Accident Analysis

This part of the course is designed to assist investigators in conducting investigations following IMO Resolution A..849 (20), the Code for Investigation of Marine Casualties and Incidents.

At the end of the course, the student will be confident in following the systematic. As stated in Section 2, Resolution A884(21) (Unit 4.6 – section 2), the essential sequence of events for investigating any accident is described in the graphic below.

Principles of Accident Analysis

Accidents are investigated to identify causes and determine actions required to prevent a recurrence. Investigators should carefully examine events and conditions that create accident or incident situations. 

In understanding sequences of events and conditions, two principles are that accidents are the results of successive events producing unintentional harm and that accidents occur during a period of work activity. A chart of events and causal factors depicts the sequence of events (and conditions linked to those events) leading to an incident and can be used to analyse that incident.

Investigators need to identify and chart or document, events, and conditions affecting each event in an accident sequence. Construction of the chart should begin as soon as relevant facts are obtained, initially preparing a skeleton chart, which can be built up as additional facts emerge. The chart of events and conditions will be useful in organising data, guiding investigation, confirming the accident sequence, assisting with identifying causes, and with the preparation of reports.

A benefit of using charts constructed by using events and conditions is that the chart illustrates multiple causes. Accidents rarely have single causes, and charts show various causes in an accident sequence. The essential sequence of events for investigating any accident is:

• Collect occurrence data.
• Determine the occurrence sequence.
• Identify unsafe acts, decisions, and unsafe conditions.

Then, for each unsafe act and decision:

• Identify the error type.
• Identify underlying factors.
• Identify potential safety problems.
• Any accident follows a sequence of events that might be depicted as a timeline or a series of parallel timelines.
• The start of the timeline can be challenging to establish, and the termination of the timeline is the time of the casualty itself.

Building a Hypothesis

One of the main reasons for using investigators skilled in the maritime disciplines is that they can empathise and will understand how an accident may have occurred. It allows an investigator to build a hypothesis as evidence is gathered. Two crucial points are:

• Any theory must be based on the evidence in front of the investigator.
• Remember the human factor and that an investigator also is susceptible to decision-making errors such as “false hypothesis or “confirmation bias.”

The need to keep an open mind while exploring any hypothesis is essential. There is always a temptation to favour facts that lean towards the way the investigator is thinking. For this reason, the ability to discuss and debate any hypothesis with peers within an administration can prevent an investigator from embracing a false theory. The techniques in this chapter will help in following objective procedures.

Ways of Looking at an Accident

Accidents and incidents need investigating and analysing to determine what happened and why and to prevent similar accidents in the future. Various models have been introduced to facilitate investigation and analysis. In 1991 Professor James Reason wrote in a paper to the 22nd Annual International Society of Air Safety Investigators:

“Many decades of air accident investigators have created well-stocked databases that, in history at least, should establish the relative significance of such causal factors as pilot error, mechanical failure, weather, inadequate maintenance, and other personnel failures. But a glance at some recent statistical analysis shows that this is far from being the case.”

A report to the Flight Safety Foundation in 1986 claimed that mechanical failure preceded by faulty maintenance was the leading cause of air accidents. However, the Chairman of the US National Transportation Safety Board (NTSB) told the press in 1987, that bad weather near airports caused 64% of major crashes in the previous five years. In 1989, a Lufthansa World Accident Survey found that cockpit crew errors accounted for 76% of all crashes.

Whom Should We Believe?

Professor James Reason suggests that we should believe none of these findings and that even though failures in maintenance, air traffic control, and crew performance were not uncommon, these factors were insufficient by themselves to create accidents. Accidents took place when malign chance with these elements in moments of system vulnerability. Reason’s model for accidents and their causes includes:

• Latent failures (fallible decisions by management and line managers).
• Preconditions (such as distractions).
• Active failures ( such as a slip or unintended deviation from a correct plan of action).
• Defences that were inadequate in preventing an occurrence.

Accident Modelling

Another model uses the concept of a “chain of events” and depicts causes leading to accidents as links of a chain, with an accident or incident occurring at the weakest link, and other links leading to the accident or incident at the weak link, being made up of latent factors, preconditions, and active failures.

Two Swedish institutions, the Lund Institute of Technology and the Karlstad Risk Centre have elaborated upon the concept of a “chain of events.” They have analyzed the capsize of the Herald of Free Enterprise by looking at the causes of isolated decisions.

They used a diagram showing areas where decisions were made without checking the effects of these decisions upon other operational areas, resulting in the capsize of the vessel.

The SHEL Model

Any operational system includes several significant elements, one of which is the human element, and all of which interact in such a way that their total effect is larger than the sum of their parts. The SHEL model helps to understand human factors. The name is taken from the initial letter of its four parts:

Liveware

In the centre of the model, the hub, is a person, the most critical as well as the most flexible component in the system. However, people are subject to considerable variation in performance between individuals, or on an individual may be subjected. It is essential to try and match the other components in the system to the Liveware to avoid stress, which will eventually lead to failure. To achieve such a match, the following characteristics are essential:

Physical Size and Shape

The design of the working environment and equipment and general ergonomic principles may have to be adapted to factors such as gender, age and ethnic characteristics.

Physical Well Being

People need food, water, air and sleep.

Input Characteristics

Humans have a sensory system for collecting information by sight, hearing, touch, and smell, all of which, either singularly or in combination, help them to respond to external physical and psychological performance over time.

Information Processing

Humans have severe limitations in short and long-term memory and some mental processing activities. In poorly designed instrumentation, these limitations have led to ambiguity and inappropriate action and deductions.

Output

The appropriate physical response, once the senses system has initiated a response and the brain has processed the perceived information, relies on all of the above and the environment.

Environmental Tolerance

Temperature, humidity, noise, time of day, light, darkness, atmospheric pressure, and aroma; all reflect on performance. A dull, stressful working environment can be expected to degrade human performance.

Liveware/Hardware

This interface is the one most often considered when regarding the human-machine (equipment) system. Because all the fact that humans are adaptable and can make allowances for less than-optimal design, deficiencies in Hardware may not be identified until after a disaster. It does not make the deficiencies any less real.

Liveware/Software

This interface is between human and system procedures (manual and checklist formats, symbology, and computer programs). In an accident, a mismatch is difficult to identify, but often stems from “contradictory software” or “misunderstood software.”

Liveware/Environment

Mismatches between humans and the environment can easily be identified at sea. Ships operate a 24-hour society all over the world. The ship’s movement, particularly in bad weather, vibration, engine noise, taken with temperature, humidity, and the need for the OOW (Officer On Watch) to operate in darkness are examples of such environmental factors.

Liveware/Liveware

It is the interface between people. Between individuals on a ship, between the bridge team and pilot, between the ship’s staff and shore management, and between port officials and the master. It is highly complex, but an essential element in all human performance.

Events and Causal Factors Charting

Events and causal factors charting is a method used by the United States Department of Energy, enabling investigators to probe deeply into events and conditions that create accidents and to determine their causes. An accident involves a sequence that occurs in the course of well-intentioned activity, culminating in an injury or damage.

The chart depicts in sequence, the events and causes leading to an accident and it can be used to analyse the accident and evaluate evidence. The chart also assists with identifying events and the conditions that affect each event in the accident sequence.

Accidents rarely have single causes, and the chart illustrates multiple causes in accident sequences.

Tests of Safe Operation

Although it is not the purpose of an investigation to attribute liability or blame, judgment of individual and collective acts of commission and omission play an important part in any accident analysis. Investigators are blessed with perfect hindsight, hindsight, which must be used to make an accurate and professional analysis of the casualty to prevent such an accident from occurring again.

It is equally important for the credibility of the investigation that the analysis (and the hindsight) should be realistic. It is useful to remember The Substitution Test mentioned earlier in this course, and it is also useful to apply the following six tests of safe operation.

• Was the casualty foreseen or foreseeable?
• Was the equipment in use fit for the purpose?
• Were the systems and procedures sufficient to maintain safe operation?
• Were the staff members fit, competent, and capable?
• Were the emergency procedures and defences effective?
• Was there a management system to monitor and improve performance?