library of Medicine (NLM), to standardize the search. Studies that measured CRC screening rates and met the inclusion/exclusion criteria were selected. Data was extracted and independently reviewed by three reviewers. Study design, population characteristics, experimental intervention, control intervention and outcomes were extracted from the selected article. Of the 649 studies identified as potentially, eleven articles were selected for the review. Four articles reported using culturally tailored interventions and seven articles did not use culturally tailored interventions. Two studies have intervention aimed at physician education. Eight articles reported statistically significant results. Because of the limited number of quality studies, no conclusive recommendations can be made regarding the contribution of culturally tailored intervention to increasing CRC screening among African Americans.

1 Department of Internal Medicine, University of Texas Medical Branch, UTMB, Galveston, TX, USA

Background
According to the American Cancer Society (ACS) [1] , colorectal cancer (CRC) is the third most common cause of cancer death in both men and women (ACS 2013). Although colorectal cancer death rates have decreased in African American men and women over the past two decades, the decline began later and has been slower among African Americans compared to whites, resulting in a widening racial disparity (ACS 2013). Since 1987, CRC screening rates among African Americans (non-Hispanic) have increased from 4% to 56% in 2013. However, this rate remains lower than screening rate in Whites (non-Hispanic) 62% (ACS 2013). The ACS 2013 CRC screening recommendations for individuals over 50 included the following: •Yearly Fecal Occult blood Test (FOBT) or Fecal Immunochemical Test (FIT) •Yearly FOBT of FIT plus flexible sigmoidoscopy every 5 years •Flexible sigmoidoscopy every 5 years •Colonoscopy every 10 years •Double-contrast barium enema every 5 years.
Inadequate patient education and poor communication with their health care provider may explain some of the racial differences in the CRC screening rate. The aims of this review are: 1.To identify and describe culturally relevant or non-culturally relevant interventions targeting African Americans that has shown to effectively increase colorectal cancer screening. 2.To identify elements in the provider-patient relationship, such as communication, trust, patient education, and provider recommendations that is associated with increased CRC screening among African Americans.
Earlier literature on colorectal cancer attributed the higher colorectal cancer death rates in the African American population to a higher proportion of advanced disease at diagnosis [2] . Despite the fact that screening would diagnose disease at an earlier stage which might lead to a better prognosis, African Americans may be less likely to undergo screening for earlier diagnosis because of decreased access to care [3] . Although white women have a higher cancer incidence rate, black women are more likely to die of cancer (US Cancer Statistics Working Group, 2005). The high cancer mortality in the black population may be related to blacks being diagnosed at later stages [4,5] racial differences [6,7] and lack of definitive treatment [6] . Poverty, racial bias, and stereotyping may play a critical role in the willingness of African Americans to seek screening for colorectal cancer. The provider-patient interaction may be negatively affected by racial discrimination and social disparities resulting in barriers in access to high quality care (ACS, 2013).
When attempting to advise patients regarding CRC screening, it is important that providers understand the cultural norms and beliefs in the population they serve, especially if the patient population is mostly of a different ethnic group [8] .
Patient education may also influence an individual to seek CRC screening. Studies have shown that African Americans may be less likely to get CRC screening because of lack of knowledge, understanding or the ability to distinguish a screening test from any other test when compared to whites, even when controlling for education levels [9] . A CRC education and screening program in a diverse low-income, uninsured population located in the state of Maryland's largest county was able to raise the screening rates in the uninsured from 13% to 52% over a two-year period [10] . Provider recommendations for CRC screening have also been shown to increase screening rates in African Americans. In a cohort of 1599 veterans in the Veterans Administration (VA) system, when CRC screening was recommended equally among African American and white non-Hispanic patients (71.0% vs. 68.2%, p=0.44), the African American patients were 1.3 times more likely than whites to receive CRC screening when recommended (36.3% vs. 28.9%, p=0.03) [11] . There is a general distrust in the African American community of what might be called "official medicine." This has been called the "Tuskegee phenomenon", a reference that stems from the 40year study that denied syphilis treatment to a group of African American men from 1932 to 1972, even after there was a known cure [12] . Distrust of physicians often leads African Americans to refuse needed procedures and sometimes to reject care [13] . A study conducted in New Orleans, compared African Americans of lower socioeconomic status to members of the Oschner Health Plan (OHP), who were approximately 80% white, and noted the difference in attitudes and beliefs about cancer and health care. Socioeconomically disadvantaged African Americans agreed 35% of the time, versus 12% of the time (p<0.001) for the OHP group to the statement, "I sometimes find it difficult to talk to doctors." The African American group agreed 35% versus 26% (p<0.001) to the statement "I sometimes do not trust what doctors tell me" [14] .
Lack of understanding of black culture may lead to distrust and patient resistance to preventive health care such as CRC screening. In efforts to increase CRC screening in African Americans, there have been several studies that address different aspects of the provider-patient relationship, and how it can influence patients' willingness to seek CRC screening. It is important to identify interventions in the provider-patient relationship paradigm that have proven to be effective in increasing CRC screening in African Americans.

Methods
Several online databases were used to identify relevant literature. These included Medline (via PUBMED, OvidMedline, and the National Library of Medicines Gateway search engine), the Cumulative Index to Nursing and Allied Health Literature (CI-NAHL), the Cochrane Database of Systematic Reviews, Web of Science (including both the Science Social Science Citation and Science Citation Index), and PsycINFO. A search was done for articles published between January 2000 and April 2013 using the Medical Subject Heading (MeSH) database, the U.S. National Library of Medicine's controlled terminology (used for indexing articles for MEDLINE/PubMed). Terms and fragments found in the MeSH database that are relevant to the topic of this treatise were 'professional-patient relationship', 'attitude', 'attitude to health', 'health knowledge', 'practice', 'cultural background', 'customs', 'social psychology', 'beliefs', 'cultural belief', 'communication barriers', 'African American/blacks' and 'colorectal neoplasm'. The MeSH terms allowed a more systematic way to search different databases for articles that use different terminology for the same concepts. Abstracts and articles from the reference lists of the chosen articles were reviewed to identify other articles that may be relevant. The quality of the articles were assessed using a modified version of the U.S. Preventive Services Task Force Hierarchy of Research Design (S. Weller, Personal communication 6/21/08) as shown in table 1.

Study selection
The following were the inclusion criteria for the selected studies: (a) studies whose sample population was greater than 50% African Americans, whose interventions addressed specific factors in the patient-provider relationship as it relates to communication, education, or trust (whether or not this was culturally relevant), (b) studies using the ACS-endorsed guidelines for CRC screening, (c) and studies measuring the rate of completed CRC screening by self-report or chart review. Exclusion criteria were studies in which the sample population was less than 50% African Americans and studies that did not stratify results by race. Eleven articles were identified and they are described in details in table 2 (supplemental material).

Main outcomes measures
The primary outcome measure was rates of CRC screening. Other areas of interest were the specific elements of each intervention, including the type of population, type of study, mode of delivery of the intervention(s), use of theory, intervention tailoring, and the identification of elements that were associated with positive outcomes. Three groups: •One-on-one education:A health educator met individually with participants in this group for 3 45-minute sessions over 3 weeks to review educational materials on CRC risk and screening.
•Group education: a health educator met with participants in groups of 4 to 14 individuals (average, 5 individuals) to review the educational material. Although the material was identical to that covered in the one-on-one counseling, there were more individuals at each meeting, and 4 sessions over 4 weeks were required to cover the material •Financial support: participants were offered financial reimbursement up to $500 for out-of-pocket expenses incurred for CRC screening, including transportation and other nonmedical expenses. A health educator was available to assist with negotiating direct payment and arranging transportation to the physician's office or medical clinic for the screening test.
•Participants attended the introductory session but received no intervention other than accepting the contents of the gift bag with the educational pamphlets.
Screening rate was 17.7% in control, 22

Data extracted
Once articles were selected for review, abstraction forms were developed by two reviewers, T. Washington (TW) and N. Shokar (NS) to discuss what data should be extracted. The data was initially extracted by TW and Maen Masadeh (MM) and the completed tables were checked by Celia Chao (CC). These reviewers (TM, MM and CC) assessed the eligibility of the chosen articles based on the exclusion and inclusion criteria described above. Each reviewer independently reviewed the articles and decided on the accuracy of the extracted data. The data that was extracted included the following: study design, setting, detailed descriptions of the intervention(s)/methods, whether the intervention was culturally tailored, number of participants, average age of the participants, percentage of African Americans, percentage of women, number of patients in each arm of the intervention, length of study and follow up, theory used if any, and outcome.

Study Selection
Overall 649 articles were initially identified by titles and abstracts and were screened. Forty six remained after the initial screening and full-text reports of these articles were reviewed by TM and MM and only 11 articles met the inclusion criteria [17][18][19][20][21][22][23][24][25] . Of the original articles, 603 were excluded because they were not relevant to the review (n=592) or because the article did not have an intervention or did not describe an intervention in the methods section (n=11). Thirty five were excluded because of one of more of the following: they reported less than 50% African American participation (n=12), or they did not stratify their data by race and, therefore, did not report the number of African Americans at all (n=8), or they did not measure rates of CRC screening at the end of their intervention (n=15), (See Figure 1).

Study design and selection
Three of the included studies were randomized controlled trials (RCT) [18,22,24] , three were randomized community trials [15][16][17] ,two studies were a pre-and post-intervention [21,23] . There was a single randomized community interventional tri-al [25] , onesingle-arm, pretest-posttest design [20] , and one nonrandomized community-based intervention study [19] .The setting for the studies varied: two studies were done in nursing homes [14,16] . Two studies were set in North Carolina, of those, one was done in churches situated in African American communities [17] and one was community-wide [19] . Two studies were in South Carolina [19,23] . Three studies were located in urban areas [18,20,24] , and one study was set in a rural area [18] . The others did not specify a rural or urban location. One study was set in an Internal Medicine clinic [21] , a specific location for the clinic was not given. One study was done in federally qualified health center that serves low socioeconomic status patients in Ohio [22] .

Study population
The study populations ranged from as large as 2167 [23] to as small as 106 [15] . Ten studies had a majority of female participants (see Table 2supplemental material), of those, two studies targeted only females [15,19] . The mean ages of the subjects ranged from 52 years [17] to 75 years [18] . Friedman and Borum's 2007 and Blumenthal et al.,2010 studies were done with a population that was 100% African American but gender or average age was not specified in Friedman and Borum et al 2007. Two studies targeted elderly African Americans [15,16] , the other studies followed the screening age recommended by the ACS of > 50 years old. Three studies targeted low-income populations [19,20,22] while others did not specify an income target.

Interventions
All but two studies incorporated patient educational materials such as posters, brochures or letters. This was done either in the control group, intervention group, or both. Four studies used videos [15][16][17]22] . Katz (2007) and his team embarked upon the task of educating entire communities by launching a community-wide, mass media campaign for 12 months using newspaper advertisements, radio announcements, and focus groups to develop culturally tailored brochures and posters. They used social learning theory (SLT), the health belief model (HBM) and the trans-theoretical model (TTM) learning theories in the design of the intervention. For a one year period, Khankari (2007) and his colleagues focused on educating individual patients and improving physician communication skills. Patients were mailed a brochure (Screen for life) and a letter explaining the importance of CRC screening just before their scheduled visit, while the physicians prepared for the visit by attending a one hour educational session to review the CRC screening guidelines and to improve their communication skills. The Friedman and Borum (2007) study was the only one whose intervention was entirely focused on education for the physicians. Using a quasi-experimental before and after study design, they incorporated didactic seminars and case-based learning sessions to educate Internal Medicine residents on appropriate screening recommendations.
Basch (2007) and associates educated their study participants by delivering tailored telephone interventions. During a 12-month period, an average of five phone calls were made to selected members in a New York City community, while simultaneously mailing a brochure (Lets Break the Silence) and a letter to the controls welcoming them to the study. Campbell (2004) and his associates decided to randomize and offer education to different African American churches to implement an intervention termed, "Wellness for African Americans Through Churches (WATCH)", instead of randomizing individual subjects. Campbell et al. (2004) divided the participating churches into four different groups. For some churches the intervention involved printed newsletters and videos (TPV) every 2 nd , 4 th , 6 th , and 9 th months. Some churches were given a lay health advisor, and other churches received both interventions. The control churches were offered health education on a variety of subjects other than CRC screening. The HBM, social cognitive theory (SCT) model and social support model (SSM), plus the stageof-change, and the trans-theoretical framework were learning theories used by Campbell (2004) and his associates. Powe (2002Powe ( , 2004 and her associates used the same 12-month, multi-component, educational intervention in both studies targeted towards educating elderly African Americans in senior citizen centers. The senior citizen centers were randomized into three groups. The first group of centers received a culturally tailored video (Telling the Story…To Live is God's Will) that discussed CRC prevention from a spiritual prospective. Acalendar, (Telling the Story), a brochure (Telling the Story… About Bowel Cancer), and a poster (Can We Talk…About Bowel Cancer) and a flyer (Doing the Kit) was also used. Another group of centers received a modified intervention which was just the culturally tailored video (Telling the Story…To Live is God's Will). The control group of centers was shown a non-culturally specific video developed by the ACS (Colorectal cancer. The Cancer No One Talks About). Each component of the intervention used images, language, food, dress, and customs that were familiar to the study population (Table 2supplemental material). Katz and Fisher (2012) randomized patients into 2 groups: the control group received CRC screening education and the intervention group received CRC screening education in addition to patient activation and barrier counseling. The intervention group watched a 12-minute video entitled "Ask your doctor about colon cancer screening", received a brochure that supplemented the video and focused on asking their provider for a CRC screening test, and received a second brochure on tips to prevent CRC (e.g., the importance of daily exercise). The patient activation section was not included in the video for the control group participants. One month after the medical visit, if a patient in the intervention group had a CRC screening test ordered but not completed, telephone barriers counseling to address patient identified CRC screening barriers was conducted. If a patient in the intervention group did not have a screening test ordered, telephone barriers counseling focused on activating them to ask for a CRC screening test by calling their provider or asking their provider at their next medical visit. Xirasagar et al (2011) compared colonoscopy rates among established patients of purposively sampled African American PCPs, both trained and untrained in colonoscopy. It was a retrospective observational study based on patient chart review of age-eligible patients of colonoscopy-trained African American PCPs (or those in-training) and patients of untrained PCPs practicing in the same geographic region. This study also compared pre-training and post-training colonoscopy rates among the PCP's patient panel. Siddiqui et al., (2011) randomly assigned participants to 1 of 4 study groups as follows: a control group (received standard care), a standard intervention (SI) group (received a mailed standard intervention i.e., a mailed CRC screening invitation letter, informational booklet, a stool blood test [SBT], and reminder letter), a tailored intervention (TI) group (received the standard intervention plus 2 tailored message pages that included brief messages addressing personal barriers to screening reflected by responses to Preventive Health Model variables measured on the baseline survey) or a tailored intervention plus reminder phone call (TIP) group. Blumenthal et al (2010) objective was to test 3 interventions to increase CRC screening rates among African Americans. Two of the interventions were educational, and the third responded to financial barriers. The educational interventions were consistent with the contention that education could increase screening. Participants were randomized into 4 cohorts: control, one-on-one education, group education and reduce out of pocket expense (participants in financial support group were offered financial reimbursement up to $500 for out-of-pocket expenses incurred for CRC screening). All eleven studies had multiple components to the interventions that were offered to the experimental groups. Six of the studies had baseline data collection, the intervention and then follow-up data collection one year after the baseline [15][16][17][19][20][21]23] , and one study had baseline data collection, the intervention, and then a six-month follow-up data collection [18] . Four studies required community cooperation to deliver the proposed intervention(s) [15][16][17]19] . The Campbell group (2007) and Siddiqui et al.,2011 used telephone outreach as the primary intervention, while Katz and his group of researchers were the only ones who used a mass-media campaign to get the message across to their target population. Two studies used specific theories to guide the delivery of their interventions [17,19] . Only Four studies were culturally tailored [15][16][17]19] .

Interventions results
In all of the studies, there appeared to have been an increase in the percentage of CRC screening participation in the intervention groups compared to the control groups. Eight had statistically significant results [15,18,[20][21][22][23][24] . Both studies by Powe andassociates (2002, 2004) had many parts to the intervention, and they both seemed to show an increase in CRC from the baseline. However, the p-value was reported only in Powe (2002), and it was significant among the groups (chi-square =29.37; df= 2;P < .0001). Although CRC screening participation was low after the intervention in the experimental group (27%), Basch (2006) and his colleagues reported that the participants who received the telephone intervention had CRC screening rates that were 4.4 times higher than the controls (95% CI 2.2-7.7) after one year. The Friedman and Borum study (2007) showed a statistically significant difference in pre-and post-intervention CRC screening using the endoscopic procedure (26.7%, n=31 vs. 59.1%, n=78; p=0.0001) but not screening rectal exams (41.4%, n=48 vs. 38.6%, n=51; p=0.6605) or FOBT (33.7%, n=46 vs. 39.7%, n=50;p=0.7748). The Wellness for African Americans Through Churches (WATCH) project reported an apparent increase in CRC screening intervention group; however, it was not statistically significant [17] . Katz (2007) and his team also reported an increase in CRC screening, but there was no statistical difference from baseline (baseline=49.3% (95% CI 46%, 52.6%) vs. a one year follow-up of 55.6% (95% CI, 47.9%, 63.0%; p=0.49). Katz  Although the financial support cohort and the One-on-One cohort rates exceeded that of the control cohort but they did not differ statistically. Of the four culturally tailored interventions [15][16][17]19] only Powe, 2002 reported a statistically significant increase in CRC screening. In all of the non-culturally tailored interventions statistically significant increases were reported in CRC screening [18,[20][21][22][23][24][25] . It is important to note that three of these studies only used insured participants [18,20,21] . These studies had the best design and received a design score of 1 (table 1) since individuals were randomized into study groups and control groups and they were not culturally tailored. Three culturally tailored studies, classified as community trials, used randomized locations instead of individuals [15][16][17] and one study classified as randomized community trial [24] were assigned a study score of 2. The designs for the other four studies [19][20][21]23] had the lowest study design score of 3, the weakest category, as they did not have a control group.

Discussion
Some investigators have suggested that increased knowledge and a realistic perception of CRC screening among African Americans could increase CRC screening rates [26][27][28] . While others suggested that culturally tailored educational interventions may be more successful at increasing the rate of CRC screening in African Americans compared to traditional outreach [29,30] . Although culturally tailored interventions are thought to be successful, there has not been a thorough review of the effectiveness of interventions that aim to increase CRC screening in the African American population. The current treatise is a review of recent studies whose interventional aims were to increase CRC screening in the African Americans. Individual-centered educational interventions (tailored or non-tailored) were found to be more successful than the interventions focused on educating entire communities. This current review suggests that increased provider education and training in communication skills play an important role in elevating CRC screening rates in African Americans, since this resulted in greater increases in CRC screening rates than community-based interventions, but does not eliminate or support the need to continue to develop more culturally tailored interventions to increase CRC screening in African Americans; however it does suggest that educational and communication gaps may be wider than the cultural gap.
Community-wide interventions may play a role in efforts to increase awareness in African American communities; however, awareness may not immediately translate to increased CRC screening participation [31] . The lower success of community intervention trials, compared with individual-centered provider-education interventions could be explained by several different factors. One could be that community-wide interventions may lack the inter-personal communication and interaction needed to properly inform someone about their individual risk and need for CRC screening. Also, the community intervention in the analyzed studies did not always provide an avenue for individuals who were interested in CRC screening but did not have the means. The interventions aimed at increasing physician awareness may have been successful for several reasons. One reason could possibly be that these studies used chart reviews to identify patients who had not been previously screened for CRC but were eligible for screening and had the means to get screened. Secondly, these studies use a pre-and post-test study design, which is an inferior study as compared to randomized community trail study design used in the culturally tailored studies. Most importantly, several other authors have documented the importance of the recommendations by physicians for CRC screening in African American patients [32][33][34] .
A limitation of this current review is that the participants in the selected studies were mostly women. Because of decreased participation of African American men, the benefit of these interventions may be limited to African American females. Another limitation is that the interventions that could be evaluated were heterogeneous and each one had multiple components; a fact that limits any conclusion about what part of the intervention was the most effective, or if one component was not effective at all. Because of the inferior quality and poor methodological quality of some of the studies available for the review, it is difficult to make any definitive assumptions about the effectiveness of recent interventions designed to increase CRC screening in African Americans. More research is needed on the effect of culturally tailored interventions on African American CRC screening. There is also a need for collaborative research involving the community and health care providers. Further research is necessary to identity the role (if any) of standardized education interventions for patient and primary care providers.