Colorectal cancer prevention for low-income, sociodemographically-diverse adults in public housing: baseline findings of a randomized controlled trial

Colorectal cancer (CRC) is the 2^nd leading cause of cancer mortality in the US [1]. Over 49,000 deaths and over 100,000 new CRC cases are projected in 2008 [2]. While absolute CRC incidence and mortality rates have declined over the past 15 years, racial/ethnic minorities, and African Americans in particular, continue to have higher incidence and death rates compared to whites [2, 3]. Recent evidence suggests that, in part because mortality rates among whites are dropping, the difference in CRC deaths between African Americans and whites has increased over the past 15 years, thus contributing to an increase in cancer-related health disparities [3]. Although rates for deaths attributable to CRC for Hispanic men and women have decreased or remained flat over this time period, similar to African Americans, they are more likely to be diagnosed with late-stage CRC and have lower survival rates compared to whites [4].

It is well-established that the majority of US cancer deaths, about 60%, could be prevented by improving lifestyle factors: reducing smoking and obesity, improving diet, increasing physical activity, and participating in screening for early cancer detection [3]. In particular, low rates of physical activity and participation in CRC screening significantly contribute to CRC incidence and death [5]. Nationally, less than 50% of adults meet physical activity recommendations of 30 minutes of moderate-intensity activity on 5 or more days per week [6]. Data from the 2006 Massachusetts BRFSS show that low-income adults (< $15,000/yr), African Americans (37%) and Latinos (44%) are least likely to meet these recommendations [7]. In addition to the aforementioned lifestyle factors, there is some evidence of a relationship between multivitamin use and chronic disease prevention. Studies show that on average, 52% of adults take multivitamins with lower intake among racial/ethnic minorities (26 - 36%) and those with less than a high school education (29-35%) [8, 9]. Recent clinical reviews of this relationship are unclear as to the overall benefit of multivitamins in reducing risk for chronic disease, such as CRC, [10] but studies show it may be helpful [11, 12].

It is estimated that screening alone can significantly reduce CRC incidence and mortality rates [13, 14]. Yet, findings from national surveys show that almost half of all adults do not meet CRC screening guidelines [15]. For adults age 50 years and older, screening options include an annual fecal occult blood test (FOBT), sigmoidoscopy every 5 years, colonoscopy every 10 years, or double contrast barium enema every 5 years. Nationally, roughly 16% of adults have had a FOBT in the past year, with a greater majority receiving either an endoscopic screening test (i.e., sigmoidoscopy or colonoscopy) within the past 10 years (56%) [16]. Current CRC screening rates are highest for whites (62%) and lower among blacks (58%) and Latinos (46%); 57% of those insured are currently screened, compared to 25% among the uninsured. There are also large disparities in CRC screening rates based on educational attainment. In 2004, only 31% of those with less than high school education were currently screened, compared to 54% of college graduates [17]. In general, screening rates are increasing across all racial/ethnic groups; however, there remains a need to increase access to CRC screening for underserved populations [16].

Racial/ethnic minorities and low-income populations report significant barriers to engaging in health promoting behaviors and getting screened for cancer. Sociodemographic characteristics and access to health care indeed determine patterns of engaging in cancer preventive behaviors. In addition to these long-established correlates of CRC prevention, social contextual factors are increasingly being examined as important influences on cancer prevention behaviors, particularly among underserved populations [18]. As part of our ongoing work to develop effective cancer prevention interventions for multi-ethnic, low-income populations, we have developed a conceptual framework for incorporating social context into the design and evaluation of interventions targeting risk behaviors (Figure 1). This model, based on the social ecological framework, draws on a range of social and behavioral theories to explicate the pathways by which social context may influence health behaviors. Social contextual factors cut across multiple levels of influence and include individual factors (e.g., material circumstances, psychosocial factors), interpersonal factors (i.e., social ties and family roles and responsibilities), organizational/systems factors (i.e., access to health care), and neighborhood and community factors (i.e., access to safe neighborhoods, transportation). These social contextual factors are shaped by a range of socio-demographic characteristics (i.e., social class, race and ethnicity) which taken together influence one's day-to-day realities and shape patterns of health behaviors and health outcomes.

Culturally appropriate, community-based interventions that focus on the day-to-day lives of socially and economically deprived populations are lacking. Public housing is increasingly being recognized as an important community-based setting for intervention delivery [19]. Public housing is comprised of low-income families, the elderly, those with disabilities, and large numbers of racial/ethnic minorities--populations that are at increased risk of poor health status. Currently there are 1.2 million households in public housing in the US [20]. Few interventions have specifically addressed the needs of those living in public housing. The few that have been conducted found that public housing residents are more likely to be obese, have chronic conditions, i.e., hypertension and diabetes, smoke, and be physically inactive [21]. However, low-income public housing environments also represent opportunities for public health improvement. It is a setting with great potential to affect the health of residents through positive mechanisms such as their peers, who are also their neighbors, and other social support networks that enable residents to share health-promoting information or easier access to health-related services such as food banks or health clinics [22]. Public housing residents are largely an insured group, given the income levels necessary to qualify for public housing [19, 23], which may address a major barrier to screening, particularly in a state like Massachusetts where the state Medicaid program covers screening. Further, residents tend to live in these communities for a long period of time [24, 25], thus making it also a relatively stable residential environment in which to conduct outreach.

Open Doors to Health is a theory-based peer-led intervention aimed at increasing CRC prevention behaviors in residents of low-income, ethnically diverse public housing communities. It sought to capitalize on social networks that exist within public housing and employed a peer-led, i.e., lay health advisor (LHA), model that used members of the target group for intervention delivery. Several behavioral studies have utilized peer-led models to increase the impact, reach, and sustainability of the interventions [26‐29]. Typically in such peer-led interventions, members of the target population are selected, trained, and supported as they implement the intervention to promote changes in cancer prevention behaviors. This model has been employed and tested most extensively in low-income, ethnically diverse groups who carry a disproportionate share of the cancer burden. The majority of peer-led cancer prevention interventions have focused on increasing rates of cancer screening and/or target other cancer-related health behaviors. Paskett et al. [29] trained LHAs, individuals known and trusted in the community, to deliver health information to increase mammography utilization among a diverse sample of rural women. Compared to a control group, women in the intervention group were more likely to have received a mammogram at 12-month follow-up and reported reduced barriers for screening. Resnciow et al. [30] conducted a LHA intervention to increase fruit and vegetable intake among African American church goers and found a statistically significant increase in intake relative to controls at 6-month follow-up. These studies found that a peer-led model can be successfully implemented in racial/ethnic minority communities and result in significant behavior change over and above standard self-help interventions.

Relatively few studies have tested the peer-led model in interventions aimed specifically at CRC prevention, and those that have had limited impact on prevention behaviors. Campbell et al. [27] conducted an intervention trial testing the efficacy of using LHAs compared to the use of tailored print information for improving CRC screening, diet, and physical activity behaviors among African American church members. Findings showed that the LHA intervention arm failed to demonstrate effectiveness in any of the target health behaviors, likely due to limited reach of the LHAs to the target audience. However, church members who did talk with a LHA were twice as likely to obtain CRC screening compared to those who did not talk to a LHA. Few studies have evaluated the impact of the lay health advisor or peer leader model in the context of implementation within low-income housing, and to our knowledge none have addressed CRC screening in this setting.

The goal of the Open Doors to Health intervention was to build upon the strengths of using a social contextual framework to improve CRC risk factors in low-income public housing residents. Using a peer-led model, we sought to work through the already-existing social networks in the housing site to address individual and interpersonal factors, as well as neighborhood and organizational factors that make adopting healthful behaviors more challenging for racial/ethnic minorities and lower income populations. The peer-led model also fosters service delivery sustainability such that after the intervention period is over trained LHAs will retain the ability to provide advice and assistance to housing residents. The purpose of this paper is to describe the Open Doors to Health study design and intervention components, and to present the demographic characteristics of the study population by age (under 50 years old/age 50 and above) and in relation to our conceptual model. This randomized controlled trial was designed to increase CRC screening, physical activity, and multivitamin use among low-income public housing residents; a key aim is to determine ways in which attending to the social context in low-income, ethnically diverse populations may address disparities in CRC preventive behaviors.

Cancer-related health disparities of incidence, morbidity, and mortality differentially affect racial/ethnic minorities and lower income groups. Recently, there has been an increased awareness of the relationship between poorer health outcomes in these groups and the social, economic, political, and cultural environments in which they live. This is particularly true for persons living in low-income public housing, environments which are often located in distressed neighborhoods with few health promoting resources. Public housing residents report poor health status, chronic disease conditions, and poor health behaviors [23, 51]. Thus it is imperative that health promotion interventions targeted and tailored to racial/ethnic minorities and low-income groups consider this constellation of risk factors in order to be effective at closing the health disparities gap. Baseline findings from Open Doors to Health (ODH) suggests that segments of this population, i.e., Hispanics and unemployed adults, are at even greater increased risk for CRC given their low socioeconomic position and low levels of physical activity.

Data from the 2006 Behavioral Risk Factor Surveillance System show that on average 91% of Massachusetts adults had health insurance--this figure is substantially lower for blacks (81%), Hispanics (76%), and those with low incomes (86%) [7]. In contrast to national estimates, the majority of ODH study participants had health insurance coverage (96%), either through public, private, or a public+private; this population most likely qualified for health care based on income and/or age requirements. This may in part explain the high rate of CRC screening in this population, in addition to CRC testing as a covered Medicare benefit nationally and as a covered Medicaid benefit in Massachusetts [15]. Comprehensive health insurance is indeed a benefit for those seeking to obtain CRC screening given the often high co-payment associated with some of the screening tests. Nationally, about 60% of adults are currently screened for CRC [16]; among our participants 66% of those age-eligible for CRC were up-to-date on their screening [50]. These findings indicate a strong relationship between access to care and screening uptake; however with almost universal health coverage, 34% of those over 50 years old were not CRC current. This indicates an opportunity for improvement and a need to better understand the mechanisms by which social contextual factors, such as the cost of co-payments, that might hinder participation in screening among those with insurance. Health insurance makes obtaining screening more financially feasible, and facilitates access to other health promoting services. However, it can be argued that certain behaviors, such as physical activity or multivitamin use are heavily influenced by individual and broader social contextual factors, such as perceived benefit, cost, time, and access to resources.

Overall, daily, regular multivitamin use was very low in this population (28%) and even lower among those under 50 years old (17%); however, use among Hispanics regardless of age (16% for Hispanics under 50 years; 31% for Hispanics age 50 and up), was much lower than all other racial/ethnic groups. This finding highlights an opportunity for intervention. Understanding and resolving barriers to multivitamin use among Hispanics may help to reduce CRC risk in this group. A 2001 study found that multivitamins use is predictive of CRC screening among women--women who took a daily supplement were 1.74-2.12 times more likely to be current with their CRC screening [52]. The inexpensive, simple act of taking a multivitamin could potentially be a cue to engaging in other health promoting behaviors. Our prior research studies focused on increasing multivitamin use among racial/ethnic minorities and lower SES groups found an almost 30% increase in multivitamin use at follow-up [45]; this increase could result in a significant (> 25%) reduction in CRC incidence as well as be an important facilitator for the reduction in other conditions such as cardiovascular disease, osteoporosis, and birth defects [53, 54].

For good health, current physical activity recommendations are for adults to engage in at least 30 minutes of moderate-intensity physical activity on at least 5 days per week--the equivalent is the accumulation of 10,000 steps per day [49]. On average, study participants had levels of activity (~5600 steps/day), far lower than recommended levels [43]. In comparing racial/ethnic groups, blacks and Hispanics were the least active, and physical activity was lowest among blacks over 50 years old (mean = 3968 steps/day), indicative of a sedentary lifestyle and increased risk for chronic health conditions and poor health outcomes. Unemployed older adults had the lowest physical activity levels overall (3691 steps/day), even lower than those disabled (4041-4710 steps/day). Studies have shown a similar relationship between unemployment and physical inactivity, possibly due to the absence of job-related or transportation-related activity [55]. They may also be more likely to have other health conditions, i.e., obesity, and comorbities that increase sedentary behaviors [56]. Fewer studies focus on these populations (i.e., disabled, unemployed) in an effort to increase their activity levels.

In Open Doors to Health, we were focusing on salient social contextual factors that may have an impact on intervention effectiveness and health behavior change, such as interpersonal and neighborhood factors. Although the majority of participants felt that their neighborhoods were safe in the daytime (85%) and agreed there were many places to go within walking distance of their home (84%), the overall mean steps per day was low. Although unable to be determined from baseline findings, it is possible that many participants have free time in the evening and would choose to engage in physical activity at that time, but feel their neighborhood is unsafe at night (63%) [57]. Participants also reported high social support from family and friends, but moderate social networks, suggesting that although the breadth of their networks may be somewhat limited, they do perceive substantial support from their social networks. This has important implications for conducting health behavior change interventions in this group. For example, women of color are more likely to engage in physical activity with others than alone, and express need for social support to be active [58, 59]. In ODH, we worked through their social networks and focused on social support to increase intervention reach and effectiveness. We sought to positively modify social norms for screening and physical activity, which are both currently very low, through on-site activities delivered and based in their community.

Targeting and working within the cohesive social structure of public housing sites in combination with efforts to improve the housing site environment represents an assets-oriented approach to improving health behaviors of public housing residents. Similar to interventions that have been effective in improving the health of worksite employees, [18, 46] in this intervention, we sought to build upon the resources and assets, i.e., residents and management, already available in the housing sites. In fact, resident service coordinators at many of the housing sites already helped residents obtain health insurance and make medical appointments prior to the intervention, likely contributing to the high rates of CRC screening in this sample. Moreover, this also represents a potentially sustainable intervention as CRC prevention information and resources are retained at the level of the housing site and can represent a long-term strategy for improving the health behaviors of racial/ethnic minority and low-income groups with health insurance.

Several study limitations and strengths should be addressed. We achieved a response rate of 53% which ranged from a low of 34% to a high of 92% across the housing sites. Nevertheless, we targeted, recruited and enrolled 1554 participants, which represents a large, ethnically-diverse underserved population. This study has limited generalizability to other populations other than low-income, urban, racial/ethnic minorities living in public housing. In an effort to accurately and objectively report CRC screening rates in this population, we also validated self-reports of CRC screening with medical record verification in a sample subgroup.

Open Doors to Health was designed to better understand and address the needs of low-income minorities in public housing to incorporate the complexities of their lives into a socially-and culturally-appropriate CRC prevention intervention. Intervention activities were specifically designed to target modifiable factors, such as social networks, neighborhood social cohesion, and access to health promoting resources (i.e., developing walking clubs) in an effort to strengthen both individual and organizational capacity for reducing cancer risk factors [18]. Few studies have conducted community-based studies in public housing communities; these data suggest areas for improvement and future opportunities for intervention development and dissemination. In particular, the high rate of participation in some CRC prevention behaviors (screening), and low participation in others (e.g. physical activity), suggests the importance of both health insurance coverage and access to preventive health programs, as all risk factors are not likely to be addressed by insurance alone. Findings from the randomized trial will determine the effectiveness of the intervention on our health-related outcomes as well as inform future avenues of research.