We used data collected through this MMCBPR pilot study to illustrate how several data elements can be integrated to describe participants’ experiences around social determinants of health using the domain of food insecurity as an example. By integrating qualitative data from written neighborhood reports, surveys, and geospatial data, we present a data model with the potential to identify specific, targeted needs and resources that emerge from participants’ experiences. In our example, data integration indicates how participants who have ever experienced food insecurity (before or after the beginning of the pandemic) were more likely to also report employment, housing, and healthcare changes during the pandemic and more likely to report everyday discrimination. Similarly, those reporting food insecurity were much more likely to mention resources and programs of any kind and to report experiences related to equity in their neighborhoods.
While this report is focused on food insecurity, it is intended to serve as an illustrative example of how this data model can inform tangible solutions across domains of structural and social determinants of health. For example, we found that those reporting both food insecurity and a higher number of other pandemic-related stressors had higher mentions of resources, programs, potential solutions, and issues related to equity and coping. This suggests that more comprehensive wrap-around services are needed for individuals who already experience stressors related to food insecurity. In addition, participants who mentioned food in neighborhood reports and who had many COVID-related stressors suggested areas for resource development related to outdoor space and health. Thus, a multilevel response to this need might include 1-community food service organizations serving as a point of contact for individuals who may benefit from other resources (healthcare and public health information) and 2-investment in outdoor spaces as sites to promote physical and mental health as well as deliver tangible health resources. Similarly, this observation stresses the importance of cross-institutional communication and connectivity to alleviate the administrative burden experienced by those who access multiple social services.
In our study, results from maps largely serve to validate the congruence of information across our data sources (e.g., mentions of food occurring most frequently in neighborhoods with higher area-level food insecurity). While we are reassured in the capacity for transformation of neighborhood reports to align with other measures of food access or experience, integration of reports or survey data with geospatial data would be most powerful when it contributes to new or emerging understanding of communities’ needs. Doing so may require smaller geographic units of analysis with larger numbers of community residents contributing reports and survey responses. For example, emergence of themes related to (non-COVID) safety in the vicinity of an existing food resource may be an early signal of an emerging access barrier or relatively frequent positive mentions of racial equity may identify exemplary aspects of existing programs. And of course, new mentions of needed resources may indicate specific areas with emerging need. We will continue to develop the relevancy of our data model to hyper-local assessment of barriers and solutions in ongoing and future studies.
A limitation inherent to our data model is that integration across data sources inevitably reduces cell sizes. Specifically, because participants were provided with multiple options to share their experiences and perspectives, there was limited overlap between data sources. This could lead to an overreliance on making inferences based on single data sources rather than on integration. Despite that, using a data integration model can help overcome the limitations of small sample sizes by leveraging information across sources of data to reinforce the validity of each. For instance, the fact that there were more mentions of food resources in qualitative neighborhood reports among those reporting food insecurity in the epidemiologic survey bolsters the validity of both data sources. Due to the small size of the participant population in this pilot study, we were unable to examine results according to individual-level factors such as race, ethnicity, immigration status, LGBTQA + status, and other individual social factors that impact individuals’ vulnerability to structural oppression. Targeted data collection to specific populations where appropriate or increasing participation would allow us to better represent those with intersecting identities.
Despite these limitations, there are several key benefits of this approach to data collection compared with other approaches that have been employed. First, collecting qualitative data using a mobile platform combined with brief, focused survey questions simultaneously with geographic information allows for continued participation compared with longer forms of data collection (e.g., in-depth interviews). Additionally, because the Streetwyze platform allows participants to share their feedback and experiences in real-time and within the context of their own neighborhood experiences, the barrier to participation is lower. This allows for sustained and low-effort participation and can make research more accessible for more diverse people and experiences.
Overall, we have demonstrated the feasibility of simultaneous, multi-modal data collection, which can be scaled to allow for a wider scope of inquiry, along with a tangible model for data integration with the potential to identify salient needs emerging from communities. The combined analysis of quantitative, qualitative, and geospatial data can be adapted by community organizations, policy makers, and researchers to improve the implementation of policies and services over time by periodically tracking and analyzing residential reports and sustaining community participation. This effort is in line with our broader intent to facilitate community-led research and community ownership of data that generates local knowledge, investment, and progress (Akom, Hope Hassberg, and Cruz; pending book chapter).