limitations of the study
The Four-County Community Health Needs Assessment was designed to provide a comprehensive baseline assessment of Northwest Arkansas citizens' perceived health care needs. It was designed specifically to capture the perspectives of the community members themselves, rather than the perceptions of the area's health care providers and professionals. This assessment was also designed to identify perceived needs in the health care delivery system and to report its findings to the Northwest Arkansas community region in the form of a community health report. This information will be utilized to improve the quality of health care in Northwest Arkansas through public health care policy, education and outreach efforts. It will also be used in the development of essential health care services, both for the general public and underserved groups within each community.
The community framework employed in this study represented both a significant strength and a significant limitation for the Northwest Arkansas Health Needs Assessment. This collaboration consisted of a coalition of the Hometown Health Improvement projects within Benton, Carroll, Madison, and Washington counties. Each county's health project served as part of the planning, oversight, and operational backbone of the project, significantly strengthening the study. This collaboration was not only consistent with the "community imbued" perspective of the health needs assessment, but it also provided the opportunity for an innovative and unique partnership between the Four Hometown Health Improvement Projects in Northwest Arkansas. The collaboration that developed between the Hometown Health Improvement Projects personified the community spirit inherent in this Health Needs Assessment and provided the means by which this ambitious project could be successfully completed.
However, this same process also provided significant challenges, which manifested themselves as limitations, to the study. Community research is a collaboration between individuals, groups and entities within a community and is designed to study or assess a phenomenon in a field-based setting. This process presumes that the goals, purpose, and methodology employed by the research project are clearly articulated, agreed upon, and empirically based. It also presumes that effective working relationships, roles and responsibilities have been clearly delineated, and that the results and recommendations that will be derived from the study will be consistent with the stated goals and objective of the project.
The research design for the Health Needs Assessment employed a multimethod data collection scheme, which consisted of the Health Needs Survey and four targeted focus groups. Survey data was collected utilizing a sample of convenience, and data collection sites were identified by members of the Four County Hometown Health Coalitions, and in conjunction with collaboration provided by the Social Work Research Center. While the sample of convenience and data collection sites employed in this study permitted a large number of individuals to be surveyed, the lack of random sampling techniques and non-representative sample surveyed, limits the conclusions that can be drawn from this assessment.
"Train the trainers" sessions were designed to ensure that volunteer survey administrators conducted the survey in a prescribed and consistent manner. However, a standardized administration of the survey was not consistently applied. The lack of standardized administration of the survey made it difficult to discern if differences in responses to survey items reflect the opinions of those surveyed, are manifestations of selection bias, or a reflection of interviewers' characteristics.
Due to the fact that the Four-County Community Health Needs Assessment is an opinion survey, it is not an empirically validated instrument. Content areas and items included in the four-County Community Health Needs Assessment Survey were selected based upon a review of the relevant literature and the Hometown Health Guide and by members of the Hometown Health Data Committee. Reliability and validity coefficients for the survey instrument were not established, nor was there an item or content analysis of survey questions.
The surveys were available in English and Spanish. While the research protocol provided for a standardized administration of the assessment tool, this procedure was not consistently followed. Aside from those surveys administered on a one-to-one basis by the interviewers (as was intended), many surveys were filled out as a group, or by the interviewees themselves. These unplanned data collection methods increase the possibility that items in the survey were misunderstood. Likewise, many of the surveys were not fully completed, and several items were answered with response sets. In these instances a response was perfunctorily selected by marking one line through or one bubble around, the same answer from the top of the page to the bottom. Surveys were also not translated into Marshallese, as was requested by members of the Marshallese community. High status members of the Marshallese community were utilized to translate and to assist in completing the survey. While this method facilitated the inclusion of a significant number of Marshal Islanders, the methodological limitations previously disclosed, and the lack of a culturally and language sensitive research instrument limits the utility of the information obtained from Marshallese respondents.
The wording of survey items required respondents to assess the importance of health care concerns in terms of the community members who reside in the same county as them. It is therefore, unclear whether results obtained from this opinion survey accurately represent the health care concerns of the community itself, or are simply reflective of the health care perceptions of the sample employed in this study.
The statistical analysis of the data was limited to means and standard deviations, an assessment of the representativeness of the sample characteristics, and the frequencies of survey items. Wilkes Lambda and Tukey post hoc analysis of survey items with salient demographic variables was also utilized. While data analysis provided the ability to generate county health opinion profiles (e.g. , most frequent responses, discrete comparison between the demographic profile and health concern content areas), it precluded the ability to develop a comprehensive multivariate analysis between demographic variables and survey data. Limitations in the sampling procedures employed in this study precluded the development of a Northwest Arkansas Regional Health Profile.
While this survey was designed to assess the health care perceptions of the Northwest Arkansas region as a whole, there was a limited attempt to canvass areas throughout the four county region. It is important to take into consideration that there were no attempts to analyze survey data based upon the location within the county in which the survey participant resided. Geographical and population representation would need to be considered in order to accurately reflect the health care perceptions of the entire region.