Forecasting intensive care unit demand during the COVID-19 pandemic: A spatial age-structured microsimulation model

  1. Sebastian Klüsener
  2. Ralf Schneider
  3. Matthias Rosenbaum-Feldbrügge
  4. Christian Dudel
  5. Elke Loichinger
  6. Nikola Sander
  7. Andreas Backhaus
  8. Emanuele Del Fava
  9. Janina Esins
  10. Martina Fischer
  11. Linus Grabenhenrich
  12. Pavel Grigoriev
  13. André Grow
  14. Jason Hilton
  15. Bastian Koller
  16. Mikko Myrskylä
  17. Francesco Scalone
  18. Martin Wolkewitz
  19. Emilio Zagheni
  20. Michael M. Resch

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The COVID-19 pandemic poses the risk of overburdening health care systems, and in particular intensive care units (ICUs). Non-pharmaceutical interventions (NPIs), ranging from wearing masks to (partial) lockdowns have been implemented as mitigation measures around the globe. However, especially severe NPIs are used with great caution due to their negative effects on the economy, social life and mental well-being. Thus, understanding the impact of the pandemic on ICU demand under alternative scenarios reflecting different levels of NPIs is vital for political decision-making on NPIs.

Objective

The aim is to support political decision-making by forecasting COVID-19-related ICU demand under alternative scenarios of COVID-19 progression reflecting different levels of NPIs. Substantial sub-national variation in COVID-19-related ICU demand requires a spatially disaggregated approach. This should not only take sub-national variation in ICU-relevant disease dynamics into account, but also variation in the population at risk including COVID-19-relevant risk characteristics (e.g. age), and factors mitigating the pandemic. The forecast provides indications for policy makers and health care stakeholders as to whether mitigation measures have to be maintained or even strengthened to prevent ICU demand from exceeding supply, or whether there is leeway to relax them.

Methods

We implement a spatial age-structured microsimulation model of the COVID-19 pandemic by extending the Susceptible-Exposed-Infectious-Recovered (SEIR) framework. The model accounts for regional variation in population age structure and in spatial diffusion pathways. In a first step, we calibrate the model by applying a genetic optimization algorithm against hospital data on ICU patients with COVID-19. In a second step, we forecast COVID-19-related ICU demand under alternative scenarios of COVID 19 progression reflecting different levels of NPIs. We apply the model to Germany and provide state-level forecasts over a 2-month period, which can be updated daily based on latest data on the progression of the pandemic.

Results

To illustrate the merits of our model, we present here “forecasts” of ICU demand for different stages of the pandemic during 2020. Our forecasts for a quiet summer phase with low infection rates identified quite some variation in potential for relaxing NPIs across the federal states. By contrast, our forecasts during a phase of quickly rising infection numbers in autumn (second wave) suggested that all federal states should implement additional NPIs. However, the identified needs for additional NPIs varied again across federal states. In addition, our model suggests that during large infection waves ICU demand would quickly exceed supply, if there were no NPIs in place to contain the virus.

Conclusion

Our results provide evidence for substantial spatial variation in (1) the effect of the pandemic on ICU demand, and (2) the potential and need for NPI adjustments at different stages of the pandemic. Forecasts with our spatial age-structured microsimulation model allow to take this spatial variation into account. The model is programmed in R and can be applied to other countries, provided that reliable data on the number of ICU patients infected with COVID-19 are available at sub-national level.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2020.12.23.20248761: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    The package ggplot2 along with functions from the packages grid, gridExtra, pracma and RColorBrewer are used to visualize the simulation results.
    ggplot2
    suggested: (ggplot2, RRID:SCR_014601)

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    Limitations of our approach include that our reduction factor μts captures any kind of factor that affects the infection risks, including also effects that cannot be controlled by policy makers or societies such as seasonal changes in weather conditions. However, further refinements of the models would offer the possibility to isolate some of these factors. Over the course of the pandemic we might obtain a better understanding of the seasonal effects, and could incorporate them into the forecasts. Reduced COVID-19 related ICU demand due to improvements in medical treatments could be accounted for by adjusting our epidemiological parameters over time, based on data from epidemiological studies. If the quality of incidence data further improves as the result of increased test capacities, we might also be able to better account for how infection intensities vary by age across sub-national regions at a given time. While focusing on ICU coverage offers the advantage that we are able to work with informative data, one potential drawback is that shifts in the COVID-19 dynamic become visible only somewhat later in the ICU occupancy than in the incidence data. This is not necessarily a disadvantage. Over the summer, Germany had experienced a number of surges in slaughterhouses among rather young healthy employees, which created local hot spots in the incidence data, but had little effect on the ICU data. Thus, focusing on the incidence data would have likely created too high estimates...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a protocol registration statement.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  2. Version published to 10.1101/2020.12.23.20248761v1 on medRxiv