Abstract
The Affordable Care Act of 2010 requires all nonprofit hospitals in the United States to conduct a Community Health Needs Assessment (CHNA) at least every 3 years. With this law in its infancy, the best practice to conduct an assessment that complies with the law is unknown. Research designs vary across states and agencies, and little is known about the reliability or representativeness of results. The rural community group model (RCGM) is a newly developed model designed for conducting assessments in rural communities. Key components of the model are disseminating surveys, conducting key informant interviews, facilitating focus groups, and integrating secondary data of county-level health behaviors and outcomes. It has been used to conduct CHNAs on more than half the critical access hospitals in North Dakota (58%). Given this large sample size, which used the same methodology, this article provides an evaluation of the model focusing on lessons learned and challenges encountered in the conduct of CHNAs. Particular strategies for assessment planners are warding off group think, monitoring against bias creep in data collection, and integrating multiple data sources to inform decision making. The model is recommended for replication in rural settings to provide meaningful feedback that allows a hospital to match long-term planning with community needs.
Keywords
In early 2010, the Affordable Care Act (ACA; Pub. L. No. 111-148) put in place comprehensive reforms designed to improve access to affordable health coverage. One of the mandates of the ACA requires all nonprofit hospitals to conduct a community health needs assessment (CHNA) once every 3 years and adopt an implementation strategy in order to meet the identified needs. Since these nonprofit hospitals were founded to meet the most pressing health needs of their communities, a vital purpose of the CHNA is to effectively assess community health needs for hospitals to develop implementation strategies to address those needs. Newly passed into law, the ACA mandate is in its infancy, and little has been written about the best methodology to conduct a CHNA.
A review of methodology currently in practice reveals that designs vary from state to state, depending on who is conducting the assessments (hospital or private consulting firm), size and location of community (urban or rural), and time line (3 months–1 year). Although many CHNA tool kits exist, few apply to rural areas (assessment.communitycommons.org). Hospitals that fail to meet the CHNA requirement are subject to a $50,000 excise tax (Internal Revenue Service, 2011). Given that these CHNAs will be routinely conducted and have a wide-ranging scope, affecting nonprofit hospitals in every state, there is critical need to review research designs and share best practices.
Models are needed that allow for effective data collection that representatively and accurately reflect community health needs. Furthermore, the model should employ preferred methods of information dissemination and informal communication networks prioritized in rural communities. The Center of Rural Health at the University of North Dakota School of Medicine and Health Sciences has developed their own assessment research design, called the Rural Community Group Model (RCGM), which is tailored to the unique context of rural communities. This mixed-methods design has been used to conduct CHNAs on 21 of the 36 critical access hospitals in the state (58%) and has yielded valuable community feedback representing broad community demographics. This article reviews the RCGMs methodologies, offers lessons learned, and identifies challenges.
Community Health Needs Assessment
Although the ACA mandate to conduct a CHNA is new, the practice of assessing community health is not. Conducting a CHNA is a common and enduring practice in health care that provides a systematic approach to identify inequalities in health services and determine priorities for the most effective use of resources (Wright, Williams, & Wilkinson, 1988). The purpose of a CHNA is to not only describe the health of a community but also pinpoint health gaps and trends that need to be addressed; once needs are identified, decisions can be made to directly enact community change (Witkin & Altschuld, 1995). Therefore, obtaining information that is based on valid, reliable data is essential for the analysis and objective evaluation of a health situation, evidence-based decision making, and programming in health (World Health Organization, 2001).
Overview of Rural Community Group Model
The RCGM is based on a model developed by the National Center for Rural Health Works (ruralhealthworks.org). Similarly, it depends on a steering committee comprised of local leaders (three to four members including hospital, public health, and other nonprofit employees) to spearhead the assessment process. Using locals is paramount to gaining entrée into rural communities where preference is given to insiders and outsiders may be perceived with caution (Little & Miller, 2007). Moreover, relying on local leadership establishes the CHNA as a collaborative process where research is done with community members (Lincoln & Guba, 1985).
It differs in that it uses a mixed-methods design, gathering community perceptions of health concerns in multiple ways: (a) distributing a survey to community members and health care professionals within the hospital’s service area, (b) interviewing key infor-mants, (c) facilitating broad-based community focus groups, and (d) collecting secondary data from multiple sources. This robust design adheres to the tenets of rigor in qualitative methods and produces a holistic picture of community health needs (Lincoln & Guba, 1985). Once this information is collected and distilled, the researcher shares it with a community focus group, which then prioritizes perceived community health needs. The list of prioritized needs serves as a guide for strategic planning and implementation. In this way, the RCGM uses a bottom-up information process that allows community members’ input to guide decisions, thereby emphasizing their role and voice in the community (Figure 1).
Bottom-Up Decision-Making Process of RCGM Assessment
The strength of the RCGM lies in its ability to invite community involvement in efficient and anonymous ways so that community members can respond openly and honestly, but it does require more time and resources. To understand the flavor and dynamics of small towns, site visits are a must. Due to the remote locations and amount of travel needed, it would be tempting to videoconference with interviewees. However, the RCGM recognizes the impact that face-to-face interaction has in building rapport, and helps outside researchers to be accepted into the community. The RCGM involves two site visits that involve interaction with dozens of community members and takes approximately 3 to 6 months.
The following sections present the key methods in the RCGM process, highlighting how they are responsive to rural communities and inform the strategic plan.
Surveys
A survey template is developed to understand the community’s use of, concerns with, and gaps in health services; solicit suggestions for better health care delivery; and collect respondent demographic data. The survey is designed with the hospital’s CEO and can be customized for each community’s unique needs. Although some CHNA consultants opt only for key informant interviews or surveys, the RCGM stresses the importance of using both as the survey allows for anonymous and confidential responses. The opportunity to dissent from the majority is especially significant in small towns where overlapping relationships and familiarity may prevent candidness (Townsend, 2007). The open-ended survey questions elicit many colorful remarks and often address conflict that is not addressed in the other methods.
Two versions of the survey are distributed to two different audiences: (a) community members and (b) health care professionals. To disseminate this survey into the broader community the RCGM relies on the steering committee. As longtime residents, they know best which locales receive the most foot traffic and which target underserved and minority populations, and leave surveys accordingly, including at bars, bowling alleys, churches, and gas stations.
Depending on the size of the community, 500 to 1,500 community member surveys are distributed in the service area. Area residents also can complete an online version of the survey, with the link publicized in the local media and on the hospital’s website; however, most rural residents prefer print copies. Of the total number of completed surveys, typically 80% per community are completed in print. The health care professional survey is available only online with the link copied on employee pay stubs and posted at staff lounges.
Interviews
Word of mouth is often the preferred mode of communication in rural communities, and community members may wear many hats, sitting on multiple boards and holding various elected positions (Prialatha & Malar Mathi, 2012). Therefore, it is important to invite community leaders who are heavily involved in the local community to participate in the assessment as they can help spread the word through their various networks. To promote local buy-in and highlight the community aspect of this assessment, the steering committee invites participants via a letter or phone call. In small towns, trust is the currency. Having locals extend the invitation engenders support, encourages participation, and sustains involvement in the assessment process.
One-on-one interviews are conducted with 6–8 key informants who represent a wide demographic and bring different perspectives; they may include representatives of the health, business and faith communities, nonprofit agencies, political bodies, and tribal leaders and must include a public health professional. Interview questions focus on perceived health needs of the community, awareness, use and delivery of local health services, and barriers to using local services.
Focus Group
In rural communities, an informal gathering of community members over dinner provides an opportunity to share concerns and validate others’ perceptions. The same questions posed during key informant interviews are asked, but the answers vary drastically. Focus groups take advantage of the group effect where shared experiences yet different interpretations provide a base to form the discussion (Lindlof & Taylor, 2011). The steering committee invites members based on their representation of diverse demographics and interaction with marginalized populations. With the understanding that those best able to access members of marginalized populations are their own peers, the RCGM asks focus group participants to take a box of surveys to pass out to their neighbors, friends, and coworkers.
Secondary Research
During early stages of the assessment process and while surveys are being collected, researchers collect secondary quantitative data that present an overview of the hospital service area’s health outcomes, behaviors, and indicators. The RCGM relies heavily on county-level data collected from more than 20 national sources by County Health Rankings, an initiative of the Robert Wood Johnson Foundation in collaboration with the University of Wisconsin Population Health Institute (www.countyhealthrankings.org). Data are also collected from North Dakota Health Care Review, Inc.; North Dakota Department of Health; KIDS COUNT; U.S. Census; National Survey of Children’s Health; and Gallup-Healthways Well-Being Index to convey children’s health, preventive care, and general area demographics.
Community Group
After data have been collected and synthesized, the researcher presents findings at a community group. To facilitate the discussion of prioritizing the needs, the researcher prepares an initial list of all potential health needs identified from the various data sets, which serves as a ballot for participants to rank their five most pressing community health needs. To narrow the list of 20 to 30 potential community needs, participants are encouraged to think of the needs in terms of greatest severity, impact, and magnitude on community health (Green & Kreuter, 2005). Once the votes have been tallied, the group is asked to review the final list to see if the needs accurately depict the community’s health needs. Some communities opt for a second ballot to differentiate between ties or reprioritize the rankings. The final ranked list is then submitted to the hospital CEO to use for strategic planning and to satisfy the objectives of the ACA.
Lessons Learned
After conducting 21 CHNAs using the same model and interacting with hundreds of community members, it is important to reflect on what has worked and what needs adjustments in the design, especially given the close ties and informal communication networks of rural communities. Group think is especially germane in rural settings where focus group participants know each other and may be more likely to represent homogenous thinking (Lindlof & Taylor, 2011). Additionally, a stigma is often perceived if community members speak against a community asset and economic generator, such as the hospital (Commins, 2012). One tactic the RCGM uses to counter group think is to promote pluralism. The Center for Rural Health researcher sends the steering committee a spreadsheet with various sectors of the population represented such as agricultural, faith, retired, family with young kids, and elected official. The steering committee must ensure it is selecting participants from each sector and then submit it to the researcher for review. The spreadsheet provides a visual display of demographics, promotes transparency, and helps discourage convenience sampling.
Although the analysis of results is a laborious process and time constraints may limit researchers to include only dominant themes, the RCGM recognizes the value of negative case analysis where researchers search for cases that refute, rather than affirm, the emergent findings (Lincoln & Guba, 1985). This process is difficult because it puts the onus on researchers to play devil’s advocate; however, it is essential to pursue rival explanations in order to claim that the conclusions are credible. This concept has relevance to the RCGM findings because, oftentimes, the most pressing concerns identified in the primary data differ vastly from the secondary data. For example, in one community the secondary data revealed screening rates for colorectal cancer, pneumonia, and influenza were significantly lower compared to state averages and national benchmarks. However, these concerns were never articulated as an expressed need by community members. Although it would be easy to disregard this data as nonrepresentative, it is important not to dismiss one data set. Additionally, using multiple methods enables the researcher to triangulate findings and examine issues of disparity.
Challenges Encountered
Although the RCGM tries to invite broad community participation that is representative of diverse demographics, it is by no means a representative sample. Using key informants volunteered by the hospital CEO may cause the participants to feel under pressure to present a particular picture of the hospital and the community health needs. The survey sampling method uses convenience and snowballing sampling. Both of these may produce nonrepresentative information. Although snowball sampling works well with a hard-to-recruit population, it produces nonprobability sampling and may segment the audience, thinking their views are representative of a random sample (Biernacki & Waldorf, 1981). To increase recruitment and retention, financial incentives have been encouraged in rural areas and may warrant future application (Brownell, Bishop, & Sindelar, 2005).
Although urban CHNA models use random sampling by phone surveying or mail distribution, this method has poor response rates in rural communities as it lacks community buy-in (Edelman, Yang, Guymon, & Olson, 2013). Small-town communities often pride themselves on their close-knit ties, and outsiders are met with skepticism. Therefore, using locals to pitch the assessment process has reframed the CHNA to showcase community members identifying and then addressing community problems. This shift emphasizes empowerment and helps facilitate change at a grassroots level. As a result, community members are more invested in improving community health and take more ownership in implementing the strategic plan.
Implications for Practice
The RCGM offers a new model for CHNAs that caters to the dynamics of rural areas where trust, word-of-mouth referrals, and insider status are valued. Using local leaders to serve as champions helps increase engagement and foster collaboration. The multiple methods of data collection target a broad audience, and the anonymity of the survey allows for candid responses, thus opening the dialogue between the hospital and the residents it serves. The community input serves as a bottom-up approach to decision making for resource allocation, program planning, and program development. Not only does this template ensure regulatory compliance and provide meaningful community feedback, it also allows hospitals to match long-term planning with community needs, thereby setting the stage for effective strategic implementations. Moreover, it offers a framework that can be customized to tailor specific community demographics and situations. The results identified by the RCGM play an influential role in the health of the community. Therefore, the RCGM serves as a data-gathering tool to inform health systems strategic planning and address priority health needs.