Data for this analysis are from a larger study on Brazilian immigrant women’s health priorities and experiences. A convenience sample was recruited from Brazilian social media pages (e.g., Facebook, WhatsApp) and through outreach by collaborating community partners (e.g., health and social service providers in Boston, MA), who posted information and a link to the study on their social media pages. Those eligible to participate were Brazilian women age 18 years or older who resided in the U.S. The online survey was offered in Portuguese and English; respondents could choose based on their language preference. The survey was translated by an American Translators Association certified translator and was reviewed by native Portuguese speakers and an expert in Portuguese linguistics to ensure that the translation was linguistically and culturally accurate. Survey respondents were provided with a financial incentive ($20 gift card) for survey completion, which took an average of 18.5 minutes. Data was collected between July and August 2020 and all study procedures were approved by the Institutional Review Board at Tufts University, Medford, MA, USA.
We assessed our primary outcome (COVID-19 vaccine intentions) by asking: “If a vaccine became available to prevent the Coronavirus, would you want to get it?” (“yes,” “no,” “don’t know”). For those who responded that they would not get the vaccine we asked, “Why not?” and respondents were able to enter free text responses. We also assessed prior testing for and diagnosis of COVID-19 (“Have you been tested for the Coronavirus? If so, what was the result?)” with response options: “I have been tested and I tested positive (I had coronavirus),” “I have been tested and I tested negative (I did not have coronavirus),” “I have been tested and I do not know the result,” or “I have not been tested.” In addition, respondents were asked, “Whether or not you have had a Coronavirus test, has a doctor or another healthcare professional diagnosed you as having or probably having Coronavirus?" (“yes,” “no,” “don’t know”).(19)
Given that the Brazilian president has downplayed the severity of the COVID-19 pandemic (20), we also assessed perceptions regarding the seriousness of the pandemic with a question asking respondents whether they thought that the pandemic was “a significant crisis,” “a serious problem but not a crisis,” “a minor problem,” or “not a problem at all.”(21). To plan future interventions, we also wanted to know about trusted source of health information, so we asked “What source do you trust the most to give you accurate up-to-date information about health?” with response options: doctor/physician, a nurse practitioner/nurse; network news; family member or friends; social media; internet (e.g., WebMD, Google); public health agencies (e.g., Centers for Disease Control and Prevention); governmental agencies or officials; pharmaceutical companies; religious or faith leaders.
Socio-demographic characteristics were assessed using items from the Brazilian Census(22), including race/ethnicity (White/Black/Indigenous/Asian/Pardo [“mixed”]), and educational attainment. We also assessed age, household income, and insurance status using items from the Behavioral Risk Factor Surveillance System (23). We obtained participant zip codes since the survey was not limited to one geographic area. We asked about the number of years lived in the U.S., because prior evidence suggests that longer time spent living in the U.S. is associated with adoption of health behaviors that are normative in the U.S. (24–26). Given that acculturation appears to change beliefs, attitudes, and values regarding health behaviors and healthcare among immigrants (27), we assessed nativity, time living in the U.S., as well as language spoken at home and with friends.
Analysis
We eliminated respondents who had missing data on vaccine intentions (n = 1), leaving a final analytic sample of N = 364. Descriptive statistics, including percentages, means, standard deviations, and ranges, were produced for all variables. We evaluated bivariate associations between vaccination intentions (outcome), key independent variables (e.g., COVID-19 experiences and perceptions, trusted sources of information) with Chi-square and ANOVA tests (19).
Before analysis, we combined categories for some variables, due to small cell sizes (n < 10). Specifically, we re-categorized answers for marital status (into “formerly married or living as married,” “married or living as married,” and “single, never married”), perceived significance of pandemic ("serious problem but not a crisis,” “a minor problem,” and “not a problem at all” combined into one category, “minor” versus “a significant crisis” relabeled as “major”), and most trusted source of health information (doctor/physician and nurse practitioner/nurse combined to “healthcare provider”; “public health agencies” and “governmental agencies” combined to “public agencies”; “social media” and “family members or friends” combined to make “social networks”; “internet” and “network news” combined “private news sources”).
Variables associated with vaccine intentions (p < 0.05) in bivariate analyses and socio-demographic characteristics were included in multivariate models. Two multinomial logistic regression models were run to examine relationships between vaccine intentions, independent variables, and socio-demographic characteristics. The models compared those who did not intend to be vaccinated with those who reported that they would get a vaccine (“yes” vs “no”), as well as with those who reported that they were unsure about vaccine intention (“unsure” vs “no”). Variables were added in blocks with the first model examining socio-demographic demographics (age, marital status, income, time in the U.S). The second model included those socio-demographics, as well as perceived significance of the pandemic, and most trusted source of health information (28, 29). Statistical significance was considered at the p < 0.05 level for the final model. Data analysis was generated using Stata software.(30)
Analysis of free text responses (reasons for not wanted vaccine) was done by thematic analysis. Based on published literature on vaccine hesitancy (31), we categorized rationale in terms of vaccine concerns (e.g., vaccine had not been fully tested, could have serious side effects, was not effective) and trust in authorities charged with developing and/or administering the vaccine.