The study utilizes data from the 34 urban municipalities of the Metropolitan Region, known as Greater Santiago. For each municipality, two mortality measures were calculated: adjusted mortality rates and potential years of life lost (PYLL). Mortality rates are adjusted by direct standardization, i.e., they are calculated considering the same age structure per administrative unit. Thus, possible confounding effects of the composition of each population are eliminated, allowing comparison between municipalities. For the PYLL, the OECD 2022 methodology was adopted, which also standardizes by age and considers a threshold of 75 years for a premature death24.
The expressions for the standardized mortality rate (SMR) and potential years of life lost (PYLL) are as follows:
$$SM{R}_{it}={\sum }_{a=0}^{K}({d}_{at}/{P}_{at})*{Pr}_{a}*\mathrm{1,000}$$
(1)
$${PYLL}_{it}={\sum }_{a=0}^{L-1}(L-a)({d}_{at}/{P}_{at})*{Pr}_{a}*\mathrm{1,000}$$
(2)
\(SM{R}_{it}\) is the standardized mortality rate per 1000 people using the direct method in the municipality “i” in period “t”. \(K\) is the maximum age reached at the time of death in the municipality “i” in period “t”. \({d}_{at}\) is the number of deaths at age “a” in the municipality “i” in period “t”. \({P}_{at}\) is the number of persons aged “a” in the municipality “i” in period “t”. \({Pr}_{a}\) is the proportion of persons aged “a” in the 2015 OECD population (reference population25). \({PYLL}_{it}\) is the potential years of life lost in the municipality “i” in period “t”. L is the upper age limit established for the calculation of the measure (75 years according to the OECD).
The weighting of deaths at early ages can be observed in the term \((L-a)\) in Eq. (2). Thus, a person who dies at 15 years of age has more weight in the PYLL index than one who dies at 70 years of age, because the first contributes to 60 years of life lost prematurely while the second has only 5 years of life lost prematurely.
Additionally, the absolute potential years of life lost is reported, which is calculated at the regional level as:
$${ABS\_PYLL}_{jt}={\sum }_{a=0}^{L-1}(L-a){d}_{at}$$
(3)
\({ABS\_PYLL}_{jt}\) is the absolute potential years of life lost for the cause of death “j” in period “t”. L is the upper age limit established for the calculation of the measure (75 years). \({d}_{at}\) is the number of deaths at age “a” in period “t”.
To study the evolution of the indicators, two approaches were followed. First, yearly data were used, considering four different periods: two before the onset of the pandemic (2018 and 2019) and two to capture mortality patterns during the pandemic (2020 and 2021). Second, indicators were calculated for each COVID-19 wave. For this purpose, we used a previously defined way to identify COVID-19 waves, based on the scale (case rate) and dynamics (changes in the number of cases) of COVID-19 incident rates26. According to this definition, a wave begins the first week with a weekly incidence rate above 70 cases per 100,000 inhabitants and with a positive growth incidence rate, while a wave ends the first week with a weekly incidence rate below 70 cases per 100,000 inhabitants and with a negative growth incidence rate for at least two consecutive weeks. Thus, for the Metropolitan Region, four study periods are identified, corresponding to four fully completed waves: a first wave from May 3, 2020 to July 26, 2020; the second wave between February 21, 2021, and July 11, 2021; the third wave from October 17, 2021 to November 21, 2021, and; the last wave covering the period between January 2, 2022 and April 24, 2022. To better capture the impact of COVID-19 deaths, two additional weeks were considered after the end of each wave to account for lags between the identification of cases and deaths.
For the analysis of health inequalities in the Greater Santiago area, the concentration index (CI) is computed. This index corresponds to a measure of economic inequality that has been adapted to measure health inequality. Conceptually, it is similar to the Gini index but it is derived from a bivariate distribution of health and social group ranking, and thus is not a measure of total inequality but captures the relationship between socioeconomic ranking and health27,28,29,30. The CI is derived from a concentration curve (\(L(p)\)) which is obtained by plotting the cumulative percentage of a health variable (y-axis) against the cumulative percentage of the population, ordered according to a socioeconomic variable (x-axis). In the hypothetical case that the health variable is evenly distributed in the population, a 45° line called the “line of equality” is drawn; on the contrary, when the distribution is unequal, the curve will lie above/below the line of equality, with a greater distance between the curve and the line meaning greater inequality in health. The CI is calculated as twice the area between the concentration curve and the line of equality31:
$$CI=1-2{\int }_{0}^{1}L(p)dp$$
(4)
All the previously mentioned scenarios can be summarized using the sign and magnitude of the CI. An index equal to 0 indicates equity in the health variable and an index with a negative/positive sign indicates that the burden of the disease is concentrated in the most economically disadvantaged/advantaged population, respectively. In addition, since the index fluctuates between -1 and 1, the closer to these limits the greater the inequity in health.
Analyzing inequalities by year and COVID-19 waves is useful to compare periods before and after the arrival of COVID-19 in Chile (years) as well as to analyze how these inequalities vary during the pandemic (waves). It should be noted that the analysis includes not only the cause of death of suspected COVID-19 cases (patients with symptoms or severe acute respiratory infection) and confirmed COVID-19 cases (patients who test positive in the RT-PCR test for SARS-CoV-2 or in the antigen test when it is a suspect case) but also the total number of deaths (from all causes). This is important to study how the COVID-19 pandemic containment measures and the classification of deaths impacted the rates of other causes.
Data sources
The number of deaths by COVID-19 was obtained from public data provided weekly by the Department of Health Statistics and Information of the Chilean Ministry of Health32. This database records all deaths that have occurred in Chile since 2016, providing the primary cause of death classified according to the International Classification of Diseases (ICD10), in addition to the age, sex, and municipality of residence of the deceased. On the other hand, the population estimated for each municipality by age stratum was obtained from the National Institute of Statistics of Chile, using population projections based on the last national census of 201733. The calculation of the municipality’s average income per household was obtained from the 2017 National Socioeconomic Characterization Survey (CASEN)34. Finally, the confirmed cases of COVID-19 came from the databases of the Chilean Ministry of Health available in a public repository of the Chilean Ministry of Science, Technology, Knowledge and Innovation35.
Ethics approval
The study uses data obtained from open databases from various public institutions. Thus, the data have the characteristic of being anonymous, secondary and aggregated. In the Methods section we mention the different sources of information from which the data were obtained32,33,34,35. Thus, due to the characteristics already mentioned, this study was not submitted to an Ethics Committee.