Introduction

The novel coronavirus disease 2019 (COVID-19) has created an unprecedented state of emergency worldwide due to its severity, rapid transmission, and proliferation across geographical boundaries. The first cluster of cases was reported from Wuhan, China, in December 2019 and quickly spread to numerous countries (Jang et al., 2021; Hearne & Nino, 2021; Nguyen et al., 2020). On February 11, 2020, the World Health Organization (WHO) identified COVID-19, also known as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), as a new and highly pathogenic virus with deadly consequences. On March 11, 2020, WHO declared COVID-19 a global pandemic due to increasing infection rates (Herne & Nino, 2021). As of June 2022, global COVID-19 cases are reported to be almost 530 million with deaths over 6 million (Johns Hopkins University & Medicine). In response to the rising rates of COVID-19 cases, WHO proposed a global action plan that highlights the importance of adopting a range of health-protective behaviors, including handwashing, social distancing, and wearing face masks (Nguyen et al., 2020; Camacho-Rivera, 2020). Although attempts to develop effective treatment and vaccines have progressed, effective management of the disease lies in preventive behavioral change to limit transmission of the virus (Breakwell et al., 2021; Li et al., 2020).

While the preventive strategies employed universally are comparable, implementation suitability, scale, and rigor have varied significantly across geographic locations and demographic subgroups shedding light on the role of demographic factors in the uptake of COVID prevention behaviors (Camacho-Rivera et al., 2020; Firouzbakht et al., 2021, Malik et al., 2020; Nguyen et al., 2020; Kim & Kim, 2020; Li et al., 2020). Further, public health and clinical researchers have indicated the disparate burden of COVID-19 morbidity and mortality among racial and ethnic minorities, older adults, and individuals with preexisting chronic health conditions (Kim & Kim, 2020; Kim & Crimmins, 2020; Malik et al., 2020). For example, clinical and population-based studies have demonstrated that non-Hispanic Black and Latino individuals and communities are at increased risk for COVID-19 exposure, morbidity, and mortality (Camacho-Rivera et al., 2020).

In further examining the uptake of vital preventive behaviors, face masks usage increased from 78% in April to 89% in June 2020 (Hutchins et al., 2020). However, other COVID-19 mitigation behaviors (e.g., handwashing, social distancing, and avoiding public or crowded places) were lowest among younger adults (aged 18–29 years old) and highest among older adults (aged over 60 years old) (Hutchins et al., 2020). Lower engagement in mitigation behaviors among younger adults might be one reason for the increased incidence of confirmed COVID-19 cases in this group compared to those over 60 years old (Boehmer et al., 2020; Kim & Crimmins, 2020). Current literature underscores education, income, and housing as critical indicators for uptake and maintenance of essential COVID-19 protective behaviors. For example, adults with a high school degree or below, household income less than $50,000, uninsured, employed, residing in rural areas, and without any chronic diseases were more likely not to adhere to COVID − 19 preventive behaviors (Islam et al., 2021). Compared with urban residents, rural residents were less likely to execute preventive behaviors, more likely to hold a negative attitude toward the effectiveness of performing preventive behaviors, and more likely to have lower levels of engagement in protective behaviors (Chen & Chen, 2020). Furthermore, rural populations face health disparities due to multiple barriers such as lack of health care resources (e.g., transportation, health insurance, providers, and facilities), geographic distance, and lower socioeconomic status (Chen & Chen, 2020). Overall, these outcomes emphasize the need to develop targeted messaging and behavior modification interventions for various demographic groups.

Numerous chronic health conditions, including hypertension, diabetes, cancer, chronic obstructive pulmonary disease, asthma, and obesity, have also been linked to a heightened risk of poor COVID-19 outcomes (Islam et al., 2021; Chen & Chen, 2020). Adults with cardiometabolic diseases and underlying respiratory conditions were more likely to report working from home or staying home because they felt unwell than those without it (Camacho-Rivera et al., 2020; Islam et al., 2021). Further, adults with immune conditions were twice more likely to report wearing a face mask when compared with individuals without immune conditions (Camacho-Rivera et al., 2020). Likewise, adults with chronic disease or living in rural areas are more likely to adhere to COVID-19 preventive behaviors (Islam et al., 2021).

Theoretical Framework

Communication through personal relationships, social networks, and communities possesses immense power to influence and significantly impact behavior change (Nitschke et al., 2020). Social networks impact an individual’s health and well-being through various factors such as social support, social influence (norms), social engagement, access to resources, and person-to-person contacts (Smith & Christakas, 2008). For example, in assessing the role of social connections with COVID-19 affected areas, one study found that social connections serve as essential cost-effective channels of information and communication about the dangers of COVID-19 thus impacting compliance with mobility restrictions (Charoenwong et al., 2020). Likewise, another study examined social connectedness in relation to stress, worry, and fatigue during the lockdown period and found that greater social connectedness was linked with lower levels of COVID-19 specific stress and worries (Nitschke et al., 2020).

Studies examining the impact of social networks have focused primarily on the influence of social networks, specifically, how norms or beliefs within an individual’s trusted network shape individual health behavior (Charoenwong et al., 2020; Smith & Christakas, 2008). An evolving body of research suggests that social norms lead to positive health outcomes because individuals are highly responsive to the behavioral choices made by individuals they trust and are more likely to adopt, model, or engage in behavior change (Bavel et al., 2020; Charoenwong et al., 2020). Earlier empirical studies in relation to COVID-19 have highlighted the importance of factors such as trust and social norms in facilitating preventive behavior depending on levels of individual risk perception (Caetano et al., 2013; Charoenwong et al., 2020). To illustrate, a recent study found that college students reported remarkably greater intentions to get the COVID-19 vaccine based on expected norms amongst their peers (Abdallah et al., 2021). Therefore, targeting well-connected individuals typically considered trustworthy and making their behavior change salient to others is an effective health promotion strategy (Bavel et al., 2020; Caetano et al., 2013; Charoenwong et al., 2020).

Communication or engagement within one’s social networks is particularly effective as it comes from familiar sources people trust (Goldberg et al., 2020; Li et al., 2020; Nitschke et al., 2020). Social connectedness through communication play a prominent role in how individuals engage with preventive behaviors. For example, Charoenwong et al. (2020) in the early stages of the pandemic found that individuals with more connections to China and Italy, two early hotspot areas, complied significantly more with imposed local government mobility restrictions. Similarly, Li et al. (2020) found that compared to individuals who did not have positive cases in their social circles, those who did were more likely to get tested for COVID-19. Women are more socially connected to diverse and larger networks, which may help explain why they are more likely to adopt preventive behaviors (Caetano et al., 2013; Smith & Christakas, 2008). Therefore, public health messaging targeting particular demographic groups and geographic regions must address diseases of epidemic potential. With a rapidly growing pandemic such as COVID-19, timely assessments of the public’s behavioral responses to the pandemic are crucial to informing public health policies, interventions, and recommendations. Given the unique epidemiological characteristics of COVID – 19, it is imperative for the public to actively engage in preventive behaviors to curb the spread of the virus. To this end, this study draws on social network theory to examine factors influencing the adoption of COVID-19 preventive behaviors.

Methods

This study used the COVID-19 Household Impact Survey (CIS) by the nonpartisan and objective research organization (NORC) at the University of Chicago for the Data Foundation (Wozniak et al., 2020). The CIS is a cross-sectional, nationally representative household survey that collects estimates for preventative behaviors, physical and mental health, economic security, and other social dynamic factors during the COVID-19 pandemic. The survey provides weekly estimates of US adults aged 18 and older nationwide and for 18 regional areas, including 10 states and 8 metropolitan statistical areas. This study uses CIS data from three time points, weeks 1–3, collected on April 20–26 (n = 8790), May 4–10 (n = 8974) and May 30 – June 8 (n = 7505), respectively (Wozniak et al., 2020). Detailed CIS study methods are reported elsewhere Wozniak et al., 2020). 25,269 participants participated in the study. Study participants with missing data on any of the study variables were excluded from the analysis (n = 502). Thus the final analysis included 19,815 study participants.

Measures

To assess COVID-19 preventive behaviors, we used participants’ responses to the following questions: Which of the following measures, if any, are you taking in response to the coronavirus? Participants were able to select all that applied from a list of 19 options. Our study focused on the three commonly recommended preventive behaviors: wearing a face mask; washing or sanitizing hands, and keeping six feet distanced from those outside my household. Only those who selected ‘yes’ to practicing all three preventive behaviors were considered compliant.

To assess social engagement, we used two variables: trust and social connectedness. To assess trust, we used participants’ self-reported responses to the following question: generally speaking, would you say that you can trust all the people, most of the people, some of the people, or none of the people in your neighborhood?. Participant’s response options included five options: all, most, some, none, don’t know. We defined participants based on their selections, excluding those who selected “don’t know”.

Social connectedness was based on participant’s self-reported response to communication frequency using the following four questions: in the past month, how often did you talk with any of your neighbors?”, “during a typical month prior to March 1, 2020, when COVID-19 began spreading in the United States, how often did you talk with any of your neighbors?, In the past month, how often did you communicate with friends and family by phone, text, email, app, or using the Internet? and during a typical month prior to March 1, 2020, when COVID-19 began spreading in the United States, how often did you communicate with friends and family by phone, text, email, app, or using the Internet?. Participant’s response options included five options: “basically every day, a few times a week, a few times a month, once a month, not all, not sure”. We excluded those who selected “not sure”.

History of chronic disease was based on participant’s self-reported response (yes/no) to the following question: Has a doctor or other health –care provider ever told you that you have any of the following: diabetes; high blood pressure or hypertension; heart disease, heart attack or stroke; asthma; chronic lung disease or COPD; bronchitis or emphysema; a cystic fibrosis; liver disease or end-stage liver disease; cancer; and a compromised immune system’?. We defined those who selected “yes” to any of the listed conditions as individuals with at least one chronic condition and those who said “no” to all as not having a chronic condition.

Risk awareness was based on participants’ self–reported response (yes/no) to the following questions: have you had a family member or close friend die from COVID-19 or respiratory illness since March 1, 2020?

Covariates

The following covariates (potential confounders) were included in the study analysis: age categories (18–29, 30–44, 45–59, 60+), gender (male, female), race/ethnicity categories (White, non-Hispanic, Black non-Hispanic, Hispanic, Other, non-Hispanic), household income (under $10,000, $10,000 – under $30,000, $30,000 to under $50,000, $50,000 to under $100,000, $100,000 or more), education categories (no high school diploma, high school graduate or equivalent, some college, BA or above), household size (one person, live by self, two person, three persons, five persons, six or more persons), and census region (northeast, midwest, south, west), population density determined based on 2010 US Census data (rural, suburban, urban), language (English, Spanish).

Data Analysis

Descriptive statistics were performed to summarize and describe the distribution of different variables. Using chi-square (χ2) test statistic, bivariate analyses were performed to compare study participants who used all COVID-19 related preventive behaviors and those who did not use all preventive behaviors by all the study variables. Logistic regression analysis was used to determine the association between safety measures use and study variables. The corresponding odds ratio (OR), 95% confidence interval (95% CI), and p-value were determined. Potential multicollinearity was assessed using the variance inflation factor. We evaluated model fit through inspection of Hosmer and Lemeshow Goodness-of-Fit Test (p = 0.164), implying that the model’s estimates fit the data at an acceptable level. Thus the regression analysis was not prone to multicollinearity. All analyses were conducted using SAS 9.4 (SAS Institute, Inc., Cary, NC). All p-values were two-sided, and statistical significance was set as p < 0.05.

Results

The majority of the study participants were 60 + old (37.8%), female (55.6%), non-Hispanic whites (72.4%), had a household income of $50k to under $100k (33.1%), had education BA or above (52.9%), household size of 2 persons (34.5%), lived in Southern region of US (36.2%), urban (80.5%), spoke English (98.8%), and had at least one chronic condition (77.2%). 44.7% reported that they trusted most people in their neighborhood, 33.4% reported that in the past month they spoke a few times a week with their neighbors. 33.2% reported that they spoke a few times a week with their neighbors during a typical month prior to March 1, 2020. Also, 67.9% reported that in the past month they communicated with their friends and family by phone basically every day. 54.9% reported that they communicated with their friends and family by phone basically every day during a typical month prior to March 1, 2020. 0.6% reported being told by a doctor or other health care provider that they had COVID-19. 0.7% reported being told by a doctor or other health care provider that someone they lived with had COVID-19. 4.7% reported that they had a family member or close friend die from COVID-19 or respiratory illness since March 1, 2020. 79.2% reported undertaking all three preventive behaviors (wear a mask, use hand sanitizer, maintain social distancing). (Table 1)

Table 1 Overall characteristics of the study population and association with safety measures
Full size table

The bivariate analysis found a significant association between the use of preventive behaviors and age, gender, household income, education, household size, region, population density, language, chronic health conditions, trust in neighbors, communication with neighbors, communication with family and friends, perform COVID test, time period, and had a family member or close friend die from COVID-19 or respiratory illness since March 1, 2020. (p < 0.05) (Table 1).

Table 2 shows multivariate logistic regression analysis results. After adjusting for potential confounders, compared to study participants age 60 + years, all other age groups were less likely to follow all three preventive behaviors: 18–29 years (Adj. OR, 95% CI, p-value) (0.44, 0.39–0.50, p < 0.001), 30–44 years (0.55, 0.49–0.61, p < 0.001), 45–59 years (0.70, 0.63–0.77, p < 0.001). Females were more likely to follow all three preventive behaviors than males after adjusting for potential confounders (1.57, 1.46–1.69, p < 0.001). After adjusting for potential confounders, compared to non-Hispanic white, all other races/ethnicities were more likely to follow all three preventive behaviors: non-Hispanic black (1.48, 1.29–1.70, p < 0.001), Hispanic (1.32, 1.15–1.51, p < 0.001), Other (1.16, 1.01–1.33, p = 0.036). After adjusting for potential confounders, compared to study participants with household income of $100k and more, all other income category groups were less likely to follow all three preventive behaviors : Under 10k (0.68, 0.57–0.82, p < 0.001), 10k-30k (0.67, 0.59–0.76, p < 0.001), 30k-50k (0.82, 0.72–0.92, p = 0.001), 50k-100k (0.83, 0.75–0.92, p < 0.001). Compared to study participants with educational degree BA or above, all other educational groups were less likely to follow all three preventive behaviors: No HS diploma (0.76, 0.62–0.94, p = 0.009), HS graduate or equivalent (0.63, 0.56–0.71, p < 0.001), some college (0.73, 0.67–0.80, p < 0.001).

Table 2 Multivariate adjusted logistic regression analysis evaluating the odds of using all three safety measures
Full size table

Compared to study participants with one person household size, those living in a two, three, and five-person household were more likely to follow all three preventive behaviors: Two persons (1.16, 1.05–1.27, p = 0.003), three persons (1.17, 1.03–1.32, p = 0.014), five persons (1.21, 1.02–1.45, p = 0.032). Compared to study participants living in South, those living in the Northeastern region were more likely to follow all three preventive behaviors (1.79, 1.57–2.04, p < 0.001), whereas those living in the Midwestern region were less likely to follow all three safety measures (0.84, 0.77–0.92, p < 0.001). Compared to study participants residing in urban areas, those living in rural and suburban areas were less likely to follow all three preventive behaviors: rural (0.67, 0.57–0.78, p < 0.001), suburban (0.80, 0.72–0.88, p < 0.001).

Compared to English-speaking study participants, Spanish speakers were more likely to follow all three preventive behaviors (1.74, 1.17–2.60, p = 0.007). Compared to study participants with no chronic condition, those with at least one chronic condition were more likely to follow all three preventive behaviors (1.42, 1.31–1.55, p < 0.001). Compared to study participants who did not trust anyone in the neighborhood, those who trusted most people in the neighborhood were more likely to follow all three preventive behaviors (1.17, 1.01–1.36, p = 0.037).

During a typical month prior to March 1, 2020, when COVID-19 began spreading in the United States compared to study participants who did not talk at all with their neighbors, those who spoke basically every day were more likely to follow all three preventive behaviors (1.30, 1.06–1.60, p = 0.012). In the past month, compared to study participants who did not communicate at all with friends and family, those who communicated basically every day (2.18, 1.33–3.59, p = 0.002) or a few times a week (1.75, 1.07–2.89, p = 0.027) were more likely to follow all three preventive behaviors. Compared to study participants whose family members or close friends did not die from COVID-19 or respiratory illness, those who died were more likely to follow all three preventive behaviors (1.43,1.16–1.76, p < 0.001).

Discussion

The current study examined factors that influence COVID19 preventive behaviors. Findings indicated that 79.2% of participants reported practicing the major COVID-19 preventive behaviors, including wearing a mask, washing hands frequently, and watching distance (keeping six feet distance).

Consistent with other similar studies (Bronfman et al., 2021; Clark et al., 2020; Hearne et Nino, 2021; Islam et al., 2021; Kim & Kim, 2020), our study suggests that being non-Hispanic white, Spanish speaking, living in urban areas, of older age, being female, having an education above a BA or income level above $100,000, was associated with the likelihood of following all three preventive behaviors. Regarding language differences, Spanish speakers were more likely to practice all three preventive behaviors. Similar to previous literature, individuals in rural areas and lower socioeconomic status experience major barriers to practicing preventive behaviors due to differential access to resources, inability to take off from work or working in occupations where avoiding close interactions with others is impossible (Chen & Chen, 2020; Hearne & Nino, 2021; Islam et al., 2021: Li et al., 2020). These factors also continue to shape vaccination access and help explain the disparities among racial-ethnic minorities (Singh et al., 2021).

Our findings revealed that individuals with one or more chronic conditions and knowing someone who has died from COVID-19 were more likely to practice all three preventive behaviors. Prior studies show that individuals with chronic conditions (Camacho-rivera et al., 2020; Chen & Chen, 2020; Islam et al., 2021) and knowing someone affected by COVID (has it/died) influences uptake of preventive behaviors due to higher risk perceptions, (Firouzbakht et al., 2021; Kim & Kim, 2020; Li et al., 2020). Furthermore, individuals living in the U.S who had more connections with China and Italy during the early days of the pandemic were more likely to comply with mobility restrictions (Charoenwong et al., 2020). Therefore, health promotion interventions that use a social network approach that relies on collaboration between public health organizations and community “links” or gatekeepers can be effective in disseminating information and adoption of preventive behaviors (Kincaid, 2000). This approach has been effective in promoting the importance of family planning and increasing the contraceptive use uptake globally, by training community links on how to facilitate discussions to clarify any misconceptions and share their adoption of modern contraceptive use (Bormet et al., 2021; Kincaid, 2000; Lowe & Moore, 2014). These community links are usually individuals within organizations already embedded in the communities and collaborating with them can help reach a wide range of community members, especially hard to reach and at risk populations. To illustrate, Borment et al. (2021) found that contraceptive uptake was found to be 1.7 times higher for women exposed to family planning messages from religious leaders than those that were not (Bormet et al., 2021).

A key finding of our study was the role of social engagement, specifically trust and social connectedness in shaping preventive behaviors. Adults more socially engaged, specifically, those that trusted most people in their neighborhood and spoke with family and friends often or a few times a week were more likely to follow all three preventive behaviors. The role of trust in predicting health behaviors has focused on trust of governmental or medical institutions or individuals (Clark et al., 2020). Trust within one’s social networks is an essential factor to explore as communication between those you trust, termed horizontal communication, can help facilitate the spread of information and behaviors to close others (Goldberg et al., 2020).

Furthermore, information that is known and shared among one’s social networks, inclusive of family or neighborhood, can guide and predict an individual’s compliance to a given behavior (Goldberg et al., 2020; Gerber et al., 2021). For example, Goldberg et al., (2020), found that a one-unit increase of perceived social norms among friends and family predicted a doubling of odds of performing various preventive behaviors, inclusive of the three hallmark preventive behaviors. Gerber et al.(2021), reported similar findings in their study participants who perceived a high level of compliance by those close to them were more likely to do all three preventive measures (Goldberg et al., 2020). The extant literature on using social and behavioral science to support the ongoing COVID-19 pandemic response illustrates that social norms are more effective when combined with the social proof of in-group members within social networks (Bavel et al., 2020). Therefore, public health messaging should be adapted to not only provide information on the importance of prevention and vaccinations, but to empower adopters, those trusted community links, to share “I did, you should too” messaging with their immediate social networks to encourage compliance. (Goldberg et al., 2020).

Limitations

Our study findings should be considered with a few limitations. First, the three safety measures were self-reported, therefore responses were subjected to recall and social responsibility bias. Secondly, the adoption of protective behaviors were measured with yes/no responses, therefore unable to explore further how often respondents practiced measures or why they did not adhere to these recommendations. As such, future qualitative studies are needed to ask more detailed questions regarding attitudes and how trust and close communication with family and friends influenced when and where these preventive behaviors were practiced. Despite these limitations, our study has some notable strengths including using a nationally representative sample of the U.S adult population that is diverse across social, demographic, and health status, all of which increased study generalizability.

Conclusion

The social norms within one’s social network can significantly influence health behaviors, such as seatbelt adherence, healthy eating, exercising, alcohol use, and intentions to receive vaccines (Sinclair & Agerstrom, 2021, Smith & Christakas, 2008). Currently, concerted COVID-19 mitigation efforts include increasing vaccinations particularly those at high risk and hard to reach populations. Interventions using a social network approach where trusted community links and adopters are empowered to facilitate discussions on the importance of vaccinations can shape social norms that influence the uptake of COVID-19 preventive behaviors and slowly curb communal spread. Future studies should consider investigating the impact of trust, communication, and language on vaccination behavior. This study adds to what is known about how and where to target messaging. Though a well-studied topic, the continuation of localized hotspots and introduction of new variants warrants continued focus on ways to encourage individuals to continue engaging in key COVID-19 preventive behaviors.