«UNDERSTANDING THE PROCESS AND KEY FACTORS OF COMMUNITY IMPLEMENTATION OF A FOOD STORE-BASED NUTRITION INTERVENTION ON THE NAVAJO NATION by Muge Qi, ...»
This study incorporated multiple stages of fieldwork between January 2007 and May 2009. The first was during the development and initiation of the NHS program, when I visited the Navajo Nation multiple times (usually about one week per visit) to assist in Food Frequency Questionnaire (FFQ) development, trainings of baseline data collectors, and NHS community workshops. During these visits, I attended meetings with NSDP representatives, and talked to some of NSDP managers and staff to understand their work and community they serve. I took notes during the meetings with NSDP representatives and community workshops, and wrote up meetings minutes and workshops reports. The second stage of fieldwork occurred during implementation of the early phases of the NHS intervention, when I attended two interventionist trainings, a capacity building workshop, meetings with NSDP leaderships, as well as Navajo Nation Human Research Review Board (NNHRRB) meetings. During this fieldwork, I closely observed interactions between the NHS academic team and NSDP managers, supervisors and interventionists and reactions of the NSDP managers, supervisors, and interventionists to the trainings, capacity building workshop, as well as the NHS implementation. I also paid close attention to emerging issues that could affect the NHS partnership and intervention implementation.
Through these experiences, I gained insight into how the NHS partnership process unfolded and what were key issues that could affect the partnership effort to implement and sustain the NHS program. Moreover, these experiences provided me opportunities to know as well as to be known by the NSDP people who were closely involved in the program. As a result, I decided to conduct semi-structured interviews with the NSDP managers and interventionists as well as with the key members of the NHS academic team during the last stage of fieldwork in the end of the NHS implementation.
3.3.2 Semi-Structured Interviews Semi-structured interviews were conducted to understand the experiences and perceptions of individuals as part of the NHS program. Based on the initial fieldwork experience, I identified the principal investigator (PI), the field coordinator, NSDP interventionists, field supervisors and program managers as potential study participants.
Then I contacted NSDP interventionists to help identify potential informants from the participating food stores. Approximately 55 people were possible to be interviewed.
Although at least two people in each store were involved in the NHS implementation, NSDP interventionists identified the store owner or store manager of each store as key store people involved during the implementation. A total of 39 individuals were identified to be interviewed for this study.
To conduct the interviews, first I sent out a letter to the NSDP top manager (director), in which I explained the purpose of interviews, who are going to be interviewed, what types of questions are to ask, and how interviews would be conducted.
The letter also explained IRB approvals for conducting the interviews, obtaining informed consent from each individual, honoring any individual’s refusal to be interviewed, and confidentiality of information provided through the interviews. The director was asked to distribute the letter to program managers, field supervisors, and interventionists. As for individuals who have left NSDP to work elsewhere, their contact information was obtained from their former co-workers and they were contacted by phone. The store owners/managers were contacted by phone after I arrived at the research site, explained to them about the purpose of this study and asked if they were interested in participating in an interview.
All but one individual identified for this study agreed to be interviewed. All signed the consent form before the interview and kept a copy of the signed consent form (Appendix A). The manager of a chain convenience store owned by a private company was not given permission to participate in the interview by their top manager. Thus, a total of 38 individuals were interviewed (Table 3.2).
A semi-structured interview guide was used during interviews. The interview guide was modified for different key stakeholders according to their main role in the program. This allowed the researcher to make best use of the limited time available to participants for the interview, exploring in detail important issues that are particularly relevant to their experiences. For example, the interview guide for interventionists specifically included questions in terms of how the intervention activities were carried
out in the field (Appendix B). For store owners/managers, interview topics included:
store recruitment, intervention implementation, barriers to implementation, coordination with interventionists, and program impact on stores and store customers.
Interviews were conducted by the researcher in a private office at their workplaces during weekdays. Signed consent was obtained from all participants. There was no audio-recording of interviews with local stakeholders because of participants’ preferences to remain completely anonymous, and because of time constraints to obtain the Navajo Nation Human Research Review Board (NNHRRB) approval for audiorecording at the time this study was conducted. I wrote down interview responses with permission, and made every effort to capture actual words or sentences used by participants. On the same day of the completion of an interview, interview responses were entered into Microsoft Word by the researcher, along with a description of the interview setting and informal conversation with the participant before and after the interview, as well as my reflection about the interview. The duration of interviews ranged from 30 minutes to 2 hours, but most of the interviews lasted about 1 hour. Two followup interviews were conducted with two NSDP managers who were more knowledgeable and reflective of the organizational decision making process to obtain answers to additional questions that were emerged from previous interviews.
Program documents can provide “a behind-the-scenes look at program processes” (Patton, 2001, p294) and help “ground an investigation in the context of the problem being investigated” (Merriam, 1992, p126). In this study, available documents were reviewed to reveal the complexity of partnership development and program implementation process, to understand the experiences and perceptions of study participants in the program context, to corroborate information from interviews, as well as to provide background detail for the study.
The primary documents reviewed for this study were relevant to the development and implementation of the NHS program. These documents included program meeting and conference call minutes, formative research reports, community workshop reports, the interventionist manual of procedures, presentation slides for training and capacity building workshops, progress updates and reports, and journal articles. A second set documents, pertaining to the development, implementation and evaluation of the AHS program, was reviewed to explore the adaptation of the previous intervention to the Navajo setting. These documents included formative research reports, community workshop reports, interventionist manual of procedures, journal articles and other publications. Documents were collected throughout the study period (Appendix C). The PI provided all official or unofficial documents generated for the NHS and AHS program.
Additionally, relevant documents were also obtained from the NSDP.
A data documentation form was created to record all the data collected for this study by category, and all documents were properly sorted and labeled (with an identifying notation) for easy access. Electronic versions of the data were stored in a password protected computer and hard copies were stored in a locked cabinet and access was limited to the researchers. NVivo (version 8), a computer-assisted qualitative data analysis software (CAQDAS), was used to assist data analysis.
The template approach (Crabtree & Miller, 1999; King, 1998) was used to analyze textual data (from interviews and documents). In template analysis, a list of predetermined codes or conceptual framework (the initial template) is applied in order to analyze textual data and modified through ongoing analysis until the researcher has achieved as full an understanding of the data as feasible (King, Carroll, Newton, & Dornan, 2002). Such a well-structured analytical process produces a coding template representing themes identified in the data and helps produce a clear, organized final account of a study (King, 2004). Template analysis was appropriate for this study for the following three reasons. (1) It allowed me to use a priori codes to help guide analysis.
Given my fieldwork with the NHS program and review of literature on dissemination and implementation, applying a priori a number of codes helped me to focus on areas of greatest relevance to the research questions. (2) Template analysis works particularly well in studies which seek to examine the perspectives of different groups within a specific context. This study sought to understand the implementation of the NHS program from the perspectives of researchers, practitioners, and food store owners. (3) A key feature of template analysis is hierarchical coding, using broad themes encompassing successive narrower, more specific themes which enable fine distinctions to be made. In this study, hierarchical coding allowed me to analyze the data at varying levels of specificity and to present important aspects of partnership and implementation process in detail.
The analytical process for this study was provided in Figure 3.2. With guidance from the fieldwork, literature review, and interview guides, I constructed a coding template through careful reading and rereading of textual data from the documents and interviews. These preliminary codes were revised multiple times working back and forth between the data and the coding template. In this process, new codes were added and some initial codes were redefined, merged with other codes, placed under different categories, or deleted by examining the meaningfulness of the themes in the light of the research questions and the accuracy of the placement of data in categories. For example, a level-two code ‘interaction with customers’ under a level-one code ‘intervention implementation’ was divided into two separate codes: ‘use of intervention materials’ (a level-two code) and ‘customer response/interest’ (a level-three code under a level-two code ‘interactive educational sessions’.
Then, I and a research assistant who had a qualitative data analysis background independently coded four sets of interview responses. The researcher and a research assistant compared coding, discussed discrepancies, and made several changes to the coding template. Again, the researcher and the research assistant coded another four interview notes independently, which resulted in further refinements to the coding template. Minor adjustments were made to the template on the basis of a detailed rereading of the textual data from the documents and interviews to produce the final version of the analytical template (Appendix D). Coded segments of text were entered into the appropriate data charts created for each code in the final version of the coding template using NVIVO 8.
I prioritized themes that were of direct relevance to the main research questions of this study and of great importance to participants by reviewing the data charts.
Representative, contextually rich quotes were identified to aid the understanding of specific points of interpretation. Because interview responses were hand-written, quotes were grammatically corrected where necessary. The findings were integrated with existing empirical and conceptual literature, and the PI as well as the field coordinator reviewed all drafts of papers. Based on their comments, revisions were made to produce final papers.
All study protocol described here was approved by both the Johns Hopkins University Institutional Review Board (IRB) and the Navajo Nation Human Research Review Board (NNHRRB).
Funding for this study came from the Center for Livable Future, Johns Hopkins University and the US Department of Agriculture Grant Number 2010-8515-20666.
Type and number of NHS participating stores by intervention store area
Study setting: Navajo Nation map
http://www.statemuseum.arizona.edu/exhibits/navajoweave/contemp/map.html Figure 3.2. Flow diagram of qualitative data analysis
Academic - community partnerships offer a model for effective implementation of evidence-based interventions in community settings by incorporating local partners’ knowledge of the host setting, with researchers’ knowledge of effective support strategies and of the intervention to be implemented. However, little is known about how the participatory partnership for implementation develops and how it affects the process of evidence-based intervention implementation. The Navajo Healthy Stores (NHS) program was a locally implemented food store-based intervention developed through extensive formative research and a community engagement process, based on a previous intervention trial. The purpose of this study was to identify the process and strategies used by the academic - community partnership to implement the NHS program. A qualitative study was conducted using a combination of face-to-face interviews with 24 key stakeholders and a review of program documents. Results indicate that the academiccommunity partnership evolved naturally through an engagement, formalization, mobilization, and maintenance process. The academic-community partnership had faced some important challenges needed to address in order to successfully move through the stages of implementation. Understanding the process and key challenges of implementation can guide academic – community partnerships in translating evidencebased interventions into sustainable, community implemented programs.