Forecasting the Long-Term Trends of Coronavirus Disease 2019 (COVID-19) Epidemic Using the Susceptible-Infectious-Recovered (SIR) Model

Round 1

Reviewer 1 Report

This paper presents a study focused on the application of the classic SIR model to predict long-term trends of Covid-19 cases for Singapore, Saudi Arabia, Philippines, and Indonesia. Despite the particular use of a well-studied approach such as SIR, the results seem interesting in the sense that how classical epidemiological models can support long-term predictions, even when they are applied to forecast the Covid-19 pandemic: a disease that leads to several new waves of cases due to the presence of variants. Therefore, I believe the manuscript has some merit for being published. However, important issues need to be resolved before further processing (see the specific comments below). 1) Results covering the use of quantitative error metrics are missing and should be provided (e.g., see references [a,b] below), even in this case where the predictions are computed under a period of months. 2) WHO dataset does not provide the R(t) compartment. How did the authors get R(t)? By computing I(t) and then applying Eq. (4)? 3) In Table 1, how did the authors determine the values for beta parameter? Values for beta in Table 1 were taken randomly or were computed from some equation? Please, state it clearly in the manuscript. 4) The authors could apply the SIR model to data for the year 2021, as new peaks of cases have been observed in several countries during 2021 so that models purely based on SIR may fail. For example, what happened to the confirmed cases/day in Figure 4 during the months of 2021? It is important to point out and discuss such a limitation of SIR, i.e., the model is not capable of producing accurate long-term predictions as well as capturing the real tendency of new waves of cases (e.g., see Fig 1.) 5) Last, the following important references about compartmental epidemic models should be cited in the paper. [a] Ramazi, P. et al. Accurate long-range forecasting of COVID-19 mortality in the USA. Scientific Reports, Nature, 11, 13822, 2021. https://doi.org/10.1038/s41598-021-91365-2 [b] Amaral, F. et al. Towards Providing Effective Data-Driven Responses to Predict the Covid-19 in São Paulo and Brazil. Sensors, MDPI, 21, 540, 2021. https://doi.org/10.3390/s21020540 Minor corrections: - Line 333: range of the 330 parameter R0 value => range of the parameter R_0 value

Author Response

Dear Reviewer 1

We have responded to your comments, our response is attached.

Thank you

Best Regards,

Authors

Author Response File: Author Response.pdf

Reviewer 2 Report

All publications helping to learn and understand the ways of COVID19 transmission are extremely valuable.

The presented work explains the SIR model and the method of data analysis in a very clear and detailed way. I have two little comments: 

in line 249 the S-shaped curve was mentioned - is it possible to add an overview chart?

There is no description of the software used on which the analysis was carried out: was the author's program used or some specialized software, e.g. R package? 

The results of the analyzes are presented clearly and, importantly, in the same way for all analyzed countries. 

The part containing the conclusions contains an objective assessment of the model used in research. The authors show its strengths and (which is common to most mathematical models) its limitations - for example, regarding a limited number of input variables. The results of the analyzes are discussed correctly and can be applied in pandemic management attempts. 

I have one little comment: 

Is it possible to add a summary, comparative statement for all countries (in the form of a table or graph) of parameters (in particular the R₀ parameter) in this part, although presenting all of them would be very helpful in the analysis of the results. 

 

 

Author Response

Dear Reviewer 2

We have responded to your comments, our response is attached.

Thank you

Best Regards,

Authors

Author Response File: Author Response.pdf

Reviewer 3 Report

Presented work is good and upto date of the current pandemic situation. Moreover, few points need to be incorporated in the main paper:

1. According to the title, presented work is related to Long Term Forecasting. But in the paper the definition and type of forecasting is missing. Kindly define the long term forecasting with detailed information and mathematical validation
2. what is the need to forecast in long term way? why not for short term and medium term forecasting. Kindly explain in detail
3. what is the main uses and application of the short-term, medium term and long term forecasting in the pandemic situation. kindly explain in detail
4. The state-of-the-art is very low. kindly enhance it to include all recent publications. then based on the review, kindly tabulate the all possible research gaps and highlight the advantage of each research gap. thereafter, explain in detail at least one research gap which has been covered/ performed/bridged in this paper
5. Kindly include the all possible bullet points related to novelty of the work in the introduction and abstract.
6. Kindly include at least one other method based results demonstration to validate the proposed method.

Author Response

Dear Reviewer 3

We have responded to your comments, our response is attached.

Thank you

Best Regards,

Authors

Author Response File: Author Response.pdf

Reviewer 4 Report

Review of "Forecasting the Long-Term Trends of Coronavirus Disease 2019 (COVID-19) Epidemic using the Susceptible-Infectious-Recovered (SIR) Model" by Kartono et al. (2021)

The authors presented an interesting paper about forecasting the log-term trend of Covid19 disease time series. They considered deterministic models (ODE) and logistic and sigmoid (non-linear) functions. The paper is well written and presented, and the results are important for Covid19 researcher. I have some comments/suggestion for authors:

1. Fix strange numbers that appear around Eq. (17).
2. Authors could include the reference Contreras-Reyes et a. (2014) at lines 288-290 for non-linear estimation.
3. Please include standard deviations for estimated parameters of Tables 1-4. Also, residual diagnostic for fitting performance could be included as Supplementary material.
4. Please include in Appendix the least squares estimators and respective variance used for model parameter estimation.
4. Delete boxes of figures.
5. This work is closed to some few asian countries. In conclusions section, please add a discussion (one paragraph) about comparison of results with the Covid19 development of other countries of Europe (Meintrup et al., 2021).

References:

Meintrup, D., Nowak-Machen, M., Borgmann, S. (2021). Nine Months of COVID-19 Pandemic in Europe: A Comparative Time Series Analysis of Cases and Fatalities in 35 Countries. International Journal of Environmental Research and Public Health 18, 6680.

Contreras-Reyes, J.E., Arellano-Valle, R.B., Canales, T.M. (2014). Comparing growth curves with asymmetric heavy-tailed errors: Application to the southern blue whiting (Micromesistius australis). Fisheries Research 159, 88-94.

Author Response

Dear Reviewer 4

We have responded to your comments, our response is attached.

Thank you

Best Regards,

Authors

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Most of my concerns were not properly addressed by the authors, especially the issue of SIR weaknesses, as the authors did not use the full 2021 data (as well as december 2020) when applying their implementation.

For instance, in Figure 5, the authors just dropped the 2021 data when the curve started to grow again. By doing this kind of intervention, it is really hard to say that the SIR model as implemented by the authors is a good match for predicting long periods of pandemic.

Therefore, the authors should apply their method to the complete data (especially Fig. 5), provide results from their own implementation to be discussed properly, rather than just discussing the limitation of the SIR in the conclusion.

Author Response

see attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Authors have incorporated the valuable comments in the main manuscript.

Author Response

see attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

All of my comments have been adressed. I don't have further comments.

Author Response

see attachment

Author Response File: Author Response.pdf