Applications of time series analysis in epidemiology: Literature review and our experience during COVID-19 pandemic

Cardiovascular presentations during the COVID-19 pandemic: an interrupted time series analysis

BACKGROUND

The broader implications of the Coronavirus disease 2019 pandemic on cardiovascular hospitalizations remain unclear. We aimed to assess trends in cardiovascular presentations during the Coronavirus disease 2019 pandemic.

METHODS

This multicentre study examined cardiovascular presentations from March 2018 to February 2023. Patients with cardiovascular presentations were identified through administrative health records using ICD-10-AM diagnosis codes. Four key study periods were analysed: T0-pre-pandemic, T1-first lockdown, T2-easing of restrictions and T3-release of restrictions and widespread vaccination. Interrupted time series analysis was used to predict weekly cardiovascular presentations, with the mean difference between actual and predicted numbers assessed for significance.

RESULTS

Overall, 116 518 patients were included across three major public hospitals in Australia. Cardiovascular presentations were significantly lower in T1 than predicted, with a mean decline of 13.1% (SD 16.2%; P = 0.004). There was a significant difference between the expected and actual number of most cardiovascular presentations in T2 and T3, apart from a significant reduction in cardiomyopathy and heart failure presentations during T3 (4.5% [SD 23.7%]; P = 0.007).

CONCLUSIONS

Cardiovascular presentations were significantly lower during the initial lockdown phase of the COVID-19 pandemic; this attenuated with easing of social restrictions and widespread vaccination, except for persistent reduction in cardiomyopathy and heart failure presentations.

[Infection epidemiological effects of school closures during the second COVID-19 pandemic wave-an exemplary analysis within a lower Bavarian region]

BACKGROUND

The rapid global spread of the SARS-CoV‑2 virus during the COVID-19 pandemic led to widespread non-pharmaceutical interventions, such as school closures, to curb infections. This study investigates the influence of school closures on the number of new infections to reduce the existing knowledge gap in this area.

METHODS

The effect of school closures starting on 16 December 2020, was quantified using COVID-19 infection data from the district and city of Landshut in Bavaria between November 2020 and January 2021. An interrupted time series (ITS) analysis was performed to analyze daily new infections among 6-18 year olds and the entire population in the region before and after the school closures using a negative binomial regression model.

RESULTS

The analysis showed that the trend of daily new COVID-19 infections among 6-18 year olds significantly decreased after the school closures (incidence rate ratio (IRR) 0.93; 95% confidence interval (CI): 0.89-0.96; p < 0.001). This was also true for the entire study population (IRR 0.95; 95% CI: 0.93-0.97; p < 0.001).

DISCUSSION

In the context of various concurrent measures to contain the COVID-19 pandemic, the study provides evidence for a significant association between school closures and reduced COVID-19 infections. By optimizing the existing incomplete evidence base on the role of school closures in pandemic control, this finding can support involved stakeholders and contribute to strengthening future approaches to pandemic prevention.