Impacts of COVID-19 pandemic through decomposition of life … – Nature.com
November 25, 2023
In the analysis, we distinguished four potential types of places of death: deaths at home, in medical facilities (hospitals or other medical facilities), in facilities of social care (social care houses), or at other places. Most CVDs deaths occurred in medical facilities (more than 50% in all the studied years). Before the pandemic, around 26% of CVDs deaths occurred at home annually. During the pandemic, this proportion increased to more than 30% in 2021. COVID-19 deaths occurred mostly (around 90%) in medical facilities (Table 1).
For the studied years, Table 2 (pre-pandemic development) and Table 3 (development during the pandemic) show the annual changes in life expectancy at birth and contributions of deaths from CVDs and COVID-19 according to places of death.
Life expectancy at birth is a summary measure of the current health status of the population under study. If the level of mortality increases, life expectancy decreases, and vice versa. According to contributions of particular causes of death, if mortality from the selected cause increases, the contribution of this cause to a change in life expectancy is negative, i.e., worsening of mortality from any cause of death negatively contributes to a life expectancy change.
Between the years 2019 and 2020, the life expectancy at birth decreased by almost a year. The second pandemic year brought a further decrease in life expectancy, by more than a year (1.03years). The year 2022 was the first year since the pandemic during which life expectancy increased again (+1.76years).
Before the pandemic, a decrease in CVDs mortality contributed significantly to the growth in life expectancy (Tables 2 and 3). Between the years 2017 and 2018, the contribution of CVDs was even higher than the overall increase in life expectancy (+0.14 and+0.17years; the positive contribution of CVDs was moderated by a mortality increase from other causes of death).
A significant change was observed in the first pandemic yearthe contribution of CVDs was negative (0.18years; mortality from CVDs increased) and supported the negative development of life expectancy in 2020. Thus, the higher level of CVDs mortality in 2020 contributed to the overall reduction in life expectancy at birth (Table 3).
In the second year of the pandemic, there was a reversal. Despite the COVID-19 pandemic peaking in 2021, the positive development in CVDs mortality was observable. Improvement in CVDs mortality helped to moderate the life expectancy decrease caused mostly by COVID-19. In 2022, also CVDs contributed significantly (+0.15years) to the overall rapid increase in life expectancy (Table 3).
The next part of Tables 2 and 3 describes the contribution of COVID-19 to life expectancy changes. In the first year of the pandemic, COVID-19 mortality led to a reduction in life expectancy at birth by 0.74years. The second pandemic year was even worse, and COVID-19 itself led to a decrease in life expectancy by 1.19years. In the final year of the pandemic, mortality from COVID-19 decreased again.
Tables 2 and 3 show also the contribution of CVDs and COVID-19 mortality according to places at death. Most of the negative contribution of COVID-19 was due to deaths in medical facilities because most of the COVID-19 deaths occurred in hospitals (9092%). Vaccination against COVID-19 was initiated during the first months of 2021, preferably from the oldest age groups, or in facilities of social care. This is reflected in the positive contributions of COVID-19 mortality at facilities of social care to life expectancy change between 2020 and 2021.
The situation was different according to CVDs. Before the pandemic, contributions of CVDs mortality were around zero at all places except for medical facilities where decreasing CVDs mortality helped to increase the overall life expectancy. Whereas the CVDs mortality rates in hospitals improved until 2019, CVDs mortality rates out-of-medical facilities were almost stable (Table 2).
In the first pandemic year (2020), the negative contribution of CVDs mortality to the overall change in life expectancy was mainly due to higher CVDs mortality at home and in social care houses. At these two places, the level of CVDs mortality worsened, i.e., the number of CVDs deaths at home or social care houses increased the most. In 2021, CVDs mortality in social care houses and in medical facilities improved and helped to decrease CVDs mortality. The contribution of CVDs mortality at home was already close to zero, but still negative (supporting the decrease in life expectancy). In the final pandemic year, 2022, contributions of CVDs mortality regardless of the place of death were positive again (Table 3).
Figure1 shows the overall development of CVDs health care provision and mortality in Czechia in time, i.e., development of the studied time seriesCVDs hospitalizations (panel A), CVDs ambulant care (B), CVDs deaths in medical facilities (C), at home, and in facilities of social care (D). The monthly data are adjusted for the length of particular months in the period January 2018December 2022. Clearly, in the time series, a strong seasonal pattern can be seen (see Fig.2Seasonal component of time series decomposition). There is also a long-term decreasing trend in some seriesabove all CVDs hospitalizations (Fig.1A) or CVDs deaths in medical facilities (Fig.1C). This trend started already before the pandemic. On the other hand, there was an increase in CVDs ambulant care before the pandemic, which was interrupted in 2020 (Fig.1B). The number of CVDs deaths at home or in facilities of social care had almost a stable trend during the studied period, however, with high variability (Fig.1D). The visible peak at the end of 2020 could be considered as an indirect effect of the pandemic and will be discussed later in more detail. The significant exceptional increases or decreases in the development are depicted in Fig.3.
Source:14,15,16, authors calculation, output of the SAS software, version 6.4.
Time series of monthly data adjusted for the length of particular months, January 2018December 2022, CVDs hospitalizations (A), CVDs ambulant care (B), deaths in medical facilities (C), deaths at home and facilities of social care (D), Czechia.
Source:14,15,16, authors calculation, output of the SAS software, version 6.4.
Seasonal component of time series decomposition (1=average month corresponding to the overall trend of the time series), January 2018December 2022, CVDs hospitalizations (A), CVDs ambulant care (B), deaths in medical facilities (C), deaths at home and facilities of social care (D), Czechia.
Source:14,15,16, authors calculation, output of the SAS software, version 6.4.
Irregular component of time series decomposition (1=average month corresponding to the overall trend and seasonal factor of the particular month of the time series), January 2018December 2022, CVDs hospitalizations (A), CVDs ambulant care (B), deaths in medical facilities (C), deaths at home and facilities of social care (D), Czechia. Note: the dotted horizontal line at value one represents the reference level corresponding to an expected values of the time series for a particular month reflecting the trend and seasonal pattern.
The seasonal patterns (Fig.2) of the analysed series are stable in time (in contrast to more traditional approaches to time series seasonal decomposition, the used X-13 methodology potentially allows for moving seasonal components, slightly developing in time). Values around one represent an average month within a year. There is a traditional decrease in the number of CVDs hospitalizations as well as CVDs ambulant care during the summer and in December each year (by some 20%, represented by values around 0.8 in Fig.2). On the other hand, the peak of CVDs mortality repeats annually in the first quartile of the year (regardless of the place of death).
Figure3 shows the irregular component of the time series decomposition. It reveals the unexpected or exceptional changes in the studied time series. Values around one correspond to expected values of the time series for a particular month reflecting the trend and seasonal pattern. For CVDs health care, there is an abrupt drop in March and April 2020, in both time series the values of provided health care decreased in those months to almost 70% of expected values (values around 0.7 in Fig.3). This decrease is partially replaced by an exceptional increase in CVDs ambulant care during the summer of 2020, however, this increase was only about 20%. Another decrease in the number of CVDs hospitalizations as well as ambulant care was observable at the end of 2020. The observable decrease in CVDs hospitalizations at the very end of 2022 may correspond to the strong flu epidemics at that time.
Several peaks of CVDs mortality were seen during the studied years. The first peak occurred at the beginning of 2018. This likely corresponds to weather conditions and flu epidemics typical for this time of the year, however stronger in the studied year. Except for these smaller or bigger peaks repeating similarly (however, not strictly regularly so as to be included in the seasonality pattern) every year during the first months (reinforcing the long-term seasonal effect), there are no strong deviations from random fluctuations around one (corresponding to average months). The change occurred with the start of the pandemicin April 2020, there was an increase in the number of CVDs deaths at home by 10% in comparison to an average April in the studied years and an increase of almost 9% in deaths in facilities of social care. This corresponds to the observed decrease in CVDs health care or hospitalization in March or April of that year. For the sake of completeness, this time, from March 12th to May 17th, 2020, a State of Emergency was declared in Czechia24.
An even more significant increase in CVDs deaths occurred, however, at the end of 2020above all in October 2020 (increase by 12% in medical fac., 15% at home, and 29% in facilities of social care) and November 2020 (increase by 45% in facilities of social care, the increase in deaths at home or in medical facilities was around 23%). At this time, the next State of Emergency was declared in Czechia in the period from October 5th, 2020, to February 14th, 202124.
The last exceptional increase in deaths at home occurred in November 2021 by 14% and at medical facilities (hospitals) in December 2022 by 12% which likely corresponds to an already mentioned flu epidemic.
More:
Impacts of COVID-19 pandemic through decomposition of life ... - Nature.com
