Adapting an Agent-Based Model of Infectious Disease Spread in an Irish County to COVID-19
Round 1
Reviewer 1 Report
General comments:
I enjoyed reviewing the manuscript of “Adapting an Agent-Based Model of Infectious Disease Spread in an Irish County to COVID-19”.
The authors started with the importance of model prediction of COVID-19 pandemic and the advantage of the agent-based models in disease spread. In general, this paper provided useful information about the application of the agent-based model to predict the COVID-19 pandemic and the influences of relative factors. The model discussed in this model could help to develop effective control strategies for new and emerging infectious diseases.
I recommend the manuscript for publication after minor revisions. Below are some detailed comments and suggestions.
- Line 47 – 51: “Thus having a model that can simulate the society, transport networks and patterns, and environment of a given region or country and that is adaptable to different diseases can reduce the time it takes to create a new model (as compared with creating a complete model from scratch) and thereby get results sooner and help guide decision making during an outbreak.”
The language of this sentence needs to be improved, the author might consider to cut this into two or more sentences. The current structure is too complicated. The first part is the requirement or features of the target model we need, and second part is the benefit of this model, no need to merge into one sentence.
- Line 69 – 70: “In this work we start by briefly discussing the COVID-19 pandemic and then discuss a number of existing COVID-19 models.”
The author could consider breaking to a new paragraph from this sentence.
- Line 123 – 126: “Comparing the results of their equation based models that do not include the spatial spread of the disease with the results of their metapopulation spatial model they determine that their spatial model forecasts that the country will reach its peak four times slower than the model that does not include spatial spread”
The language of this sentence needs to be improved, at least we need a break before “they determine that their spatial model forecast ...”. Also, structure here is too much complicated.
- Line 294 - 295: “Additionally, while there is a existing level of immunity to measles in the population through vaccinations and those who have previously had measles, there is no existing immunity for COVID-19.”
Change “a existing” to “an existing”. Also, remove the word “while”
- Page 7, at the Reverso Context: “The model is run 300 times to account for the stochasticity in the agent-based model. This is based off of methods in [? ].”
Please double-check and show the references.
- For the results and discussion, make sure you define the time steps. it was only shown in Figure 2, but it might be better to be defined in the method part, as well as for the other figures.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
- More than 200000 papers have been published that deal with COVID 19: https://www.nature.com/articles/d41586-020-03564-y
- Many deal with modelling (same reference). Thus,
- presented paper is way too long
- importantly, model is missing
- what is its contribution to above set of publications?
- For publication, the model and (exemplary) data (rest provided online) must be shown, advantage over other models when applied to the dataset presented, text much shortened (<10 pages with references).
-
content of the manuscript
The authors adjust parameters of an existing agent-based epidemic model (used for analysing measles' outbreaks) to the COVID-19 outbreak. -
Manuscript’s strengths and weaknesses;
A novel contribution is missing:
- The (math) model is not shown and explained and it has been published before.
- A thorough comparison with existing models to show the benefits/advantages of the proposed model is missing. Why should one use the proposed model and not another model?
- Moreover, All models (at least that I know of) claim to represent data satisfactorily well, while, here, a thorough statistical analysis of the data fitting procedure is missing. It seems to me as if the authors use the same data set for fitting and predicting.
- On the other hand, if the authors rather wish to present a procedure that can be transferred to other processes then without the (math) model and (an algorithmic) description of how to adjust it, it is difficult for me not only to evaluate the authors' claims but also to see the merit of the paper when it comes to transferring their approach. -
major recommendations for the improvement of the manuscript;
- the authors need to provide novelty. Data-Fitting isn't one, unless, the model that had already been extensively used is better than others in some respects when it comes to analysing/predicting COVID19 (show!), or the data-fitting is novel and robust, or the model has been applied for policymaking, etc
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
This paper shows the feasibility of transforming an agent-based model for the spread of measles to a model for the spread of COVID-19. It is shown that when porting an existing model to a new disease we can alter a few key disease parameters and capture the dynamics of the other disease. The differences in the results are different diseases and parameters. These differences highlight the complex nature of the system and how that complexity needs to be considered when simulating the infectious disease dynamics. The results are interesting. However, there are many technical difficulties and I recommend a rejection and revision. Some comments are as follows.
1. In the paragraph starting line 35, region, age, mobility etc have been taken into consideration. However, some key elements such as immunization has not been mentioned.
2. The novelty and contribution of the work is not sufficiently discussed. It only becomes clear when we read the whole paper.
3. In agent-based epidemic spreading models, an important factor is the awareness to disease. This aspect is overlooked in the work.
4. The adpatation from measles to covid-19 should be justified. Do you have any realistic evidence or calculation that can support this treatment in the particular setting of the current work?
5. Line 349-352, it is not clear how R_0 is obtained in this work. It is of course a key parameter in theoretical models, but it has different formula for different models.
6. Figure 5 and Figure 7 have too small fonts. Please enlarge them if possible.
7. The conclusion is disappointing. The reflection is not critical.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Dear authors,
Many thanks for taking our suggestions into considerations and, in our opinion, making better clear the merit of the paper.
Kind regards
Reviewer 3 Report
I am satisfied with the authors' revision and the responses. The paper can be accepted now.
