Barriers and facilitators of adherence to social distancing recommendations during COVID-19 among a large international sample of adults

Background Social distancing measures (e.g., avoiding travel, limiting physical contact with people outside of one’s household, and maintaining a 1 or 2-metre distance between self and others when in public, depending on the country) are the primary strategies used to prevent transmission of the SARS-Cov-2 virus that causes COVID-19. Given that there is no effective treatment or vaccine for COVID-19, it is important to identify barriers and facilitators to adherence to social distancing to inform ongoing and future public health campaigns. Method This cross-sectional study was conducted online with a convenience sample of English-speaking adults. The survey was administered over the course of three weeks (March 30 –April 16, 2020) when social distancing measures were well-enforced in North America and Europe. Participants were asked to complete measures assessing socio-demographic characteristics, psychological constructs, including motivations to engage in social distancing, prosocial attitudes, distress, and social distancing behaviors. Descriptive (mean, standard deviation, percentage) and inferential statistics (logistic regression) were used to describes endorsement rates for various motivations, rates of adherence to social distancing recommendations, and predictors of adherence. Results Data were collected from 2013 adults living primarily in North America and Europe. Most frequently endorsed motivations to engage in social distancing (or facilitators) included “I want to protect others” (86%), “I want to protect myself” (84%), and I feel a sense of responsibility to protect our community” (84%). Most frequently endorsed motivations against social distancing (or barriers) included “There are many people walking on the streets in my area” (31%), “I have friends or family who need me to run errands for them” (25%), “I don’t trust the messages my government provides about the pandemic” (13%), and “I feel stressed when I am alone or in isolation” (13%). Adherence to social distancing recommendations ranged from 45% for “working from home or remotely” to 90% for “avoiding crowded places/non-essential travel”, with men and younger individuals (18–24 years) showing lower adherence compared to women and older individuals. Conclusion This study found that adherence to social distancing recommendations vary depending on the behaviour, with none of the surveyed behaviours showing perfect adherence. Strongest facilitators included wanting to protect the self, feeling a responsibility to protect the community, and being able to work/study remotely; strongest barriers included having friends or family who needed help with running errands and socializing in order to avoid feeling lonely. Future interventions to improve adherence to social distancing measures should couple individual-level strategies targeting key barriers to social distancing identified herein, with effective institutional measures and public health interventions. Public health campaigns should continue to highlight compassionate attitudes towards social distancing.

The authors say "Finally, given that men are more likely to die from COVID and older adults are at higher risk of being infected by Sars-Cov-2, it is likely that gender and age could differentially impact adherence to social distancing behaviours.". Can the author cite studies showing that men are more likely to die from COVID-19 and older adults are at higher risk of being infected by Sars-Cov-2? Also, I wonder if the fact that men die more than women and that older adults are more likely to be infected compared to their younger counterparts is the only reason why gender and age could impact adherence to social distancing behaviours differently. It would be good if the authors could expand on the other plausible reasons why we could find heterogeneous results. What does the literature tell about women behaving more cautiously than men, and why may adherence to preventative health behaviours vary by sex and age? Also, if men are more likely to die from COVID-19, they are likely to be more adherent to social distancing recommendations, but even the authors find that men are less adherent to social distancing recommendations.
Partially addressed. The authors have not commented on the last issue 'If men are more likely to die from COVID-19, they are likely to be more adherent to social distancing recommendations, but even the authors find that men are less adherent to social distancing recommendations', and I think they could provide some explanations for why they find that men are less adherent to social distancing recommendations.
The authors say "In the context of the COVID-19 pandemic, it seems reasonable to assume that individual reasons to adhere to social distancing measures (e.g., desire to protect self and others) as well as external circumstances or motivators (e.g., workplace/school conducted remotely) contribute to engagement in and adherence to preventative behaviours, such as social distancing. In addition, individual characteristics, such as demographic and psychological profile (educational level, health literacy, anxiety/stress, empathy towards others) might also play a role in adherence. Finally, given that men are more likely to die from COVID and older adults are at higher risk of being infected by Sars-Cov-2, it is likely that gender and age could differentially impact adherence to social distancing behaviours.". One factor that is lacking in this paragraph is the family. An individual is part of a family. An individual may live with their partner, their kids, etc. They may also live with the most vulnerable people in this pandemic, such as their old parents or a partner with a pre-existing illness. The family composition may be an important socio-demographic predictor of social distancing behavioural outcomes. It would be good to include this factor in the analysis, or, if not available, at least discuss it in the background section.
Partially addressed. A discussion of this factor is still lacking in the background section. The authors should consider adding this factor to the paragraph discussing the sociodemographic variables that might play a role in adherence to social and physical distancing.

Methods
The survey was piloted on 15 individuals whose data were not included in the analysis. It would be good to mention what if this is in line with what is usually done in the literature. Are surveys usually piloted on more/less than 15 individuals? And, what was their assessment of the survey, did they find it easy to complete? Satisfactory.
I could not find the list of motivations for social distancing and social distancing behaviours in this section. The authors should here refer to Table 2 and Table 3 from the results section to allow the reader to know the motivations for social distancing and social distancing behaviours.

Satisfactory.
P. 9 The authors conceptualised adherence to social distancing as "always" endorsing the behaviour (coded as "1") and nonadherence as behaviour endorsed less often than "always", including "never", "sometimes", or "often" response choices (coded as "0"). It would be good to specify why "never" was treated the same way as "sometimes" and "often" and what this could imply for your results. If the reason is purely methodological, then I wonder why not using a tobit model. If conceptual, please specify. Also, please give an example of a behaviour where the "not applicable" option could be used.
Partially addressed. The authors have not commented on the last issue: "please give an example of a behaviour where the "not applicable" option could be used." Does the model include variables for country of residence? Because countries took different approaches (even within the same category 'moderate rules' / 'strict rules', there are differences in the measures adopted), the behaviours may also vary by country.
Partially addressed. Measures taken were different across countries even when countries fell under the same category, and this should be taken into account in the model, are results robust to including country fixed effects (as opposed to countries with strictly enforced guidelines for social distancing etc.)? P. 9 The authors say: "During data collection, recommendations and policies for social distancing differed by region or country but did not change within one region or country, hence our regression models did not account for timing of survey completion.". If, on the one hand, recommendations and policies did not change, on the other, the number of cases and deaths have increased over the period of analysis and this might have changed people's behaviours by for example increasing their adherence to the social distancing measures. It would be good if the authors could account for the passage of time in their analysis.
Not addressed. The passage of time should be taken into account in the model estimation, or the authors should at least check that results are robust to this inclusion.

Results
The results section is very difficult to follow because the results are presented as if they were reported on a presentation with bullet points. The whole section is organised in a similar fashion: "Endorsement rates for the four sets of motivations "for" (facilitators) and "against" (barriers) social distancing are included in Table 2. Highest endorsement rates were found for the following facilitators of social distancing: "I want to protect myself" (84%) and "I want to avoid spreading the virus to others" (83%) (individual-level facilitators); "I want to protect others" (86%) and "I feel a sense of responsibility to protect our community" (84%) (interpersonal-level); "My workplace/ school recommended we practice social distancing" (54%) and "My workplace /school conducts operations remotely" (51%) (organizational-level); "Restaurants in my area are closed for eating-in" (95%) and "Community centers and recreational facilities in my area are closed" (94%) (community-level).". The authors should find a better way to present the results because the way it stands now is not ok.

Satisfactory.
The organizational-level motivations against social distancing stand out for having the lowest endorsement rates, i.e. "My workplace/ school recommended we practice social distancing" (54%) and "My workplace /school conducts operations remotely" (51%). Can the authors speculate why we get such low rates in this cluster?