Why is it difficult to accurately predict the COVID-19 epidemic?

Longitudinal study of interventional radiology activity in a large metropolitan Italian tertiary care hospital: how the COVID-19 pandemic emergency has changed our activity

Objectives

To retrospectively analyze interventional radiology (IR) activity changes in the COVID-19 era and to describe how to safely and effectively reorganize IR activity.

Methods

All IR procedures performed between January 30 and April 8, 2020 (COVID-era group) and the same 2019 period (non-COVID-era group) were retrospectively included and compared. A sub-analysis for the lockdown period (LDP: 11 March–8 April) was also conducted. Demographic, hospitalization, clinical, and procedural data were obtained for both groups and statistically compared with univariable analysis.

Results

A total of 1496 procedures (non-COVID era, 825; COVID era, 671) performed in 1226 patients (64.9 ± 15.1 years, 618 women) were included. The number of procedures decreased by 18.6% between 2019 and 2020 (825 vs 671, p < .001), with a reduction by 48.2% in LDP (188 vs 363, p < .0001). In the LDP COVID era, bedside procedures were preferred (p = .013), with an increase in procedures from the intensive care unit compared with the emergency department and outpatients (p = .048), and an increased activity for oncological patients (p = .003). No incidents of cross-infection of non-infected from infected patients and no evidence of COVID-19 infection of healthcare workers in the IR service was registered.

Conclusions

Coronavirus disease outbreak changed the interventional radiology activity with an overall reduction in the number of procedures. However, this study confirms that interventional radiology continuum of care can be safely performed also during the pandemic, following defined measures and protocols, taking care of all patients.

Key Points

• Coronavirus disease pandemic determined a reduction of interventional radiology activity as compared to the same period of the previous year.

• Interventional radiology procedures for life-threatening conditions and non-deferrable oncologic treatments were prioritized as opposed to elective procedures.

• Strict adoption of safe procedures allowed us to have until now no incidents of cross-infection of non-infected from infected patients and no evidence of COVID-19 infection of HCWs in the IR service.

Simulation of the Final Size of the Evolution Curve of Coronavirus Epidemic in Morocco using the SIR Model

Since the epidemic of COVID-19 was declared in Wuhan, Hubei Province of China, and other parts of the world, several studies have been carried out over several regions to observe the development of the epidemic, to predict its duration, and to estimate its final size, using complex models such as the SEIR model or the simpler ones such as the SIR model. These studies showed that the SIR model is much more efficient than the SEIR model; therefore, we are applying this model in the Kingdom of Morocco since the appearance of the first case on 2 March 2020, with the objective of predicting the final size of the epidemic.

Multiple Ensemble Neural Network Models with Fuzzy Response Aggregation for Predicting COVID-19 Time Series: The Case of Mexico

In this paper, a multiple ensemble neural network model with fuzzy response aggregation for the COVID-19 time series is presented. Ensemble neural networks are composed of a set of modules, which are used to produce several predictions under different conditions. The modules are simple neural networks. Fuzzy logic is then used to aggregate the responses of several predictor modules, in this way, improving the final prediction by combining the outputs of the modules in an intelligent way. Fuzzy logic handles the uncertainty in the process of making a final decision about the prediction. The complete model was tested for the case of predicting the COVID-19 time series in Mexico, at the level of the states and the whole country. The simulation results of the multiple ensemble neural network models with fuzzy response integration show very good predicted values in the validation data set. In fact, the prediction errors of the multiple ensemble neural networks are significantly lower than using traditional monolithic neural networks, in this way showing the advantages of the proposed approach.

A Mathematical Model for the COVID-19 Outbreak and Its Applications

A mathematical model based on nonlinear ordinary differential equations is proposed for quantitative description of the outbreak of the novel coronavirus pandemic. The model possesses remarkable properties, such as as full integrability. The comparison with the public data shows that exact solutions of the model (with the correctly specified parameters) lead to the results, which are in good agreement with the measured data in China and Austria. Prediction of the total number of the COVID-19 cases is discussed and examples are presented using the measured data in Austria, France, and Poland. Some generalizations of the model are suggested as well.

Prediction of Epidemic Peak and Infected Cases for COVID-19 Disease in Malaysia, 2020

The coronavirus COVID-19 has recently started to spread rapidly in Malaysia. The number of total infected cases has increased to 3662 on 05 April 2020, leading to the country being placed under lockdown. As the main public concern is whether the current situation will continue for the next few months, this study aims to predict the epidemic peak using the Susceptible-Exposed-Infectious-Recovered (SEIR) model, with incorporation of the mortality cases. The infection rate was estimated using the Genetic Algorithm (GA), while the Adaptive Neuro-Fuzzy Inference System (ANFIS) model was used to provide short-time forecasting of the number of infected cases. The results show that the estimated infection rate is 0.228 ± 0.013, while the basic reproductive number is 2.28 ± 0.13. The epidemic peak of COVID-19 in Malaysia could be reached on 26 July 2020, with an uncertain period of 30 days (12 July-11 August). Possible interventions by the government to reduce the infection rate by 25% over two or three months would delay the epidemic peak by 30 and 46 days, respectively. The forecasting results using the ANFIS model show a low Normalized Root Mean Square Error (NRMSE) of 0.041; a low Mean Absolute Percentage Error (MAPE) of 2.45%; and a high coefficient of determination (R2) of 0.9964. The results also show that an intervention has a great effect on delaying the epidemic peak and a longer intervention period would reduce the epidemic size at the peak. The study provides important information for public health providers and the government to control the COVID-19 epidemic.

A Predictive Model for Patient Census and Ventilator Requirements at Individual Hospitals During the Coronavirus Disease 2019 (COVID-19) Pandemic: A Preliminary Technical Report

During the initial wave of the coronavirus disease 2019 (COVID-19) pandemic, many hospitals struggled to forecast bed capacity and the number of mechanical ventilators they needed to have available. Numerous epidemiological models forecast regional or national peak bed and ventilator needs, but these are not suitable for predictions at the hospital level. We developed an analytical model to assist hospitals in determining their census and ventilator requirements for COVID-19 patients during future periods of the pandemic, by using their data. This model is based on (1) projection of future daily admissions using counts from the previous seven days, (2) lengths of stay and duration of mechanical ventilation, and (3) the percentage of inpatients requiring mechanical ventilation. The implementation is done within an Excel (Microsoft, Redmond, WA) workbook without the use of add-ins or macro programming. The model inputs for each currently hospitalized patient with COVID-19 are the duration of hospitalization, whether the patient is currently receiving or has previously received mechanical ventilation, and the duration of the current ventilation episode, if applicable. Data validity and internal consistency are checked within the workbook, and errors are identified. Durations of care (length of hospital stay and duration of mechanical ventilation) are generated by fitting a two-parameter Weibull distribution to the hospital's historical data from the initial phase of the pandemic (incorporating censoring due to ongoing care), for which we provide source code in the R programming language (R Foundation for Statistical Computing, Vienna, Austria). Conditional distributions are then calculated using the hospital's current data. The output of the model is nearly instantaneous, producing an estimate of the census and the number of ventilators required in one, three, and seven days following the date on which the simulation is run. Given that the pandemic is ongoing, and a second surge of cases is expected with the reopening of the economy, having such a tool to predict resource needs for hospital planning purposes has been useful. A major benefit to individual hospitals from such modeling has been to provide reassurance to state and local governments that the hospitals have sufficient resources available to meet anticipated needs for new COVID-19 patients without having to set aside substantially greater numbers of beds or ventilators for such care. Such ongoing activity is important for the economic recovery of hospitals that have been hard-hit economically by the shutdown in elective surgery and other patient care activities. The modeling software is freely available at https://FDshort.com/COVID19, and its parameters can easily be modified by end-users.

Temporal Dynamics of COVID-19 Outbreak and Future Projections: A Data-Driven Approach

Long-term predictions for an ongoing epidemic are typically performed using epidemiological models that predict the timing of the peak in infections followed by its decay using non-linear fits from the available data. The curves predicted by these methods typically follow a Gaussian distribution with a decay rate of infections similar to the climbing rate before the peak. However, as seen from the recent COVID-19 data from the US and European countries, the decay in the number of infections is much slower than their increase before the peak. Therefore, the estimates of the final epidemic size from these models are often underpredicted. In this work, we propose two data-driven models to improve the forecasts of the epidemic during its decay. These two models use Gaussian and piecewise-linear fits of the infection rate respectively during the deceleration phase, if available, to project the future course of the pandemic. For countries, which are not yet in the decline phase, these models use the peak predicted by epidemiological models but correct the infection rate to incorporate a realistic slow decline based on the trends from the recent data. Finally, a comparative study of predictions using both epidemiological and data-driven models is presented for a few most affected countries.

Estimation of COVID-19 dynamics "on a back-of-envelope": Does the simplest SIR model provide quantitative parameters and predictions?

Basing on existence of the mathematically sequential reduction of the three-compartmental (Susceptible-Infected-Recovered/Removed) model to the Verhulst (logistic) equation with the parameters determined by the basic characteristic of epidemic process, this model is tested in application to the recent data on COVID-19 outbreak reported by the European Centre for Disease Prevention and Control. It is shown that such a simple model adequately reproduces the epidemic dynamics not only qualitatively but for a number of countries quantitatively with a high degree of correlation that allows to use it for predictive estimations. In addition, some features of SIR model are discussed in the context, how its parameters and conditions reflect measures attempted for the disease growth prevention that is also clearly indicated by deviations from such model solutions.

100 Days of COVID-19 in India: Current and Future Trends

COVID-19 classified under emerging infectious diseases has spread across the world resulting in many casualties. India reported its first case on 30th January 2020 and has completed 100 days in this pandemic period. The government of India has issued stringent guidelines and imposed a lockdown for long periods to ensure the practising of social distancing. This paper reviews and discusses the current trends in the confirmed cases of India in comparison with the other prominent countries around the globe. A novel approach using a sigmoid function to predict and forecast the trends for cases in India are also presented in this work. By placing the current time on the sigmoid curve, forecasting the total number of confirmed cases by the end of the pandemic is made. If proper measures and stringent guidelines are not followed, India may have to endure a total confirmed case of up to 4.4 lakh. The prediction also suggests that 99.9% of the pandemic may end in India by 27th July 2020. The best possible approach is to undertake preventive measures by strictly adhering to the guidelines and policies set by the government. Performing hand hygiene, practicing social distancing, surveillance and isolation is the only means to break the chain of transmission and control the pandemic.

Forecasting the Spread of COVID-19 in Kuwait Using Compartmental and Logistic Regression Models

The state of Kuwait is facing a substantial challenge in responding to the spread of the novel coronavirus 2019 (COVID-19). The government’s decision to repatriate stranded citizens back to Kuwait from various COVID-19 epicenters has generated a great concern. It has heightened the need for prediction models to estimate the epidemic size. Mathematical modeling plays a pivotal role in predicting the spread of infectious diseases to enable policymakers to implement various health and safety measures to contain the spread. This research presents a forecast of the COVID-19 epidemic size in Kuwait based on the confirmed data. Deterministic and stochastic modeling approaches were used to estimate the size of COVID-19 spread in Kuwait and determine its ending phase. In addition, various simulation scenarios were conducted to demonstrate the effectiveness of nonpharmaceutical intervention measures, particularly with time-varying infection rates and individual contact numbers. Results indicate that, with data until 19 April 2020 and before the repatriation plan, the estimated reproduction number in Kuwait is 2.2. It also confirms the efficiency of the containment measures of the state of Kuwait to control the spread even after the repatriation plan. The results show that a high contact rate among the population implies that the epidemic peak value is yet to be reached and that more strict intervention measures must be incorporated

To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic

Face mask use by the general public for limiting the spread of the COVID-19 pandemic is controversial, though increasingly recommended, and the potential of this intervention is not well understood. We develop a compartmental model for assessing the community-wide impact of mask use by the general, asymptomatic public, a portion of which may be asymptomatically infectious. Model simulations, using data relevant to COVID-19 dynamics in the US states of New York and Washington, suggest that broad adoption of even relatively ineffective face masks may meaningfully reduce community transmission of COVID-19 and decrease peak hospitalizations and deaths. Moreover, mask use decreases the effective transmission rate in nearly linear proportion to the product of mask effectiveness (as a fraction of potentially infectious contacts blocked) and coverage rate (as a fraction of the general population), while the impact on epidemiologic outcomes (death, hospitalizations) is highly nonlinear, indicating masks could synergize with other non-pharmaceutical measures. Notably, masks are found to be useful with respect to both preventing illness in healthy persons and preventing asymptomatic transmission. Hypothetical mask adoption scenarios, for Washington and New York state, suggest that immediate near universal (80%) adoption of moderately (50%) effective masks could prevent on the order of 17-45% of projected deaths over two months in New York, while decreasing the peak daily death rate by 34-58%, absent other changes in epidemic dynamics. Even very weak masks (20% effective) can still be useful if the underlying transmission rate is relatively low or decreasing: In Washington, where baseline transmission is much less intense, 80% adoption of such masks could reduce mortality by 24-65% (and peak deaths 15-69%), compared to 2-9% mortality reduction in New York (peak death reduction 9-18%). Our results suggest use of face masks by the general public is potentially of high value in curtailing community transmission and the burden of the pandemic. The community-wide benefits are likely to be greatest when face masks are used in conjunction with other non-pharmaceutical practices (such as social-distancing), and when adoption is nearly universal (nation-wide) and compliance is high.

COVID-19: The unreasonable effectiveness of simple models

•

COVID-19 outbreak belongs to the simple universality class of the SIR model and extensions thereof.

•

The unpredictable non-stationarity of the testing frames behind the figures reported by national authorities sets a fundamental limitation to any theoretical approach.

•

The time evolution of the reporting rates controls the occurrence of the apparent epidemic peak, which typically follows the true one in countries that were not vigorous enough in their testing at the onset of the outbreak.

When the novel coronavirus disease SARS-CoV2 (COVID-19) was officially declared a pandemic by the WHO in March 2020, the scientific community had already braced up in the effort of making sense of the fast-growing wealth of data gathered by national authorities all over the world. However, despite the diversity of novel theoretical approaches and the comprehensiveness of many widely established models, the official figures that recount the course of the outbreak still sketch a largely elusive and intimidating picture. Here we show unambiguously that the dynamics of the COVID-19 outbreak belongs to the simple universality class of the SIR model and extensions thereof. Our analysis naturally leads us to establish that there exists a fundamental limitation to any theoretical approach, namely the unpredictable non-stationarity of the testing frames behind the reported figures. However, we show how such bias can be quantified self-consistently and employed to mine useful and accurate information from the data. In particular, we describe how the time evolution of the reporting rates controls the occurrence of the apparent epidemic peak, which typically follows the true one in countries that were not vigorous enough in their testing at the onset of the outbreak. The importance of testing early and resolutely appears as a natural corollary of our analysis, as countries that tested massively at the start clearly had their true peak earlier and less deaths overall.