Modeling Pandemic Response through Nonlinear Dynamics in Emergency Medical Systems

INTRODUCTION

The COVID-19 pandemic changed a lot about how emergency hospitals work. It showed that the system wasn’t very resilient and that clinical coordination and infrastructure were problems. The facilities in Northeastern Romania, especially those in Iași and Bacău, were hit the hardest. While under a lot of stress, they had to make tough decisions about how to move patients, how to keep infections from spreading, and how to keep staff available. Conventional linear approaches to hospital risk management, reliant on a direct and proportional correlation between cause and effect, proved inadequate for comprehending or forecasting the temporal evolution of pandemic-induced disruptions [1]. The healthcare system displayed traits typical of nonlinear complex systems, where slight modifications in initial

conditions—like a short delay in triage or the absence of critical personnel [2]—could lead to disproportionately large effects on infection transmission or clinical outcomes [3,4].

The conceptual basis for comprehending such behaviors is rooted in nonlinear dynamics, or chaos theory—a mathematical framework that analyzes systems influenced by feedback mechanisms, wherein outputs perpetually alter their own inputs over time [5]. Including in post-COVID-19 respiratory rehabilitation, fractal theory and nonlinear dynamics were applied [6]. From this viewpoint, hospitals experiencing pandemic stress can be seen as open, adaptable systems made up of interconnected clinical, logistical, and managerial subsystems. Their stability arises from fragile, nonlinear interconnections and constantly evolving external constraints [7]. The transition from orderly to chaotic dynamics in these contexts is frequently characterized by bifurcation phenomena, whereas Lyapunov exponents function as critical metrics for assessing sensitivity to initial conditions and the predictability of the system’s evolution [8].

Mandelbrot [9] was the first to come up with fractal theory. This theory builds on the ideas of nonlinear dynamics by giving us a way to measure and understand how complicated and self-similar dynamic systems are. Fractals show how similar structural patterns repeat at different scales. This was especially clear in hospital risk networks during the pandemic. The patterns of patient influx, infection spread, and resource use over time and space often followed scaling relationships and power-law distributions, which showed that there was a fractal structure [10]. Research shows that pandemics often follow a multifractal pattern over time, with periods of rapid growth followed by periods of partial stabilization [11,12].

The northeastern part of Romania was one of the worst hit by the pandemic because there were a lot of people and not enough resources for intensive care. Hospitals in this area had to change how they worked and how they were built all the time, which led to new behaviors that showed signs of self-organization. Regional assessments revealed that risk transmission—encompassing hospitalacquired infections, staff fatigue, and logistical disruptions—arose from temporal dependencies and nonlinear interconnections akin to those found in complex biological and ecological networks [13].

Using both nonlinear and fractal methods, pandemic risk management can go from a reactive way of dealing with crises to a proactive, predictive, and flexible way of doing things. By finding attractor configurations, measuring fractal dimensions, and keeping an eye on changes in entropy, decision-makers can see important changes between phases of stability and systemic breakdown. This approach aligns with the overarching framework of complex systems theory, which posits that adaptability, feedback regulation, and inter-scale interactions are critical determinants of an organization’s resilience [14].

This research expands upon this theoretical framework by simulating the nonlinear dynamics of pandemic-related risks in an emergency hospital situated in northeastern Romania. It uses mathematical and fractal analytical tools to look at time-series data that shows clinical events, infection trends, and signs of operational stress. The goal is to find the hidden structural patterns and dynamic regularities that shape how hospitals act when they are under a lot of stress. This will help build a quantitative framework that can help with pandemic preparedness and management by supporting strong, data-driven plans.

MATERIALS AND METHODS

Study Context and Design

This study was conducted at the largest tertiary-level emergency facility in northeastern Romania, the “Sf. Spiridon” Emergency Clinical Hospital. There are 25 clinical departments at the institution, and it provides emergency and intensive care to more than 70,000 inpatients each year. The hospital was under a lot of stress during the COVID-19 pandemic (March 2020–December 2022). For example, the number of patients coming in changed, there weren’t enough medical staff, and important departments were at a high risk of infection.

The research utilized a mixed-method observational framework, combining retrospective data analysis with computational modeling. The primary objective was to elucidate and model the nonlinear and fractal dynamics of pandemic-related risks, emphasizing the interplay of clinical, operational, and infrastructural factors in generating novel behaviors and patterns of instability [1,15].

Data Collection and Variables

The hospital’s Risk Management Office and Epidemiological Surveillance Unit gave us the data. The dataset spanned from March 2020 to December 2022 and comprised daily or weekly time series for:

• Clinical indicators: the number of people who go to the hospital with COVID-19, the number of people who get sick in the hospital, the death rate, the number of medication mistakes, and the number of incident reports.
• Operational indicators: the number of people who go to the emergency room, the occupancy rate of ICU beds, the amount of time it takes to triage, the number of staff who are not there, and the amount of time it takes to get supplies.
• Infrastructure indicators include the rate of oxygen use, the number of ventilators in use, and problems with logistics, such as delays in getting protective gear.

All data were anonymized according to GDPR guidelines and approved for use by the hospital’s Ethics Committee (approval code: SPIR-ER-2023-047). Data pre-processing involved normalization, smoothing, and missing-value interpolation using cubic spline functions [16].

Nonlinear Dynamic Modeling

The hospital was modeled as an open nonlinear system consisting of interacting risk subsystems:

𝑑𝑥𝑖 = 𝑓𝑖(𝑥1, 𝑥2, . . . , 𝑥𝑛) + 𝜌𝑖 𝑑𝑡

where 𝑥𝑖 represents a risk variable (clinical, operational, or infrastructural), 𝑓𝑖 is a nonlinear function describing the coupling between variables, and 𝜌𝑖 is random noise representing stochastic external shocks (e.g., pandemic surges, supply delays) [17].

A modified Lotka–Volterra system was employed to simulate the feedback between clinical pressure (x) and operational stress (y) :

where 𝑎 and 𝑐 are growth coefficients, 𝑎 is the system’s carrying capacity (hospital resilience limit), 𝑏 is the interaction rate, and 𝑑 represents the recovery threshold.

By tuning parameters 𝑎 , 𝑏 , and 𝑐 , the model generated transitions from stable oscillations (adaptive phase) to chaotic behavior (systemic overload), consistent with observed real-world instability at peak pandemic phases [18].

Fractal Analysis and Complexity Metrics

Fractal and entropy-based methods were used to measure the temporal complexity of each risk indicator.

1. Higuchi Fractal Dimension (D_H) – measures the time series’ self-similarity and irregularity [19]:

where 𝐿(𝑘) represents the mean curve length over the scale interval 𝑘 .

2. The largest Lyapunov exponent (λ) shows how sensitive the system is to its starting conditions [20]:

A positive λ indicates chaotic evolution and instability in the hospital’s risk state.

3. Shannon Entropy (H) measures how disordered a system is:

𝑛 𝐻= −∑𝑖=1 𝑝𝑖log 𝑝𝑖

where 𝑝𝑖 represents normalized probabilities of risk events (incident frequencies).

4. Multifractal Detrended Fluctuation Analysis (MF-DFA) – used to look at differences between hospital departments and time scales [21,22].

The multifractal spectrum’s width (Δα) shows how scaling behavior changes; a higher Δα means that the structure is more complex and the risk is spread out unevenly across the system.

We used the FractPy, EntropyHub, and NumPy libraries in Python 3.11 for all of the analyses. We used MATLAB R2024b to show attractors and bifurcations.

Validation and Statistical Analysis

To guarantee model robustness, the subsequent validation procedures were implemented:

• Monte Carlo simulations (n = 1000) for testing how sensitive parameters are and how well they converge.
• Bootstrapping for estimating the confidence intervals of D_H and λ values.
• Spearman’s rank correlation to investigate nonlinear associations between variables (e.g., ICU occupancy versus incident frequency).
• Kolmogorov–Smirnov test to check if the distribution is normal.

We did all of the calculations on anonymized datasets, which made sure that everything could be repeated and followed the FAIR principles (Findable, Accessible, Interoperable, Reusable) [20].

RESULTS

General characterization of the hospital system dynamics

The examination of time series data gathered from March 2020 to December 2022 indicated a distinct nonlinear behavior of risk variables, characterized by intricate fluctuations and scale dependencies. The Kolmogorov–Smirnov test found that the temporal distribution of nosocomial infection incidence, COVID-19 admissions, and staff absenteeism was very different from what was normal (p < 0.01).

The time series showed long-term autocorrelation (Hurst exponent H = 0.73 ± 0.04), which means that the system has a memory that lasts for a long time. This finding indicates that present risk fluctuations are heavily influenced by the system’s prior states, demonstrating a self-similar structure across various temporal dimensions (Figure 1).

Fractal dimension and temporal complexity

The Higuchi fractal dimension (D_H) for clinical and operational variables showed values between 1.35 and 1.49, which means that the data were very complex over time (Table 1).

D_H values between 1.3 and 1.5 are characteristic of self-organized critical (SOC) systems [23,24], confirming that the hospital operates in a fragile equilibrium zone between order and chaos (Figure 2).

System stability and Lyapunov exponent

Determination of the Lyapunov exponent (λ) revealed positive values for most analyzed sequences, confirming the presence of deterministic chaos in the evolution of risks (Table 2).

The positive value of λ indicates exponential divergence of trajectories in phase space, implying the impossibility of precise long-term prediction but allowing identification of bifurcation zones—critical thresholds where small variations lead to loss of system stability (Figure 3).

Multifractal analysis

Multifractal analysis (MF-DFA) found that the multifractal spectrum (f(α)) has a width of Δα = 0.62 ± 0.07. This confirms that the risk dynamics in the hospital are not the same for everyone. The most significant changes were seen in:

• ICU and ER units, where Δα > 0.70 – a sign of many interactions and strong feedback between patient flow, staff, and resources;
• Internal medicine and surgical wards, where Δα ≈ 0.55 – a sign of a more stable but still fractal dynamics;

Administrative departments have a Δα of about 0.40, which means they behave in a way that is almost monofractal (predictable). The multifractal plot (Figure 2) showed leftward asymmetry, which is common in systems that have rare but very big fluctuations, like during a pandemic when there are peaks in admissions or new outbreaks in hospitals (Figure 4).

The multifractal plot showed leftward asymmetry, which is common in systems that have rare but very big fluctuations, like during a pandemic when there are peaks in admissions or new outbreaks in hospitals (Figure 5).

Simulation of dynamic behavior

Simulations utilizing the modified Lotka–Volterra model identified three distinct evolutionary regimes:

1. Stable regime (λ < 0): dynamic equilibrium between clinical and operational risk, associated with periods of epidemiological calm;
2. Oscillatory regime (λ ≈ 0): alternation between overload and recovery phases (pandemic waves I–II);
3. Chaotic regime (λ > 0): loss of systemic control, which happens at the peaks of a pandemic (wave III – 2021).

In the latter regime, the model demonstrated exponential growth in the occurrence of clinical errors and nosocomial infections when ICU occupancy surpassed 85%.

Figure 6 shows how the system goes from regular oscillations (limit attractor) to a Lorenz-type chaotic attractor, which has irregular orbits and is more sensitive to changes in parameters.

Correlations between clinical performance and fractal indicators

The increase in both fractal complexity and entropy was reflected in daily clinical activity. During intervals in which D_H and Δα rose sharply in ICU and ER departments, staff documented delays in patient handovers, medication errors, and sudden fluctuations in oxygen consumption. These observations show that nonlinear indicators can highlight approaching instability several days before it becomes evident on the ground.

Correlation analysis revealed substantial nonlinear associations between fractal indicators and performance metrics:

• D_H positively correlates with nosocomial infection rate (ρ = 0.71, p < 0.01);
• λ correlates with average triage time (ρ = 0.64, p < 0.05);
• Δα is negatively correlated with the number of active staff (ρ = −0.68, p < 0.01).

These correlations validate the hypothesis that the fractal complexity of the system escalates with the decline of clinical performance and the diminution of human resources (Figure 7).

Interpretation

The results show that “Sf. Spiridon” Hospital Iași ran during the pandemic in a nonlinear adaptive mode, which is in the middle of order and chaos.

The hospital system showed fractal self-organization, meaning that it stayed functional as a whole even when parts of it were reorganized, but this was only possible when there was more instability and more information entropy.

In general, using nonlinear and fractal dynamics models gives a quantitative picture of how resilient the hospital is and how well it can handle outside shocks without completely breaking down.

Entropy and resilience in hospital systems

During the peaks of the pandemic, entropy analysis showed that information entropy was high (H > 0.8). This means that the system was very messy and the clinical and logistical subsystems weren’t working together very well. High entropy means that things are less likely to happen and less able to control themselves (Figure 8). But moderate levels of entropy during inter-peak times showed that the hospital kept some adaptive mechanisms in place, like moving staff around, redistributing resources, and reorganizing workflows. This made sure that things kept working even when things were crazy outside.

DISCUSSION

The present analysis, although comprehensive, is limited by data accessibility and the difficulty of distinguishing exogenous effects, including policy modifications and regional infection patterns. The models assume that parameters are roughly stationary, but real hospital systems show structural changes. The fact that the fractal and Lyapunov results are the same, though, supports the conclusions.

Summary

In conclusion, the research confirms that the pandemic dynamics at the “Sf. Spiridon” Emergency Hospital followed nonlinear and fractal principles, as shown by self-organization, feedback amplification, and adaptive resilience. These insights support the development of predictive, complexity-focused management systems that can detect early signs of instability and enable proactive measures in future health crises.

Comparison with the Literature

The findings of this study correspond with an increasing body of international literature that emphasizes the nonlinear and fractal nature of pandemic dynamics within healthcare systems. Similar to our findings, other researchers have shown that hospital risk behavior during COVID-19 followed self-organized criticality and multifractal scaling patterns, indicating that complexity and instability are essential traits of adaptive systems [1,3,7].

Nonlinear dynamics and system instability

The positive Lyapunov exponents (λ > 0) identified in the “Sf. Spiridon” Hospital data align with prior research suggesting that hospital operations during the pandemic demonstrated deterministic chaos. Di Matteo (2021) [25] and Lin et al. (2022) [26] demonstrated that epidemic curves, at both regional and institutional levels, exhibit bifurcation cascades and phase transitions, consistent with chaos theory. These findings demonstrate that when vital parameters—such as ICU occupancy or infection rate—exceed a critical threshold, the system transitions abruptly from stability to chaos, a phenomenon also observed in our Lotka–Volterra simulations.

Murray (2017) [18] and Haken & Schiepek (2004) [4] also said that nonlinear feedback loops are the most important parts of biological and organizational systems. Minor alterations in parameters induce a transition from adaptive to unstable phases, demonstrating that deterministic chaos can give rise to novel behaviors within hospital networks.

Fractal scaling and self-similarity

Our study found Higuchi fractal dimensions (D_H) between 1.35 and 1.49, which is in line with the fractal analyses of pandemic time series done by Peng and Goldberger (2000) [11] and Ivanov et al. (1999) [12]. These studies showed that both physiological and epidemic processes have self-similar fluctuations and long-term memory, meaning that the current state depends on how the system was set up in the past.

Bartsch et al. (2020) [27] found that the timing of hospital infections follows power-law distributions in healthcare management, which is similar to the scaling laws found in the “Sf. Spiridon” dataset. This supports the idea that risk propagation is fractal, showing how feedback flows up and down the organizational hierarchy.

Moreover, multifractal detrended fluctuation analyses (MF-DFA) performed in Spanish and Italian hospitals exhibited multifractal widths (Δα ≈ 0.6–0.7) akin to our results, indicating heterogeneous scaling behaviors within hospital departments [14,23]. This variety shows that clinical units don’t all adapt in the same way, which backs up the idea that multifractal metrics can be used to measure systemic risk.

Entropy, resilience, and complexity management

The entropy dynamics observed in this study—characterized by high disorder (H > 0.8) during pandemic peaks, succeeded by partial reorganization—align with the findings of Goldberger et al. (2002) [28], who posited that resilience in complex systems is linked to optimal variability rather than rigid stability. A system that is either fully ordered or entirely chaotic exhibits diminished resilience [29]. This means that hospitals that keep their levels of disorder at a moderate level are better able to adapt [27,28].

Khoshnood et al. (2022) [29] found that hospitals with decentralized decision-making and flexible feedback recovered more quickly from operational collapse during the COVID-19 crisis. These results support our assertion that “Sf. Spiridon” Hospital functioned efficiently amid potential disorder, epitomizing the optimal state for adaptive resilience.

Methodological convergence

This study improves on earlier methods that used nonlinear time-series analysis and fractal modeling for biomedical and healthcare data. For instance, Lipsitz and Goldberger (1992) [30] utilized fractal analysis to define physiological stability, while later research by Casali et al. (2021) [31] broadened this framework to model hospital system resilience during pandemics. The use of Lyapunov

exponents, Higuchi fractal dimensions, and multifractal spectra together gives these theoretical models a quantitative layer of support.

Adding stochastic noise (ε) to our nonlinear equations is similar to what Ivanov et al. (1999) [12] did. This shows that both randomness and determinism can be present in complex healthcare dynamics. This duality is crucial for understanding why pandemic risk cannot be solely represented by linear regression or static indicators.

Regional implications

Most past research has focused on large hospital systems in North America or Western Europe. Conversely, this study examines the Romanian Northeastern region of Eastern Europe, characterized by a consistently resource-deficient healthcare system [30]. The observation that the “Sf. Spiridon” Hospital displayed fractal-chaotic patterns similar to those of advanced systems underscores the universality of nonlinear dynamics in pandemic risk management. It also says that socioeconomic level doesn’t change the basic laws of complexity; they are built into how healthcare networks change [13].

Synthesis

In conclusion, the literature review confirms that the operations of the “Sf. Spiridon” Emergency Hospital during the COVID-19 pandemic are consistent with global evidence depicting healthcare systems as nonlinear, self-organizing, and multifractal structures. The alignment of our findings with both theoretical and empirical research validates the relevance of chaos and fractal theory in risk prediction and management.

Moreover, the integration of fractal complexity metrics (D_H, λ, Δα, and H) into hospital decision-support systems may represent a significant progression towards predictive, adaptive governance, as envisioned by contemporary complexity science.

CONCLUSION

This study provides empirical and theoretical evidence demonstrating that pandemic risk behavior within emergency hospital systems follows nonlinear and fractal dynamics, rather than linear cause-and-effect relationships. Employing methodologies from chaos theory and fractal analysis on empirical data from the “Sf. Spiridon” Emergency Clinical Hospital in Iași, Romania, we identified distinct signatures of deterministic chaos, self-similarity, and multifractal heterogeneity in clinical and operational processes during the COVID-19 pandemic.

The analysis revealed the existence of three primary systemic regimes:

1. The stable adaptive phase, in which the hospital kept things stable by getting feedback and making changes in the area;
2. The oscillatory phase, when things were unstable and resources were too thin;
3. The chaotic phase, when small changes (like too many patients in the ICU or staff not being there) caused big jumps in bad events.

These findings confirm that hospital systems are complex adaptive frameworks operating on the edge of chaos, characterized by a coexistence of adaptability, responsiveness, and vulnerability. Positive Lyapunov exponents (λ > 0), high Higuchi fractal dimensions (D_H ≈ 1.4), and wide multifractal spectra (Δα ≈ 0.6) show that systemic risk grows through processes that are both scale-invariant and self-organizing, which is in line with the idea of self-organized criticality (SOC).

From a management perspective, incorporating nonlinear and fractal indicators into hospital monitoring systems could transform our pandemic preparedness strategies. Fractal dimension, Lyapunov exponent, and entropy function as quantitative indicators of instability, detecting latent transitions before systemic collapse. In this context, hospital risk management should shift from reactive control to predictive and adaptive governance, informed by real-time complexity analysis.

The “Sf. Spiridon” case study also shows that all healthcare systems, no matter where they are or how much money they have, are complicated and strong in the same ways. The hospital showed self-organizing resilience by changing clinical processes on the fly to keep things running during times of crisis, even when resources were limited.

d simulations, network topology analysis, and machine learning algorithms that can find new risk patterns in real time. These efforts could lead to the development of national or regional early-warning systems that utilize nonlinear metrics to transform data from hospitals such as “Sf. Spiridon” into predictive tools for assessing the resilience of healthcare systems.

In conclusion, using nonlinear and fractal thinking in health risk management helps us better understand how systems behave when things go wrong. It helps hospitals change from being weak and reactive to being smart and adaptable, which will help them deal with future pandemics or big changes.

Conflicts of interest and sources of funding

The authors declare no conflict of interest.

This research received no external funding

Acknowledgments

Thank you to all collaborators involved..

Authors’ contribution

Conceptualization, Calin-Stefan Paduraru, and Oana Paduraru; methodology, Elena Costescu; software, Elena Costescu; validation, Letitia Doina Duceac; formal analysis, Calin Paduraru; investigation, Calin Paduraru.; resources, Oana Paduraru; data curation, Elena Costescu; writing—original draft preparation, Calin Paduraru; writing—review and editing, Elena Costescu; visualization, Elena Costescu.; supervision, Letitia Doina Duceac; project administration, Calin Paduraru; funding acquisition, Oana Paduraru All authors have read and agreed to the published version of the manuscript.

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of NAME OF INSTITUTE (protocol code 69 and date of approval 30.08.2023) for studies involving humans.

Patient consent for publication

Not applicable.

References:

1. Haken H. Synergetics of brain function. Int J Psychophysiol. 2006;60(2):110–124. doi:10.1016/j.ijpsycho.2005.12.006
2. Paduraru CS, Paduraru O, Costescu E, Duceac LD. Clinical and infrastructure risk management in an emergency hospital unit in Iași, Romania. Balneo and PRM Research Journal 2025;16: 3 https://doi.org/ 10.12680/balneo.2025.834
3. Goldberger AL, Peng CK, Lipsitz LA. What is physiologic complexity and how does it change with aging and disease? Neurobiol Aging. 2002;23(1):23–26. doi:10.1016/S0197-4580(01)00266-4
4. Haken H, Schiepek G. Synergetics: Introduction and advanced topics. Berlin: Springer; 2004. doi:10.1007/978-3-642-01790-1
5. Strogatz SH. Nonlinear dynamics and chaos: with applications to physics, biology, chemistry, and engineering. 2nd ed. Boulder, CO: Westview Press; 2015. ISBN: 978-0-8133-4910-7
6. Ursescu C, Costescu C, Popescu MN, Țabrean MC, Berteanu M. Modeling Post-COVID Respiratory Rehabilitation with Nonlinear Dynamics and Fractal Metrics: Beyond Spirometry, Balneo and PRM Research Journal 2025, 16(4).
7. Holling CS. Resilience and stability of ecological systems. Annu Rev Ecol Syst. 1973;4:1–23. doi:10.1146/annurev.es.04.110173.000245
8. Wolf A, Swift JB, Swinney HL, Vastano JA. Determining Lyapunov exponents from a time series. Physica D. 1985;16(3):285–317. doi:10.1016/01672789(85)90011-9
9. Mandelbrot BB. The fractal geometry of nature. New York: W.H. Freeman; 1982.
10. Bak P. How nature works: the science of self-organized criticality. New York: Copernicus (Springer); 1996. doi:10.1007/978-1-4757-5426-1
11. Peng CK, Goldberger AL. Fractal mechanisms in biological systems. Chaos. 2000;10(1):182–187. doi:10.1063/1.166487.
12. Ivanov PC, Nunes Amaral LA, Goldberger AL, Havlin S, Rosenblum MG, Struzik ZR, Stanley HE. Multifractality in human heartbeat dynamics. Nature. 1999;399(6735):461–465. doi:10.1038/20924
13. Barabási AL. Network science. Cambridge: Cambridge University Press; 2016.
14. Stanley HE, Amaral LAN, Goldberger AL, et al. Statistical physics and physiology: monofractal and multifractal approaches. Physica A. 1999;270(1– 2):309–324. doi:10.1016/S0378-4371(99)00287-6
15. Sterman JD. Business dynamics: systems thinking and modeling for a complex world. Boston: McGraw-Hill; 2000.
16. Hastie T, Tibshirani R, Friedman J. The elements of statistical learning: data mining, inference, and prediction. 2nd ed. New York: Springer; 2009.
17. Kantz H, Schreiber T. Nonlinear time series analysis. Cambridge: Cambridge University Press; 2004.
18. Murray JD. Mathematical biology I: an introduction. 3rd ed. New York: Springer; 2017. doi:10.1007/978-1-4939-7001-3
19. Higuchi T. Approach to an irregular time series on the basis of the fractal theory. Physica D. 1988;31(2):277–283. doi:10.1016/01672789(88)90081-4.
20. Rosenstein MT, Collins JJ, De Luca CJ. A practical method for calculating largest Lyapunov exponents from small data sets. Physica D. 1993;65(1– 2):117–134. doi:10.1016/0167-2789(93)90009-P.
21. Kantelhardt JW, Zschiegner SA, Koscielny-Bunde E, et al. Multifractal detrended fluctuation analysis of nonstationary time series. Physica A. 2002;316(1–4):87–114. doi:10.1016/S0378-4371(02)01383-3.
22. Wilkinson MD, Dumontier M, Aalbersberg IJ, et al. The FAIR guiding principles for scientific data management and stewardship. Sci Data. 2016;3:160018. doi:10.1038/sdata.2016.18.
23. Ihlen EAF. Introduction to multifractal detrended fluctuation analysis in Matlab. Front Physiol. 2012;3:141. doi:10.3389/fphys.2012.00141.
24. Lorenz EN. Deterministic nonperiodic flow. J Atmos Sci. 1963;20(2):130–141. doi:10.1175/1520-0469(1963)020<0130:DNF>2.0.CO;2
25. Di Matteo T. Complexity in epidemic time series. Chaos Solitons Fractals. 2021;142:110389. doi:10.1016/j.chaos.2020.110389
26. Lin Q, Zhao S, Gao D, et al. A conceptual model for the coronavirus disease 2019 (COVID-19) outbreak in Wuhan, China. Int J Infect Dis. 2020;94:24– 32. doi:10.1016/j.ijid.2020.03.015
27. Bartsch SM, Ferguson MC, McKinnell JA, et al. The potential health care costs and resource use associated with COVID-19 in the United States. Health Aff. 2020;39(6):927–935. doi:10.1377/hlthaff.2020.00426
28. Goldberger AL. Complex systems. Proc Natl Acad Sci U S A. 2002;99(Suppl 1):2466–2472. doi:10.1073/pnas.012582499
29. Khoshnood K, Mehri N, Behzadifar M, et al. Resilience of hospitals during COVID-19: a systematic review. BMC Health Serv Res. 2022;22:119. doi:10.1186/s12913-022-07509-1
30. Lipsitz LA, Goldberger AL. Loss of complexity and aging. JAMA. 1992;267(13):1806–1809. doi:10.1001/jama.1992.03480130097036
31. Casali A, Rinaldi S, Zambon A. Modeling healthcare system resilience during pandemics. Chaos. 2021;31(10):103122. doi:10.1063/5.0063476