Food security is defined as “enough food for an active, healthy life” for all members of a household [1]. In contrast, a household is considered to be food insecure when the provisions outlined for food security are limited by money or lack of resources [1]. Food insecurity is associated with poorer outcomes in the prevention and management of chronic cardiometabolic diseases such as diabetes, hypertension, obesity, and cardiovascular disease [2-5]. These disease outcomes are a product of the compensatory behaviors that members of food-insecure households employ to avoid hunger, which can include relying on nutritionally poor, calorie-dense foods [6].

The prevalence of food insecurity in the United States has increased during the COVID-19 pandemic, impacting an estimated 54 million households [7], as many household members experienced periods of under- or unemployment. In attempts to ensure that US households had access to food, many nutrition assistance programs evolved during the COVID-19 pandemic. These evolutions included the rapid expansion of the US Department of Agriculture’s pilot program for Supplemental Nutrition Assistance Program (SNAP) benefits to be used online [8]. As internet and social media play a growing role in how individuals access food and nutrition information [9], it is important to consider how the food environment—defined as the interface where people interact with the wider food system—becomes increasingly digital [10]. A scoping review of studies of the digital food environment demonstrated how social networking platforms can be used to shape food culture, drive trends, and also serve as an abundant source of health and nutrition information [9].

Social networking platforms, and the online communities they foster, have been capitalized on in various ways in efforts to address chronic cardiometabolic diseases and food insecurity. Facebook is one of the largest online social networking platforms, which was used by over 2 billion people each month in 2019 and 2020 [11]. Facebook has been used to facilitate emotional and informational support via peer-to-peer and caregiver-to-patient interactions, leading to improved disease self-management among patients with diabetes [12]. However, these approaches to focus on cardiometabolic diseases using social media do not appear to have yet been adapted to address food insecurity. In a scoping review of 39 studies of digital technology use in food assistance programs, only a single study described the use of social media by a food pantry to communicate with clientele [13]. Another single observational study found that consuming information about COVID-19 online was associated with individuals’ concerns about food insecurity [14]; however, it remains uncertain how social media may be used to discuss, and ultimately address, food insecurity.

Native Hawaiian and Pacific Islander (NHPI) communities are a priority population for consideration as the digital food environment continues to evolve in the United States. NHPI adults experience a disproportionately higher prevalence of cardiovascular disease, diabetes, hypertension, and obesity compared to other demographic groups [15,16]. Additionally, a recent study showed that NHPI adults experience a higher prevalence of food insecurity (21%) than their non-Hispanic white counterparts (8%) [17]. A population-based survey of primarily NHPI adults (N=637) in low-income households found that very low food security was strongly associated with greater odds of both hypertension and diabetes [18]. Interventions to address these dual burdens of cardiometabolic disease and food insecurity among NHPI adults are needed. NHPI populations have distinct cultural characteristics, including values, history of foreign colonization and exploitation, and emigration to the continental United States, warranting culturally informed interventions [19-21].

Healthy Hearts Among Pacific Islanders (HHAPI) is an educational program created by and for NHPI people. The goal of HHAPI is to utilize culturally grounded evidence-based interventions to mitigate cardiometabolic conditions in the NHPI population via self-management of hypertension and diabetes at the individual, family, and policy levels. At the individual and family levels, HHAPI began in 2016 with in-person hypertension management classes hosted in community organizations in the Pacific Northwest, with one of the largest NHPI populations outside of Hawaii [22], which then shifted online in 2020 due to the COVID-19 pandemic. At the policy level, HHAPI initially intended to influence grocery store labeling of foods as either being low in sodium or high in potassium. Given the unanticipated changes with the COVID-19 pandemic, including the growing importance of social media, the HHAPI project shifted to observing the online food environment and discussions of food insecurity among NHPI people living in the Pacific Northwest, as an opportunity to better understand social media as a potential mechanism to decrease food insecurity and cardiometabolic conditions among NHPI adults.

The purpose of this study was to explore the online presence (static website, social media accounts) of NHPI-serving community groups and cultural or health-focused organizations in Washington state, and to compare the nature (eg, educational, event promotion, resource sharing) and frequency of food insecurity–related messages posted on their social media sites from March through June 2019 and from March through June 2020. We also evaluated the amount and type of engagement (ie, likes, shares, comments) on food insecurity–related messages by the nature of the post.

The sample of online data was extracted from publicly accessible Facebook pages and groups. On Facebook, individuals make posts, comprised of text, image(s), video(s), and/or hyperlink(s) to other webpages, and other individuals can engage with the post. Engagement includes commenting with their own text, image(s), video(s), and/or hyperlink(s), or by clicking the “reaction” button to show one of seven possible reactions (Like, Love, Care, Haha, Wow, Sad, and Angry) as a small, animated symbol listed under the original post [23,24]. Facebook posts can be made on an individual’s profile; a page representing a business, organization, or event; or within a group. Pages for businesses, organizations, or events allow for staff or volunteers to make posts to communicate with their target audience. In contrast, Facebook groups are online spaces that allow for any group members to make posts to communicate around shared interests or identities, allowing for a multidirectional conversation. Facebook groups can be public and visible to any Facebook user, or they can be private, restricted to Facebook users who have applied to join the group and have been formally accepted by group administrators.

Facebook business pages and groups focused on serving NHPI community members in the Pacific Northwest region were identified through consultation with three staff members who identified as Native Hawaiian or Pacific Islander and lived in the Pacific Northwest, a region with large communities of NHPI people. Each group’s Facebook page or group was reviewed to ascertain corresponding details about each organization’s mission or objective. To help maintain confidentiality, the names of the groups and organizations are not included in this manuscript.

The study was certified as “Exempt” from approval by the Washington State University Institutional Review Board (IRB) on March 22, 2021 (IRB #18784).

The list of keywords used to identify posts related to food security was generated in three steps. First, the keywords were identified through a review of related literature [25-29], which included the terms food desert, food insecurity, food security, food supply, food access, food sufficiency, and food insufficiency. Next, the list was supplemented with plain-language synonyms or related terms to include free food, free lunch, free breakfast, free dinner, free plate, free meal, food pick-up, boxed meals, boxed food, food aid, food kitchen, food distribution, meal center, food donation, hunger, and hungry. Lastly, the list of terms was reviewed by a member of the research team with expertise in food security and experience in conducting systematic literature reviews. Based on their suggestion, the final list of keywords included additional terms referring to food assistance programs (food stamp, food bank, food shelf, food pantry, food drive, soup kitchen, food closet, WIC [Special Supplemental Nutrition Program for Women, Infants, and Children], SNAP, EBT [Electronic Benefits Transfer], SNAP-Ed [SNAP Education], senior meals, and school lunch).

One research team member searched all identified Facebook pages and groups using each keyword as the search term. The results were restricted to posts made in the months of March through June in the years 2019 and 2020 to compare social media engagement before the COVID-19 pandemic to the same period during the pandemic. The researcher conducting the searches was not a member of any of the Facebook groups or a follower of the Facebook pages searched. All searches were logged and the date of any resulting posts was documented. Each post was then assessed for its relevance to food security. Posts were considered related to food security if they contained information about food assistance opportunities/events or about individual or community needs for food due to income or access constraints. If the post was considered relevant to food security, additional data were documented about the content of the post (ie, text, web links, images, videos) as well as the amount and type of engagement (ie, comments, shares, and reactions). All posts that were not considered relevant were excluded from coding and analysis. These searches were conducted between June and August 2021.

To identify and quantify the nature of posts identified, a content analytical approach was used to assess patterns of posts and related engagement by their categorical nature [30]. This process began after all searches were completed between September and October 2021. Three members of the research team independently reviewed all of the relevant posts identified, and created memos that reflected possible codes, corresponding definitions, and example posts from the data set. Two members of the research team reviewed the relevant literature to identify possible literature-based codes that have been used in prior research on food security, public health, or online information exchange. These memos and potential literature-based codes were discussed by the three researchers as a group and used to create an initial codebook that was used by all three researchers to independently code the data. The agreement of these independently coded data was evaluated according to the percentage agreement between coders. Differences in interpretation or application of the codebook during the coding process were evaluated and discussed as a group to refine the codebook and maximize clarity and specificity. The final codebook was used by the group to review and finalize codes for all posts and accompanying comments that characterized the nature of information shared and audience engagement.

To describe the online presence of the NHPI-serving groups and organizations identified, the purpose or mission of the organization’s page or group was broadly categorized. Means and frequencies were used to describe the number of users engaged by each page or group. To assess the nature and engagement of posts, the number of posts identified per page by year and types of engagement (eg, reactions, shares) were enumerated. Codes resulting from the content analysis were described and compared by year, page type, and engagement. Qualitative coding was organized in a spreadsheet (Excel, Microsoft Office 365) and quantitative descriptive analyses were conducted in STATA/MP 17.0 (StataCorp, LP).

Seven relevant Facebook pages (n=5, 71%) or groups (n=2, 29%) were identified. Broadly, the stated missions or objectives of these organizations and groups created a welcoming online space to build community among Pacific Islanders, share resources, promote events or local businesses, preserve culture, and/or address social or health inequities. All groups or pages had, on average, 4229 (SD 2731) followers or members. Facebook groups had, on average, a greater number of members (mean 6010, SD 2279) than pages had followers (mean 3516, SD 2769).

Across all groups or pages, 1594 posts were identified using all search terms during the selected time periods. Of these, 278 posts were duplicates (ie, different search terms resulted in the same post being identified) and were removed. Of the 1316 nonduplicated posts, 88 (6.69%) were considered relevant to food security. Nonrelevant posts included announcements for surveys and information about hours of local businesses, among other topics. Among the 88 relevant posts, a subset of posts (n=8) were reposted 2-4 times on different dates and/or in different groups with identical text content. These posts were treated as unique and remained in the data set to reflect their greater potential reach. In 2019, there were 8 relevant posts identified in contrast to 80 relevant posts identified in 2020. Across the 7 groups or pages, a range of 0-45 relevant posts were identified with a mean of 12.6 posts (16.2 SD) per group or page. Of the 88 posts, 12 (14%) had comments associated with the posts, which were qualitatively analyzed and are described below.

A total of 20 codes for posts and 7 codes for comments were developed and used to characterize the data set. Codes were not mutually exclusive, and each post had a mean number of 3.8 (SD 1.7) codes applied. For the posts in 2019, a total of 15 codes were applied (mean of 1.9 per post), whereas the posts in 2020 had a total of 321 codes applied to them (mean of 4.0 per post). Among posts with comments, a mean of 1.8 (SD 0.8) codes were applied to the comments. After coding was complete, the concepts for posts were broadly organized into 6 categories, with some codes represented in more than one category, to assist with interpretation and presentation of results. Given the small number of comments included in the data set, all 7 codes for comments were interpreted and presented together.

There were similarities and differences in the themes of the comments (Table 4) when compared to the larger data set of posts. One key similarity was in the salient gratitude and support reflected by users’ comments, mirroring the optimism and community support indicated in the posts. On the one post that included critical jests about food assistance, comments echoed this sentiment. Similar jests were not present in comments on any other included posts. While there were a few posts relevant to literature-based codes (educational and promotion of event or business), there were no comments with relevant content. Another point of divergence was in self-disclosures of food insecurity experiences. A single disclosure of food insecurity within a comment was explicit about the individual’s family having gone hungry in the past.

Among the analyzed posts, 74% (65/88) had some form of engagement via comments, reactions (Like, Love, Care, Haha, Wow, Sad, and Angry), or shares. Posts coded as reflecting community gratitude or support had descriptively more engagement (mean 19.9, 95% CI 11.2-28.5) in comparison to those that were not (mean 6.1, 95% CI 1.7-10.4). Engagement did not appreciably differ when all other broad code categories were compared. However, posts coded as incorporating culture into their text also had higher amounts of engagement (mean 26.8, 95% CI 12.7-40.9) than those that did not (mean 5.3, 95% CI 3.0-7.7).