Tailored Incident Investigation Protocols: A Critically Needed Practice

Round 1

Reviewer 1 Report

The English usage in the paper is fine - there is no need to improve the English language.

I am bothered that the paper does no refer to a number of standard and key OSH aspects.  The first would be ISO 45001, the International Standard on OSH Management Systems - the need for organization specific Investigation Processes is a requirement.

Next, I notice that the Incident Investigation does not ask the question of whether a Hazard Assessment of the equipment has been performed, and whether the assessment was reviewed prior to the work.  Arc Flash Hazard Assessments are covered by NFPA 70E and IEEE1584.  Again, we are back at standards.

There is no mention of a review of the required paperwork - if one is opening up such a device, there should be supervisor approval and a signing off on the work to be done, before the work starts.  Certainly, this is part of the procedure, but it is a separate part of the procedure and it shifts the focus from worker faults to supervisor faults.

After the incident investigation, the findings need to return to the Hazard Assessment stage, and to determine whether the Hazard Assessment needs to be updated.  The paper speaks about using the findings, but one needs to be more specific - and to integrate this into the Plan, Do, Check, Act loop of the ISO OSH standard.

Finally, the construction industry has high accident/incident rates globally.  The construction sector has a weak safety culture, relative to other sectors.  A significant aspect contributing to this lack of culture is the use of contractors, rather than employees.  Contractor management is key - and a "certified contractor" may not be aware of all of the site specific aspects.  The incident investigation only identifies this if it is aware of the aspect.  It is easy to blame a contractor for their own error, rather than to identify the need for active management of the construction site.

I believe that the paper needs to go deeper into a number of these areas.

Author Response

Dear Reviewer 1, 

Thank you so much for your constructive comments. The revised version of the article is much stronger because of you. Please see the detailed responses in the attached file.  

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear author

Overall, the manuscript shows potential. However, there are a lot that need to improved before it can be recommended for publication.

Abstract

1. The methodology is unclear.

2. Key findings are not presented.

Introduction

1. can be improved by critically discussing the incident investigation type and systems.

3.1 may be due to wrong numbering.

Research Methodology

1. Numbering of sub-topics

2. What is the scope of the study? 

3. How many respondents are in total and how are they being selected? Any idea of the population?

4. What is the type of analysis being applied in achieving the findings?

Findings

1. Are the tables shown after the 2 rounds of analysis?

Discussions

1. The findings can be said as expected and not something new. Can it be expanded by looking at both the unsafe conditions and acts at the simultaneously?

2. The number of references can be increased to support the findings and discussions of the manuscript.

Author Response

Dear Reviewer 2, 

Thank you so much for your constructive comments. Please see the detailed responses in the attached file.  

Author Response File: Author Response.pdf

Reviewer 3 Report

The objective of the paper is relevant, but the methodology is rather superficial. I suggest some improvements:

Authors should use a representation (fault-tree, bow-tie, ...) of root and direct causes to facilitate use of the method in the field; see for instance Chevreau et al., 2006, Organizing learning processes on risks by using the bow-tie representation, Journal of hazardous materials, 130 (2006), pp 276-283

This methodology uses a fixed list of causes, which can prevent users to update the list and by that, reduce the learning value of the method. See for example

Drupsteen et al. (2015) Assessing propensity to learn from safety-related events, Safety Science, 71, 28–38

This method is limited to causes to improve prevention of accidents; the study would benefit adressing also protection measures to reduce severity.

Finally, the learning process from incident analysis would also benefit from an extension of the methodology to adress probabilities of causes from the compilation of incidents (at the level of workshop, plant or company) which would help managers to set up the hierarchy of measures (training, control, prevention/protection measures and devices, ...)

Author Response

Dear Reviewer 3, 

Thank you so much for your constructive comments. The revised version of the article is much stronger because of you. Please see the detailed responses in the attached file.  

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed the issues presented in the review.

Reviewer 2 Report

Dear authors,

Most of the comments have been addressed. Please repair the abstract by summarizing the research methodology with significant details such as research design, sampling, etc.

Thank you.

Reviewer 3 Report

Authors have greatly improved the paper which can be published