gms | German Medical Science

Introduction: In Germany and many other countries, the COVID-19 pandemic has impacted heavily on health systems. A major goal of governmental interventions focused on preventing an overload of intensive care unit capacity - especially in the light of the limited number of ventilators available for treating severe COVID-19 cases: When demand for ventilation surpasses availability, some form of selection and triage is necessary.

The aim of our study is to compare two different triage strategies in the context of COVID-19: triage based on maximizing (1) survival probability and (2) life expectancy of the patients.

Methods: We used the agent-based simulation system AbstractSwarm (a graphical modeling and simulation software) [1], [2] to create different hypothetical scenarios of patient demand and ventilator availability, according to recent COVID-19 numbers from the University Medical Center Mainz, resulting in triage and non-triage scenarios. The data for the patients’ ages, gender distribution, survival probabilities and other criteria were extracted from the literature and the German federal statistics agency [3], [4].

Simulations were modeled with up to 14 times the recent number of COVID-19 patients that needed to be ventilated at our hospital at end of March 2021. Algorithms that favored either survival or life expectancy were compared: The former allocates the ventilator to the patient with maximal survival probability, the latter favors the youngest patient. Results are presented in terms of comparative plots.

Results: Progressive overload of the health system increasingly impacts both survival and life expectancy of patients requiring ventilation. Moreover, we find that the two different triage strategies had limited impact on outcome parameters.

In a tenfold overload scenario, employing a triage strategy based on maximum survival probability, only 71.5% of the lives and 78.5% of the patient agents’ lifetime were salvaged compared to a corresponding no-triage scenario with a sufficient number of ventilators. Using maximum expected lifetime as a triage strategy, 71.7% of the lives and 77.1% of the lifetime were saved according to our simulation (see Figure 1 [Fig. 1]).

In this model, increasing the number of younger patient agents (< 60 years) by 20% (leading to an average age of 60.8 instead of 67.2 years) did not change the outcome based on the two triage strategies.

Discussion: In our hypothetical cohort of COVID-19 patients, using different triage strategies had minimal impact on the outcome. A reason for these findings could be the advanced average age of the actual patients on which our models were based. Further reasons may have been the relatively high overall mortality of ventilated patients as well as the fact that younger patients having with a higher expectancy of life usually also have a higher survival probability.

Conclusion: According to our simulation experiments, avoiding a triage scenario by assuring an adequate number of ventilators is the most effective measure to save lives and lifetime. Once initiated, the triage strategy in itself plays a minor role. Hypothetical models can help to understand the potential impact of real-life triage scenarios.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.