The Investigation Process Research Resource Site
A Pro Bono site with hundreds of resources for Investigation Investigators
Home Page Site Guidance FAQs Old News Site inputs Forums
. . . .. . . . . . . . last updated 3/23/07


to advance the
State-of-the-Art of
investigations, through
investigation process

Research Resources:

Search site for::

Launched Aug 26 1996.



Research Report

Reported by K. Priestly WS Atkins (UK) 5/2/97



WS Atkins, in conjunction with the UK Health and Safety Executive (HSE) is undertaking a study to examine the cost effectiveness of good health and safety management. This involves the development of a number of methods to measure health and safety management within an organisation, including auditing techniques and tools to identify the root causes and costs of accidents occurring within the organisation.

These methods build upon the work previously undertaken by the HSE, linking the two publications Successful Health & Safety Management (HS(G)65) and The Cost Of Accidents at Work (HS(G)96) and seeks to test the hypothesis that those companies who excel in safety management also have lower associated costs. The main objective of the research is to undertake a number of case studies using the methodology which will examine the cost effectiveness of health and safety management systems.

Recently a pilot study has been undertaken in a UK company to test these data collection methods, before initiating the full study. The following notes are intended to give a brief overview of our experience of trying to encourage an organisation to undertake root causes analysis of health and safety events within their organisation.


The model of a system for health and safety management developed by the HSE and widely publicised by them formed the basis of the methods subsequently developed. It was a requirement of our sponsors that the findings be related to this model. This model is shown in .

Figure 1: HSE's Health and Safety Management Model

The root causes analysis process followed a number of iterative stages to identify what were termed `Level 1' and `Level 2' management system failures, see . These terms were derived from an expanded version of the HSE's model shown in .

Investigations required interviews, inspections, observations and technical consultations to be conducted to gather evidence about the incident. Events and causal factors charting was used as a method to assist the investigation and identify which Level 2 `risk control systems' had failed. Investigators were required to document which systems were identified as having failed and the evidence on which these decisions were made.

Figure 2: Overall Root Causes Analysis Process


Corporate Acceptance
Written Statements
Practical Policy

Allocation of


Structure and

Accountability Supervision

Identification of HS Objectives

Job Descriptions


Recruitment, Selection/Placement

Key Personnel


(Professional) H&S Advice


Techniques - Information Flows

Visible Behaviour

Verbal Systems



H&S Committees

Workforce Involvement


Hazard Analysis and Risk

Assessment HS Plans

Performance Standards

Setting Priorities

Achievement of and Maintenance of

Risk Controls

Active Monitoring

Workplace Inspections

Checks on Compliance

Incident Investigation


Product/Process Design - Plant Design
/Construction Installation



Selection and Control of Contractors


Start Up/Shut Down

Operation - Working Procedures

Maintenance - Planned


Planned Plant Inspection

Transport (on site) - Materials

Handling and Storage

Pressure Systems

Electrical Safety

Work at Height

Confined Spaces

Control of Change

Permits to Work

Emergency Plans and Procedures


Occupational Health

Product and Service Design

Packaging, Labeling

Storage/Transport (off-site)

Off-site Risks/Plans

Disposal (and Pollution Control)


Figure 2.3.: Expanded Version of the HSE's Health and Safety Management Model

The project team were then required to take the investigations a step further for all the Level 2 systems selected to identify the problems in the management (Level 1) arrangements for those systems.

The problems with the Level 1 management arrangements were defined as the root causes and investigators were expected to identify weaknesses at this level. For each of the elements in the Level 1 management arrangements, see (e.g. organising, planning, implementing etc.), four root causes were defined as follows:

* no system was defined i.e. there is nothing in place to address a particular risk * the system was poorly defined and implemented according to the definition i.e. the system to address a risk has been implemented but it is poorly conceived and defined. * the system was adequately defined but not implemented according to the definition i.e. there is a system defined, but this is not what is implemented * the system was adequately defined but it was not implemented i.e. there is a system defined but it has not been implemented at all.
Within an element only one root cause could be selected, but there was no limit to how many elements could be selected.

Events Investigated

One major assumption of the study is the concept that weakness in high level management systems will underlie many accidents at different levels, e.g. loss time events, dangerous occurrences and near misses. In order to gather as much data as possible for the study during the limited time available, all classes of events that incurred a cost were investigated, both those resulting in harm and those with the potential for harm.

Investigation Experience

The following is a summary of some of the problems encountered with the investigation process:

* Once the Events and Causal Factors chart had been completed the Project Team frequently got side tracked into seeking solutions to the problems identified at this Level 2 stage, rather than investigating further to Level 1 failures. * Where no hard evidence could be obtained it was necessary to examine root causes through discussion and group consensus of the Project Team. This occurred when: - events were associated with the design and installation of old plant and equipment. No information on this equipment was available in terms of a design specification or the procurement process at that time and typically there were no staff available to interview who had been involved in the design and installation of that plant. - systems were undocumented and had evolved either informally or through a corporate/central function with which relationships had changed, or which no longer existed. * The Project Team members did not attempt to identify root causes at Level 1 without the input of the researchers. Even after a number of sessions where analysis had been undertaken as a group, the RCA analysis was not performed alone. One of the team members in particular, only felt comfortable discussing `safe' issues, i.e. those which were not particularly critical of management, for example hazard reporting or inspection systems, maintenance systems. He was also influenced by his expectations of the cost of addressing the failures in the systems identified, preferring only to look at those which he considered it would be `reasonable' to address. * When fatalities or major accidents occur, it is expected by all concerned that there will be an investigation that will require difficult questions to be answered and verified by other sources. Furthermore, it is probably expected that management will come under scrutiny and hopefully this will be a very infrequent event, possibly only occurring once in an individual's career. With this study the investigations were undertaken of lower order events, which typically constituted first aid injuries, near misses or dangerous occurrences. The greater frequency and the lesser consequences of the events under investigation resulted in the members of the project team experiencing some anxiety about the interviews that were required: - With only three departments in the study, certain managers and team leaders were being interviewed repeatedly, there was concern that they would be discouraged from reporting events - Asking in-depth questions about the management of systems in place and searching for root causes at Level 1 could, at times, feel out of proportion to the scale of the event, for example, a first-aid injury resulting in costs of approximately [sterling]25. * Only one LTA occurred during the study. It was understood that the lady involved in the incident was considering taking out a civil claim against the company. While this was the case the Group Safety & Environment manager advised that the RCA should not be conducted due to concerns about disclosure of any documentation/evidence that might affect the outcome of such a case.


A comparison of the frequency of selection of the safety management elements through the root causes analysis process showed a good relationship to the findings from our other measurement tool the health and safety management audit. Both measurement systems revealed that the company had weaknesses in the main areas of Organising (Control), Organising (Communication) and Planning.

The relationship of the cost profile to the audit results was not so strong due to the skewing effect of a small number of high cost events. Potentially this measure might reveal significant differences between high frequency / low cost events (typically surgery attendances in this study) and low frequency / high cost events (typically non-injury events with the potential for harm). This is something that will be explored as the case studies progress.

The accusation of researcher bias could be levied at the research results to date. However, the client, who monitored our methods throughout the study, was pleased with the outcome and is looking to progress the case studies in the coming year. This will require a larger research team which will offer the potential to use different personnel on the different measurement methods in each organisation, which will hopefully address this criticism.