Identifying the Predictors of Patient-Centered Communication by Machine Learning Methods

Round 1

Reviewer 1 Report

This manuscript by Wu et al studied four machine learning methods to identify important Patient-centered communication predictors based on the nationally representative data of the National Health Information Trends Survey (HINTS) 2019-2020. Nevertheless, the four machine learning methods selected in this study appear arbitrary, and the likelihood of these findings influencing clinical PCC prediction and clinical decision-making is low. 

In addition, The type of strategy presented in the paper is not new.  There have been many related papers published in this field over the past two decades, and the information revealed herein is not beyond the published works. 

The other main problem of the manuscript is its references. In fact, almost all of the cited references in this work are outdated and only a few references are within the most recent three years. Numerous latest and significant publications in this field are missing and must be referenced and discussed in this work. I strongly recommend the authors update the references.

Overall, the information revealed herein is not beyond the published works and I am not able to recommend the publication of this manuscript in Processes.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The study of Wu and colleagues aimed to study the patient communication quality through the use of machine and deep learning approaches. The study employs a reliable data set with h2o tools. The main concern regards the feature interpretability. This should be stressed since represents a crucial point in health informatics and computer aided diagnostic tools. A study that Authors should review for stress the aforementioned aspect in their work is: 

"Identification of Neurodegenerative Diseases From Gait Rhythm Through Time Domain and Time-Dependent Spectral Descriptors." IEEE Journal of Biomedical and Health Informatics (2022)

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

1)      The flow diagram, describing the overall activities of the research, is required.

2)      While describing the machine learning algorithms, the values of the hyperparameters have been outlined randomly. An optimization algorithm must be used at first to get the optimal values of the hyperparameters rather than choosing them randomly. The pseudocode of the algorithm can also be added to make everything clear to the reader.

3)      The dataset is not briefly described. A demographic chart of the dataset will give the reader a clear idea of its composition.

4)      The findings are not clear. The outcome of the research should be outlined clearly.

5)      Machine learning algorithms have been used, but no graphical visualization (ROC curve, precision-recall curve, recall-decision boundary curve, confusion matrix) has been added to support the outcome of the proposed method.

6)      Only P-value, MAE, RMSE, RMSLE, MAPE, and precision have been added as measures of machine learning-based research here. Other modern evaluation measures, such as the Kappa index, accuracy, and area under the ROC curve (AUC), can be added to prove the novelty of the proposed methodology. The evaluation measures mentioned above should be added to the full form if they are used first.

7)      What is the real implementation of this research? The end-users (doctors and clinical staff) may find it ambiguous. So, a decision-support system can be added to assist the end users.

8)      The conclusion section does not contain enough statistics to support the proposed technique. The future direction of this research should be clearly outlined at the end of the conclusion.

 

9)      Additional proofreading is required to improve the structure of the sentences and remove unwanted grammatical mistakes.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

 

The current article titled “Identifying the Predictors of Patient-Centered Communication 1 by Machine Learning Methods” Ref: 2040268 deals with an important subject with understandable language. Few/minor revisions are recommended.

- Enhancement of the conclusion section is highly needed to highlight the observations attained relative to the aim of the study.

- Figure(s) exhibiting the model may be useful for the impact of the study.

-List of abbreviations used in the study should be added.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

This research study shows the value of combining feature selection with machine learning approaches to identify broad variables that could enhance PCC prediction and clinical decision-making, influence future PCC prediction research, and improve patient-centered care.

My comments here are concerned solely with the organization of the manuscript. Consideration of these points will, I believe, lead to an improved report that better illustrates the key concepts and conclusions.

1.     The abstract should be rewritten and improved by combining the objectives, short methodology, main review findings, and prospective application.

2.     It requires further explanation with regard to the Problem Definition, and more technical highlights, and kindly signify the broad objectives as a sub-section after the finish of Introduction.

3.     Summarize the related works in the form of a table.

4.     How author is finding fifteen variables as important predictors?

5.     Elaborate more on the Dataset, what are the highlights and what is the significance of the dataset choosen for this research paper.

6.     Authors need to clearly discuss their proposed architecture, and proposed methodology, and justify their approach, If possible, they can summarize their proposed method in an algorithmic method. Additionally, authors can justify their analysis or results that show the achieved parameters with respect to the use cases and be able to verify their results benchmarked against related existing literature.

7.     Try to concise the conclusion and discuss the future plans with respect to the research state of progress and its limitations.

8.     Author used 34 references but in few references, some details are not available like page number, volume, issue, year etc.

9.     The authors should add other exciting works as there are many research papers in this similar research area. Like:

·       Vishnu Vandana Kolisetty, Dharmendra Singh Rajput*, "A REVIEW ON THE SIGNIFICANCE OF MACHINE LEARNING FOR DATA ANALYSIS IN BIG DATA", Jordanian Journal of Computers and Information Technology (JJCIT) ,Volume 06, Number 01, pp. 41 - 57, March 2020, doi: 10.5455/jjcit.71-1564729835.

·       Chowdhary, C. L., Khare, N., Patel, H., Koppu, S., Kaluri, R., & Rajput, D. S. (2022). Past, present and future of gene feature selection for breast cancer classification–a survey. International Journal of Engineering Systems Modelling and Simulation, 13(2), 140-153.

In a conclusion, the technical content is good. I am accepting an article with minor revision for publication in this journal.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors responded to the reviewers' comments and revised the manuscript (mainly in the Introduction and Discussion). Nevertheless, they did not really revise these key questions raised by reviewers. In my view, the academic novelty and contribution of this work are still relatively limited.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The methodology and ML classifiers are very popular. Still lack of contribution; suggest improving the quality of the paper.  

Author Response

Please see the attachment.

Author Response File: Author Response.docx