Predicting Hospital Admissions to Reduce Crowding in the Emergency Departments
Round 1
Reviewer 1 Report
The paper presents an interesting topic and significantly related to the scope of the journal.
The GBM method was adopted to predict the ED admission decision.
Overall, the paper is well written and easy to follow.
The main remarks and suggestions, are summarized below, others are left in the attached file.
- The paper reported that the admission rate is around 11% from the observations, and the proposed model gives an AUC (100-89=11)% as a false negative admission decision rate... which refer to the overhead work that Administrative staff will carry on with no real admission.
Are these numbers equal in term of Administrative work, which is the main motivation in your work Line 101 "
the administrative staff would be able to carry out this process while the patient goes 101 through the ED visit in a simultaneous rather than sequential way." , i.e, it seems that what you save is what you spend.
- can you elaborate on what are the extra Admin works are needed above the Admin work on Trigger phase ( to give the reader a full picture of the admin work)
-A reader from the abstract and Introduction might feel that the paper will examine different methods to predict the ED admissions. However, the paper selected the GMB, only based on other studies results, which is a bit biased, since datasets are different. I suggest if applicable to elaborate on this.. or justify, and to compare GBM method with other methods (e.g, mentioned in ref.26.); or use their dataset if data accessible .
- Table 1: would you provided how the factors are deducted.
- several claims were introduced with no prior citation or proof ( e.g, line 72, 109, ..etc, see the attached file)
- references style should be revised, and to consider more recent works.
- make sure that terminology (e.g, Trigger, admission process ) and abbreviations (AUC, PNV..) are introduced and provided
-Please refer to the attached file for further comments.
Comments for author File: 
Comments.pdf
Author Response
Thank you very much for your comments and suggestions, attached we upload the responses.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Comments for author File: Comments.pdf
Author Response
Thank you very much for your comments and suggestions, attached we submit the responses. The manuscript was clearly improved.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Overall, the authors have provided sufficient clarification and updates.
However, it is recommended to explain (explicitly) why other datasets, which have been used in the literature works, cant be used in your research for the seek of comparison.
Author Response
Thank you for the comment. Any dataset can be used with that methodology. However, what in our opinion makes unique our work is that includes the data from 60 different institutions on a complete National Healthcare System. What is interesting for us, as a company (beHIT), of this work is that by knowing the potential crowding and occupation we can redirect patients from one ED to another.
So, your comment will be considered on a future work to include a better comparison between data sets.
Thank you very much.
