Page 34 - IJPS-9-2

International Journal of
Population Studies COVID-19 and intersectionality in Brazil

of 2020 (Ministry of Health [Brazil], 2020). The study of residence.” “Chest X-ray” – often required for clinical
period was chosen to align with the official declaration investigations of respiratory syndromes and part of the
of COVID-19 community transmission in Brazil, starting protocol for COVID-19 case follow-up – was considered
from the 12 epidemiological week and concluding with an access marker, as it assumes contact with and use of
th
the final data provided by the Brazilian government on public or private health facilities.
June 16, 2020.
Clinical severity markers, such as belonging to a
The target population for this study (n = 284,928) COVID-19 risk group, having been hospitalized, and
consisted of individuals selected based on notifications of having been admitted to an ICU and placed on ventilatory
SARS obtained from the data input of SIVEP-Influenza. support, have suggested the need for healthcare workers to
Specifically, individuals included in the study exhibited provide more appropriate treatment for patients’ clinical-
clinical conditions consistent with flu-like syndromes etiological conditions. Accordingly, it was hypothesized
(fever and cough or sore throat), along with respiratory that there would be a greater probability of undergoing
distress (dyspnea, oxygen saturation below 95% in ambient diagnostic testing among individuals with these clinical
air, or difficulty breathing). severity markers, which are considered potential
Notifications that contained “Missing” or “Unknown” confounding factors in the association being investigated
data regarding diagnostic tests, race, and gender were in this study.
excluded from the study, as these variables were of interest Clinical severity marker covariables were treated
to the study. Additionally, notifications representing race dichotomous (Unrecord/Record). It is important to note
as “Asian descendant” and “Indigenous” populations that the “risk group” variable was specified, considering
were excluded from the study (these populations are the risk groups for SARS, including a positive response for
underrepresented in the databases and may suffer other at least one of the following variables: current pregnancy,
forms of racism). This methodological option was chosen current puerperium, current or previous history of
following the technique employed in a previous study heart disease, hematogenous disease, Down syndrome,
examining ethnicity and gender in Brazil (Smolen et al., hepatopathy, asthma, diabetes, neuropathy, pneumopathy,
2018). immunodepression, nephropathy, obesity, and other
2.2. Measurements and conceptual framework comorbidities (Cen et al., 2020).
The outcome of the present study was determined Education has been identified as one of the
by assessing the failure to perform diagnostic testing determinants of socioeconomic position and influences
for COVID-19 and other respiratory viruses. It was access to material resources and health services (van
determined by the question, “Was a sample collected for Gaans & Dent, 2018). Accordingly, the present study
diagnostic testing?” employed “education” as a proxy for social class. The
conceptual model applied in the present study (Figure 1)
In the study, the variables, namely “race (ethnicity)” thus assumes that “education” has the potential to modify
and “gender,” were considered to be exposures of interest. the impact of the association between race and gender
The categories of “pardo (mixed-race Brazilians)” and on access to diagnostic testing. While the “missing
“black” were combined to form a new category: “black,” data” category was excluded from the analysis of other
in line with the theoretical assumptions of the study. The variables, it was retained for the variable “education” to
other covariables were included in the model due to their examine the statistical modeling’s behavior in relation
potential role as confounding and mediation variables, to the “unknown” stratum. It is important to note that
which could have an impact on the outcome of the the category “does not apply” was assigned to cases of
“diagnostic test.” These covariables were further subdivided children aged 7 years old or younger.
into markers following the relevant literature, as depicted
in Figure 1, which shows the conceptual framework of the 2.3. Statistical analyses
study. The dataset was subjected to, initially univariate, statistical
The variables “age,” “geographical region of residence,” processing using Stata SE 15 software to better understand
and “chest X-ray” were grouped into a theoretical block the data distribution, as well as to calculate proportions
of access markers. The variable “age” was represented in along with their corresponding 95% confidence intervals.
the study as both a continuous variable and a categorical Crude odds ratios (cOR) and p-values were calculated
variable, with the numerical form of age being used using the chi-square test during bivariate analysis. All
in the final model. Access to health-care services is variables included in the conceptual framework were
commonly associated with “age” and “geographical region incorporated in the multiple logistic regression analysis.

Volume 9 Issue 2 (2023) 28 https://doi.org/10.36922/ijps.0865