Page 32 - IJPS-5-2

Maternal characteristics and neonatal mortality in Brazil

Figure 1. Flowchart of linkage process with data from SIM and SINASC, Brazil.

For the present investigation, we considered the following maternal characteristics: age (10-14, 15-19, 20-24, 25-29,
30-34, 35-39, 40-44, 45-49, 50, or more years), schooling (0, 1-3, 4-7, 8-11, 12, or more years of study), marital status
(single, married/stable relationship, widowed, and separated/divorced), and race/skin color (white, black/brown, and yellow/
indigenous). About 50% of newborns had missing data on race/skin color and were therefore excluded from the analysis.
Descriptive statistics were performed for all variables analyzed in the study. The NMR was estimated by dividing the
number of deaths of infants less than 28 days of age by the number of live births per 1,000. Decision trees constitute a method
for identifying homogenous subgroups of the population. As only three categorical variables were used in the present study,
we created three decision trees through a descriptive analysis between the NMR and the following combinations: marital status
and years of schooling, age and marital status, and years of schooling and age. Each of the visualizations enables identifying
the relationship between a variable and the NMR and the relationship between two variables combined and the NMR, resulting
in a more specific classification. The other visualization methods used were line graphs, bar plots, and boxplots.
The methods used to treat the data and produce visualizations were the Python programming language (3.6) along
with the Pandas, Matplotlib, and Seaborn modules. LibreOffice Calc and Google Drawings were used for the displays.
3. Results

Between 2006 and 2016, a considerable decrease in infant mortality occurred in Brazil. However, while the post-neonatal
mortality dropped by 27%, neonatal mortality decreased by only 20% in the same period, reaching the nine deaths per
every 1,000 births, as shown in Figure 2.
Figure 3 shows the descriptive characteristics of births in Brazil between 2006 and 2016 and the NMR according to
mother’s age, schooling, and marital status. Most mothers had 8-11 years of schooling (50.5%) and, despite the difference,
mortality rates were similar for mothers with 1-11 years of schooling, whereas mortality rates decreased notably when
mothers had 12 or more years of schooling, which corresponded to 16.2% of the sample. Although a small part of the
population (1.3%), mothers with no education had a much higher risk of losing their infants in the first 28 days, with a rate
at least 18% higher in comparison to the other groups.
In the dataset, the percentage of missing data on marital status was less than 2%. Unmarried women surpassed married
women and those in a stable relationship, accounting for 61.2% of the total. The risk of neonatal mortality was lower when
the mothers were married or in a stable relationship.

26 International Journal of Population Studies | 2019, Volume 5, Issue 2