Social and cultural context shapes how communities perceive health, well-being, and risk. Risk perception can change over time as a product of new information and improved understanding. We investigate risk perception and protective behaviors in Southeast Alaska during the first year of the COVID-19 pandemic. Surveys were circulated at two time points: (1) April-June 2020, before COVID-19 reached epidemic levels in the region, and (2) November 2020-February 2021. Ordinary least squares (OLS) regression models were used to analyze how demographic characteristics of the respondent population influenced risk perception. OLS models were again used to predict how individuals engaged in protective behaviors while controlling for risk perceptions. Controlling for demographic characteristics, risk perception increased as age increased for perceived risk of getting sick and dying from COVID-19, males perceived lower risk in general for all tested variables, and Alaska Native respondents perceived higher risk than non-Alaska Native respondents. Controlling for risk perception, results for protective behaviors were mixed; however, the strongest association identified was that knowing someone with a positive COVID-19 diagnosis increased protective behaviors. Between the two time points, risk perceptions increased significantly for variables related to oneself, others, and community members becoming infected with COVID-19. Protective behaviors like traveling less than normal, buying more cleaning products, and engaging in more subsistence gathering significantly increased. Identifying patterns of risk perception and protective behaviors, and especially how they change over time, are critical to developing place-specific public health recommendations, action, and preparedness plans against future infectious threats.

Public health messaging in the United States during the COVID-19 pandemic has been variable in effectiveness. Different levels of governmental institutions have different goals and methods; it is unclear how messaging from these disparate levels is received, trusted, and implemented. We investigate the degrees of trust of Alaska Native and non-Alaska Native people in Southeast Alaska for the U.S. federal, Alaska state, and local Southeast Alaskan governments to parse how Southeast Alaskans feel about relative preparation, actions, and public health messaging. We use data collected in two waves of a regional survey: the first in April-June 2020, and the second in November 2020-February 2021. Results indicate that trust in the federal government was significantly lower than in the Alaska state government for each time period, and trust in both federal and state government significantly decreased between the two periods. Trust in the local governments of Southeast Alaska were significantly higher than both state and federal levels, and increased between the two survey waves, albeit insignificantly. We discuss potential drivers of these observations and outline how this can be leveraged for more comprehensive research into how relatively small communities with a large Native presence perceive public health messaging from different sources.

Stalemates frequently obstruct disaster risk management related initiatives, including related to risk financing. They can arise from misaligned stakeholder objectives and a cognitive effort to avoid decision-making under uncertainty. Participatory action research techniques can be useful for overcoming stalemates but have not been examined in the context of disaster risk management. To fill this gap, we explored how participatory techniques could overcome one disaster risk financing related stalemate: the case of landslide insurance in a highly landslide-prone location, Sitka, Alaska. Landslide insurance in Sitka is a classic stalemate, as diverse stakeholders see the value of offering a landslide insurance policy, but their respective concerns and information gaps stymie their ability to feel confident in decision-making making landslide insurance generally unavailable. Utilizing a series of inclusive interviews and workshops as our participatory techniques, respondents describe how the workshop engendered confidence in the prospect of landslide insurance, specifically by brokering new relationships and serving as a catalyst for new public-private partnerships while effectively reframing the landslide risk financing conundrum. However, they also note that a new market for landslide insurance requires sufficient state or Federal funding support, improved probabilistic landslide models, and sustained consumer demand. This suggests that participatory techniques can play an enabling role in overcoming stalemates, but only if certain conditions are in place. These findings may have relevance to policymakers facing stalemate-related obstacles in disaster risk management.

Several behavioral, social, and public health interventions, such as suicide/HIV prevention or community preparedness against natural disasters, leverage social network information to maximize outreach. Algorithmic influence maximization techniques have been proposed to aid with the choice of "peer leaders" or "influencers" in such interventions. Yet, traditional algorithms for influence maximization have not been designed with these interventions in mind. As a result, they may disproportionately exclude minority communities from the benefits of the intervention. This has motivated research on fair influence maximization. Existing techniques come with two major drawbacks. First, they require committing to a single fairness measure. Second, these measures are typically imposed as strict constraints leading to undesirable properties such as wastage of resources. To address these shortcomings, we provide a principled characterization of the properties that a fair influence maximization algorithm should satisfy. In particular, we propose a framework based on social welfare theory, wherein the cardinal utilities derived by each community are aggregated using the isoelastic social welfare functions. Under this framework, the trade-off between fairness and efficiency can be controlled by a single inequality aversion design parameter. We then show under what circumstances our proposed principles can be satisfied by a welfare function. The resulting optimization problem is monotone and submodular and can be solved efficiently with optimality guarantees. Our framework encompasses as special cases leximin and proportional fairness. Extensive experiments on synthetic and real world datasets including a case study on landslide risk management demonstrate the efficacy of the proposed framework.

We conducted a rapid, mixed-methods assessment to understand how COVID-19 affected Latinx sexual minority men (LSMM) and transgender women (LTGW). Using a computer-assisted telephone interviewing software, one interviewer called 52 participants (randomly sampled from a larger HIV prevention pilot study aiming to increase HIV knowledge and testing frequency; n = 36 LSMM and n = 16 LTGW) between 04/27/20-05/18/20. We quantified core domains using the Epidemic-Pandemic Impacts Inventory scale and provided important context through open-ended qualitative questions assessing: 1) COVID-19 infection history and experiences with quarantine; 2) Health and healthcare access; 3) Employment and economic impact of COVID-19. Participants reported increases in physical conflict or verbal arguments with a partner (13.5%) or other adult(s) (19.2%) due to stressors associated with the safer-at-home order. Participants also reported increased alcohol consumption (23.1%), problems with sleep (67.3%) and mental health (78.4%). Further, disruptions in access to Pre-Exposure Prophylaxis or PrEP-a daily pill to prevent HIV-occurred (33.3% of 18 participants who reported being on PrEP). Many said they received less medical attention than usual (34.6%), and LTGW reported delays in critical gender-affirming hormones/procedures. Half of the participants lost their jobs (50.0%); many undocumented participants relayed additional financial concerns because they did not qualify for financial assistance. Though no COVID-19 infections were noted, COVID-19 dramatically impacted other aspects of health and overall wellbeing of LSMM and LTGW. Public health responses should address the stressors faced by LSMM and LTGW during the COVID-19 pandemic and the impact on wellbeing.

To the Editors: INTRODUCTION Mental and sexual health disparities have been well documented among lesbian, gay, bisexual, and transgender (LGBT) communities1–3 and are often further exaggerated for Latinx LGBT. For example, increasingly restrictive US immigration policies have heightened the unwillingness of many Latinx sexual minority men (SMM) and transgender women (TGW) to engage in formal systems such as health care.4 COVID-19 has the potential to widen existing mental and sexual health disparities in general,5,6 and among Latinx LGBT communities in particular,7 yet there remains limited empirical data, and most existing studies evaluate outcomes only during the pandemic, which confounds the pre-COVID-19 status with the (additional) impact of COVID-19. We provide much-needed, longitudinal information on how COVID-19 is impacting the mental and sexual health of Latinx SMM and TGW across Los Angeles County. METHODS We evaluate the impact of COVID-19 longitudinally based on prepandemic data from participant responses in a NIH-funded study [R34 MH109373] collected between August 29, 2018, and July 03, 2019, and from a randomized subset of 50 of the same participants gathered during the stay-at-home-order between May 1, 2020, and May 18, 2020. Detailed information on the parent study is published elsewhere8,9; thus, we provide only a brief summary below. All participants were originally recruited through our partner Bienestar Human Services Inc., a community-based organization focused on serving Latinx and LGBT communities with 6 offices across Los Angeles County. Inclusion criteria were being HIV negative, self-identifying as Latinx and SMM or TGW, 18 years of age or older, owning or having regular mobile phone access, and fluency in English or Spanish. The institutional review board approval was received for all study protocols and materials. A random subsample of 50 participants of the parent study were called, read a copy of the informed consent document, and asked to provide verbal consent. The survey took an average of 20–30 minutes to complete. The primary outcomes are based on validated measures of mental health (assessed using the Kessler 10 interviewer-administered nonspecific distress scale10) and sexual health (based on select questions from the AIDS Risk Behavior Assessment11). Basic descriptive statistics and longitudinal comparisons (difference-of-means tests and difference-in-proportions tests) were conducted (paired t-tests and z-tests) using Stata version 15.1.12 Because of the small sample size, we abstain from conducting statistical tests for differences between Latinx SMM and TGW. The qualitative data were recorded, transcribed, translated (for those conducted in Spanish), and then coded using Dedoose (Version 8.3.21).13 Two team members (M.I. and S.M.) jointly coded a random set of 21 transcripts and achieved a pooled Cohen14's Kappa of 98% indicating “good agreement.” A directed content analysis15 was used to provide context for key themes. RESULTS The sample included 35 Latinx SMM and 15 Latinx TGW. We found no statistically significant differences on key sociodemographic characteristics between the study sample reported here and the parent study. Both groups show similar age (mean age of 40.2 years for Latinx SMM vs. 38.5 for Latinx TGW) and relationship status (65.7% of Latinx SMM single vs. 73.3% of Latinx TGW) (Table 1). The 2 groups differ in their incomes (22.9% of Latinx SMM had an annual income of over $35,000 vs. 6.7% of Latinx TGW), education (82.9% of Latinx SMM have at least a high school education vs. 46.7% of Latinx TGW), and a higher proportion of full-time employment (65.7% of Latinx SMM vs. 46.7% of Latinx TGW). More Latinx SMM are US citizens or permanent residents (65.7% of Latinx SMM vs. 53.3% of Latinx TGW), but the same proportion had undocumented residency status (20.0% Latinx SMM vs. 20.0% of Latinx TGW). Although no participant reported testing positive for COVID-19, many reported increased risk of complications if infected with COVID-19 (40.0% of Latinx SMM vs. 20.0% of Latinx TGW). TABLE 1. - Sociodemographic Characteristics of and Changes in Mental and Sexual Health Among Latinx SMM and TGW (n = 50) Variable Pre-COVID-19 Survey z-stat COVID-19 Survey SMM TGW Pooled Pooled SMM TGW Age 39.7 40.2 38.5 At least a high school education 82.9 46.7 72 Full-time employed 65.7 46.7 60 Annual income >$35,000 22.9 6.7 18 Relationship status Single 68 65.7 73.3 Married or living with a partner 30 31.4 26.7 Divorced/separated/widowed 2 2.9 0 Documentation status US citizen or permanent resident 62 65.7 53.3 Deferred action for childhood arrivals (DACA), asylum, and U visas for victims of violent crimes 10 5.7 20 Undocumented 20 20 20 Others 8 8.6 6.7 COVID-19 characteristics Have a risk factor increasing potential complications of COVID-19 34 40 20 Currently or previously had COVID-19 symptoms 14 14 13.3 Tested positive for COVID-19 0 0 0 Mental health Fraction likely well 74 93 80 −2.62‡ 56 58 50 Sexual health Condomless sex 29 53 36 −1.90* 19 14 31 Perceived HIV risk from condomless sex 18 −2.73‡ 2 The value displayed for paired z-tests are the differences in means between surveys.‡, †, and * indicate significance at the alpha = 0.01, 0.05, and 0.1 levels, respectively. Mental Health Table 1 shows significant declines in mental health due to COVID-19. Before the pandemic, about 80% of respondents fell into the category of “likely to be well,” decreasing to 56% during the stay-at-home-order (P-value = 0.01), with Latinx TGW experiencing a larger drop from 93% to 50%. Importantly, there is a statistically significant change in 7 of 10 items of the scale, indicating a broad decrease in mental health. Qualitative findings illustrate how much of the distress was tied to the financial hardship associated with job loss. One Latinx SMM participant said, “Well the stress, the anguish, in the news that so many people are dying… also what's going on with the economy, because they say the economy has gotten worse. Everything is getting worse and there is more of this in future plans” (age 54). At the same time, some said that thinking about leaving their house to return to work increased their fear of contracting COVID-19. Some reflected that staying at home gave them the opportunity to process their emotions, whereas others said having so much time, without access to their traditional coping mechanisms—and in particular the support of their friends and family— was overwhelming. Sexual Health In Table 1, the fraction of people reporting condomless sex dropped from 36% to 19% (P-value = 0.06); prepandemic more than one quarter (29%) of Latinx SMM respondents reported condomless sex compared with 53% among Latinx TGW, decreasing to 14% and 31%, respectively (because of the small sample size we do not know whether the differences are statistically significant). Furthermore, among those instances of condomless sex, the perceived risk of contracting HIV also decreased significantly, from 18% saying there was some risk of infection (“almost no risk” was classified as knowing that the partner either is HIV-negative or virally suppressed) prepandemic to 2% during the pandemic (P-value = 0.01). Qualitative results revealed that most participants reduced the number of their sexual partners to avoid the risk of getting infected with COVID-19: “I don't want to get COVID-19; so, I'd rather not expose myself to other individuals…” (LSMM age 51). Some Latinx SMM and TGW reported increased fears associated with COVID-19 to continue engaging in sex work: “[With respect to sex work, I feel] a little fear because I still have had to do it when nothing's opened and I'm exposed to people, but I have to support myself” (Latinx TGW age 37). DISCUSSION To the best of our knowledge, this is the first study to leverage longitudinal data to isolate the impact of COVID-19 on the mental and sexual health of Latinx SMM and TGW by comparing data collected during California's stay-at-home-order with their situation approximately 8–18 months before the pandemic's start. With respect to mental health, we find a significant decline across mental health outcomes that seem to be more pronounced among Latinx TGW. We also find a decrease in sexual behaviors such as condomless sex, although a small minority of respondents did report engaging in sex work to counter their financial strain while also engaging in risk reduction techniques. For example, 1 participant noted that to safely continue engaging in sex work, they reduced their number of sexual partners and avoided home visits. Our findings contribute to a currently limited literature on the impact of COVID-19 among LGBT populations and Latinx LGBT specifically. Existing studies have focused on people living with HIV,16,17 with the exception of 1 study based on a cross-sectional online survey among SMM (n = 1051)—not Latinx specifically—that reported adverse impacts on general wellbeing and that half reduced the number of sexual partners.18 Another small study among Latinx SMM (n = 12)19 reported that mental health concerns were exacerbated by financial loss; a decrease in sexual behaviors was also noted. Our study contributes to this literature and enlarges the evidence base by confirming similar results among Latinx SMM and TGW in Los Angeles County, based on an assessment of critical mental and sexual health outcomes prepandemic compared with data collected during the stay-at-home-order. Our study has significant strengths and provides critical insights based on a longitudinal study design that allows us to more clearly isolate the impact of COVID-19. There are also limitations: first, it is based on a relatively small sample of 50 participants, preventing us from statistically evaluating differences between Latinx SMM versus TGW. We cannot compare the changes observed to a control group to investigate to what extent they may be driven by factors other than COVID-19. The outcomes are self-reported and as such may be subject to social desirability bias. Furthermore, the prepandemic data collection happened over a relatively extended time period, so we therefore cannot exclude that other events may have also negatively influenced their mental health. However, the qualitative data convincingly support the paramount role of COVID-19. Finally, our focus on Latinx populations limits our ability to concretely state if and how the deterioration in sexual and mental health outcomes within our sample is reflective of widening gaps in disparities with other racial/ethnic populations. Our results suggest public health responses should address unique COVID-19 stressors on mental health such as anxiety because of job loss or isolation from social support networks. The pandemic also presents opportunities, as some health departments have already shown20 to promote the enjoyment of safer sex and mental health support through virtual spaces while simultaneously reducing the risk of spreading COVID-19.