Guidelines for Investigation of Logistics Incidents and Identifying Root Causes

# ISSUE 1 - JULY 2015

Introduction
Scope and Objective
1. Incident Investigation
2. The Incident Investigation Process
3. The Root Cause Analysis Method for Logistics Operations
4. Corrective Actions
5. Examples
- 5.1 Example 1
- 5.2 Example 2
Contact List
DISCLAIMER

Introduction

An accident is a sudden event that is not planned or intended and that causes damage or injury. An incident is a sudden event that is not planned or intended and that causes damage or injury or has the potential to do so.

In this guideline both words are used interchangeably because the process to investigate and to take corrective actions is the same.

There exists plenty of guidance on how to investigate incidents. Most chemical companies have identified and developed their own standard method for investigating on-site events. There is a however a need for industry guidance for the investigation of off-site logistics events, to assist Logistics Service Providers (LSP’s) in carrying out incident investigations. The availability of industry guidance should promote more uniformity and provide a common methodology for LSP’s independent of the customer. It will help both transport companies and chemical companies in continuously improving their safety performance by learning from incidents.

Scope and Objective

This guideline focuses on the investigation of incidents and near incidents (near misses) that happen in the off-site transport and related handling of chemical products.
It covers all modes of transport, loading/unloading and operations at terminals, warehouses and tank cleaning stations.

Most chemical manufacturers have identified their investigation method of choice for on-site events. They can, depending on the nature of the incident, decide to use their on-site incident investigation method for on-site logistics events or use the logistics-specific method described in this document.

The objective of this document is to provide guidance on how to carry out an incident investigation, identify the root causes and the corrective actions to prevent reoccurrence. The guideline is aimed for use by all parties in the supply chain: chemical manufacturers, transport companies, distributors, storage companies, tank cleaning stations, etc.

Incident Investigation

1.1 What is an Incident Investigation and Root Cause Analysis

An incident investigation is a process conducted for the purpose of incident prevention which includes the gathering and analysis of information, drawing of conclusions, including determination of causes and, when appropriate, making of safety recommendations.

A root cause analysis (RCA) is a method that allows identification of the true causes of incidents, with the aim of preventing these root causes so that they are not repeated over and over again. It helps to move from goals to clear action plans. The investigation of an incident should always include a root cause analysis.

1.2 Why Perform an Incident Investigation and Root Cause Analysis

There are a number of reasons why an incident investigation and root cause analysis is performed after an incident:

Eliminating the root cause means stopping it from happening again
It is a structured problem-solving technique - an agreed approach that determines underlying causes
It provides permanent solutions

It should also be:

Part of the policy and goals of the organization
A process to provide long term improvement
A powerful vehicle for training people

The analysis of a root cause is a mindset, it takes more time at first but is a 'high return on investment’, eliminating ongoing firefighting. Applying the incident investigation process, including RCA, will structurally lead to an improvement cycle of a company’s management system, processes and barriers used to manage its health, safety, security and environmental risks.

1.3 When to Perform an Incident Investigation and Root Cause Analysis

2 days – 2 weeks rule

The root cause analysis needs to be initiated after no more than two days. The first hours will be used to perform the emergency response and to secure the incident location. Meanwhile, the mind needs time to settle into a no-blame mode necessary for a successful incident investigation.

The incident investigation needs to be concluded and reported within 14 days. After 14 days the feedback becomes more guesswork than factual as witnesses will start forgetting important elements and will start ‘remembering’ things that are actually assumptions.

Each organization should have a clear policy as to which level of root cause analysis is needed depending on the severity of an event. This policy is related to the effects that an incident has for the company in question, like injuries, damage to the environment, material damage or damage to reputation.

The policy defining at which level a root cause analysis is performed should be part of the company’s Safety Management System. A detailed RCA should not only be carried out for severe incidents but also for ‘high potential incidents’ or ‘high learning value incidents’.

1.4 Incident Severity and Categorization

According to the Pareto logic, incidents with increasing severity occur with decreasing frequency in a cascaded design.

1 Fatality

400 Lost time injuries

20,000 minor injuries

240,000 near misses

2,000,000 unsafe acts

Sources: Heinrich, HSE, John Ormond

The Pareto concept also recognizes that 20% of the incidents cause 80% of the damage. By combining those two concepts, incidents can be categorized in several classes. An example of categorization of severity could be:

Major
Moderate
Minor
Near misses
Unsafe circumstances or acts

Incidents commonly cause more than one effect. One can group effects in categories such as:

Human impact
Environment impact
Property and equipment loss

In addition, incidents can result in effects such as:

Financial impact to other parties
Media attention
Reputational damage
Public disruption

For each of the selected categories, clear definitions of severity need to be defined by the company. High potential incidents are incidents which could have caused more severe consequences. These potential consequences are to be considered as well. Investigation of near misses is an obvious application of this principle. More information can be found in different books and on the internet. For example, see the DuPont website (the real cost of safety).

The Incident Investigation Process

When an incident occurs the first step should always be to mitigate the risk of consequential damage, but if possible also to secure the information from the incident as soon and completely as possible.

The second step is to inform the other parties involved in the operation related to the incident. In a logistics operation there is typically more than one company affected by the incident (e.g. the chemical supplier of the product, the customer, the sub-contractor in charge of the operation). The number of different companies that are involved can be a constraint to the incident investigation process. The process should be designed so that it leads to good learning from the incident for all the parties involved.

Based on the criteria defined by the company as described in section 1, it should be determined to what extent the incident needs to be investigated and analyzed.

2.1 Parties Involved

In logistics incidents, there is always more than one party involved. All of them need to be involved in the incident investigation or as a minimum be informed about the outcome. The organization that is in control of the operation when the incident occurs is the party that should lead the incident investigation, unless otherwise agreed. For fatal incidents and for incidents on the public road, the authorities may carry out an official investigation. In this case, all parties will be expected to cooperate in that investigation.

In case of an incident, any liability must be determined via the claims process. The claims process can negatively influence the incident investigation process because both processes have a different focus, possibly opposite to each other. Whereas the claims process is looking for the possibility to attribute liability and blame, the incident investigation process is aiming to prevent the incident from happening again and this goal can only be reached if the notion of blame is left aside.

The first step is to define which parties are involved in a particular incident, who is directly impacted and who needs to be informed. Parties should agree on a common description of the issue, agree upon the type of root cause analysis that needs to be conducted and who should undertake it.

A communication process between the different parties involved must be agreed upon along with the methodology of the incident investigation.

2.2 Setting Up an Incident Investigation Team

Agreement on the incident investigation team composition is essential. A multifunctional team with appropriate skills and ownership should be convened. It can consist of a core team with additional supporting team members.

There must be an “owner” of the incident investigation process in the organization. The “owner” should ensure an effective team composition:

Define who in the chain leads the incident investigation. The leader should be close to, but not part of the line of responsibility for the incident.
Team members must be aligned to the business process.
The resulting corrective actions should be discussed with the operational leader of the unit where the incident took place.
Include a trained facilitator for the root cause analysis process independent and not involved in the incident itself.
The incident investigation team members have to be sufficiently trained or guided through the process.
People involved in the incident should never be part of the team but should be interviewed as part of the investigation process.

For small companies, fulfilling all the above requirements can be an issue. Depending on the severity of the incident, support could be sought from the consigning chemical company or from an independent external investigator. There are different commercial training courses in the market for root cause analysis. Most of them are linked to a certain methodology.

Any root cause analysis method is based on the sequence of events leading up to the incident and should look at the facts and evidence in a structured manner. A distinction between direct causes and root causes should be made.

2.3 Description of the Incident

A detailed description of the incident is essential to ensure a comprehensive analysis and to provide the basic input for the final documentation. The report model of RID/ADR section 1.8.5 can be used as guidance for that purpose.

The incident description should contain the following elements:

When, where, what happened and who was involved:
- Affected operation
- Date/time and location of occurrence
- Environmental conditions such as topography and weather conditions
- Description of the incident as accurate and complete as possible, formulated in a way that it can be understood by anyone who is not involved: what occurred and the primary effect.
Affected elements:
- Product involved and estimated quantity of loss of product
- Type and material of containment (steel tank, plastic drum, …)
- Type of failure of the means of containment
Consequences:
- Personal injury
- Loss of product
- Material/environmental damage
- Evacuation of persons, closure of public roads
- Impact on production and supply performance (delay, customer satisfaction)

2.4 Gathering Evidence and Facts

If it is safe and possible, the evidence should be gathered at the scene of the incident. Look for evidence in people, processes, paper, and parts. When gathering evidence:

Keep an open mind on all the potential activities, situations, or circumstances that can lead to the effect without jumping to conclusions.
Obtain a factual and as complete as possible description of the incident by gathering evidence.
Record only facts, not opinions, and do this as soon as possible.
People involved in the incident are a very important information source.
Pictures are a helpful tool. Include CCTV and onboard camera recording, if available.
Make a drawing of the incident scene.
Unusual or substandard information requires further investigation.

It is important not to allocate blame during the evidence-gathering process in order to ensure that facts and real root cause(s) are identified.

2.5 Root Cause Analysis

After the fact-finding process, in which the investigator should refrain from ‘jumping to conclusions’, it is time for the actual analysis of the facts: the root cause analysis. It is important that this stage of the investigation is performed as a team effort. The team plays an important role and all core team members should be present to perform the incident investigation. The effectiveness of the preventive and corrective actions that will be decided upon will depend on this.

In the supply chain process multiple parties are involved and they are not always part of the investigation team. The investigation is performed on the process of the party carrying out the investigation. If during the investigation it is found out that information from another party is relevant for the investigation of the root cause, this should be reflected in the report without jumping to conclusions on the process of the other party involved. Support should be sought from the consignor or contract party for further investigation.

During the analysis, it is possible that not all evidence is available. In that case, one should go back to the previous step (2.4 Gathering evidence and facts)

The analysis can lead to multiple causes, as well as ‘contributing factors’. Ask the question: is it necessary and sufficient to contribute to the incident? All causes should be investigated up to a level where there is certainty that they are/are not contributing. In this phase, an open mind should be retained. When the process is finished, these events can be put into a schedule which can serve as a reporting tool. It should result in the description of a chain of events that were necessary and sufficient to lead to the incident and the effects.

2.6 Corrective and Preventive Actions

When the root cause(s) are identified, corrective and/or preventive actions should be defined. The actions must be such that the root cause(s) are prevented from happening again. Once the corrective actions have been identified, an implementation plan should be established and communicated.

For examples of corrective and preventive actions see Section 4. The actions must be SMART (Specific, Measurable, Assignable, Realistic, Time-related) and reasonably practical to implement. The effectiveness of the action should be reviewed within a defined timeframe after implementation.

2.7 Reporting an Incident Investigation to the Parties Involved

Intermediate, final, and complete reports should be shared with the stakeholders. The report should be written such that it is easy to understand for a non-specialist or someone who was not involved in the incident investigation. The level of detail should be such that any common industry practice is sufficiently challenged. It should offer a basis to improve the safety management of the organization. After the root cause investigation is completed, the risk assessment of the logistics processes should be reviewed to add the learning.

Legal departments should be consulted on which information can or cannot be shared with third parties. If an investigation is performed by the authorities this can also raise limitations on what can be reported.

A standardized reporting and documentation of the incidents and the related root causes are essential to facilitate a systematic analysis across incidents and to evaluate and cluster common causes.

The following reports should be issued:

An immediate incident notification to the stakeholders
An investigation report for the stakeholders
A report for sharing the learning via Cefic (optional)

2.7.1 Immediate Incident Notification to the Stakeholders

The immediate incident notification to the customer (chemical manufacturer) is usually a quick call which is followed up within 24 hours by the ‘first incident report’. It is necessary to cover the following basic information about the incident (this applies both for dangerous and non-dangerous goods):

Affected transport mode
Date, time and location of occurrence
Topography and particular weather conditions
Short description of occurrence (5 to 10 lines of text)
Product(s) involved
Consequences, e.g. personal injury, loss of product and if appropriate the estimated quantity of spilled product, material/environmental damage, evacuation of persons, closure of public traffic routes

The format of the report might be specified by the charterer or the authorities. For dangerous goods, the ADR/RID demands a report of the incident using a form as specified in section 1.8.5 which can also be used to report the incident to the chemical manufacturer (both for dangerous and non-dangerous goods incidents) in the absence of any prescribed reporting format.

2.7.2 Investigation Report to the Stakeholders

For the reporting of the root cause analysis and the applied corrective actions for risk mitigation, it is recommended to follow the structured approach as described in sections 3 and 4 of this Guidance.

It is recommended that the report is supported by pictures and drawings and to present the different kinds of causes as well as other contributing factors.

The report can be used to share the findings with the stakeholders. The format to present findings and conclusions from the report should be adapted to the audience.

The content of the report should include:

a) The updated information included in the immediate incident notification to the stakeholder b) Immediate actions c) Impact of the incident (see section 1.4) d) For high potential incidents: description of potential consequences e) Chronology and description of events, circumstances, and facts with clarifying illustrations (pictures, plans, drawings) f) Root cause analysis: clarifying all elements that were both necessary and sufficient for this incident to occur (see section 2.5) g) Corrective actions and action plan