Accidents, causes, investigation and prevention: book summary

Accidents: causes, investigation and prevention by Doctor James Thornhill is an introductory book to the Author’s same named course that also goes by the acronym ACIP. This book is a deep dive into a professional process of what happens after injury, accident, fatality or other severe incident has already happened.

What should be done, investigated, communicated, reported and by whom? Moreover, how can we prevent those adverse events in the first place? This blog post aims to give a short summary to these questions but to understand the authors thought more deeply, we recommend reading the book.

Three ideas to remember:

  1. The aim of accident investigation is to prevent accidents, to save lives and to save money.
  2. A structured investigation procedure improves results and recommendations
  3. Accident investigation is not only about human errors but looking at systems


Some terminology first how Dr. Thornhill decides to define important HSE terms. In the book when talking about accidents, they cover events that result in loss. Loss can be a fatality, injury, ill health, damage or other type of loss. When an incident is mentioned, it refers to an event where loss did not take place but had potential for it (near miss) or conditions and circumstances where in place for bad outcome (undesired circumstance). Adverse events cover both accidents and incidents. 

Herbert Heinrich’s triangle is familiar to almost all who work within Safety Management. The author raises an interesting misconception on the research Heinrich conducted. One of the study’s results was that 88% of accidents were caused by human error. However, the majority of the reports were conducted by supervisors and considering the safety climate of the 1930s: who else would these supervisors blame if not their employees?

Of course how safety is seen today is a lot different from how it was in the 1930’s and what often changes things is society’s attitude. Accidents and other negative safety outcomes are not as acceptable today as they used to be. At least part of this we can thank Henrich and his team who found that for every major accident, there are a lot of incidents taking place. These results are not isolated to just Henrich’s study but repeated on multiple occasions. For example Frank Bird’s team found in 1969 that for every major accident, there are 600 incidents. Another study conducted by ConocoPhillips Marine found out that for every fatal accident more than 300 000 at risk behaviours take place. If only organisations could investigate more of the minor incidents instead of costly accidents.

What hasn’t changed though is that people’s assessment of risk is rarely accurate. Not going into detail on that topic as it contains many major and even more minor factors, the author points out that we are too often more afraid of unknown risks which we have little control over, but we do accept risks which we know and can control more.

To reduce risks and more specifically unsafe acts, some organisations have used the Three E’s model which stands for Engineering, education and enforcement. This model is especially important in Behaviour Based Safety programs that heavily depend on this thought. It is however a model that focuses a lot on just human actions and originally based on the Heinrich’s models outputs of almost nine out of ten accidents happening because of human errors. A good accident investigator understands this and doesn’t immediately jump into conclusions of blaming people.

Another interesting theory the author elucidates is the Error Theory by James Reason. This theory explains three basic types of errors: lapses, slips, mistakes. Slips are attention related execution failures when a person for example focuses too much on something and is inattentive to a failure happening because of that. Lapse on the other hand is memory related execution failure. An example of lapse is when you enter a hotel, hang your jacket in the wardrobe but forget that it’s there. Slips and lapses are often called together as skill-based failures whereas mistakes are called rule-based failures. Hence, a mistake happens when you do what is planned or intended to but the result is not what you expected.

Violations are also rule-based failures, where a person deliberately disobeys the rules or plans without the intention of any harm though. Sabotage is otherwise the same except that in that case there is clear intention to do harm. A lot of terms and concepts already here but one more: latent errors are something that is present in the system but remains hidden until particular circumstances make it visible and important. Good accident investigation obviously aims to find these too and then revise and correct the system.

The author also spends a good deal of the book to Introduce transactional analysis, which we will not cover in this summary. To people interested in parent, adult and child ego states of the theory, a good place to start is Wikipedia and of course the book itself.

Before digging deeper into what are the benefits and costs of accidents investigation, it’s good to remember that accidents are (almost) always unexpected and unintended and take place at the end of a sequence of events. It is the job of the investigative team to uncover the reasons and causes of those events.

Practical guide for setting up an incident reporting process

Why does it then make sense to investigate accidents? Some reasons that Dr. Thornhill lists are as follows:

  • Reduce likelihood of future accidents or incidents
  • Make improvements management system
  • Generate operating efficiency and in the end profits
  • Minimise dangers to people
  • Educate employees
  • Prevent staff attrition
  • Research and learn
  • Comply with laws and regulations

Accidents obviously create costs (both direct and indirect) for the company and to the individual. Some direct costs include

  • Downtime in production and lost revenue
  • Service disruptions and lost revenue
  • Absenteeism and medical expenses of the injured
  • Repair or replacement costs of damaged machinery and equipment
  • Early retirements of the injured or coworkers
  • Grown insurance premiums
  • Fines from regulators

Indirect costs are more difficult to track and make transparent, but they do nonetheless exists and consist for example:

  • Bad morale and increased attrition and in the end hiring costs
  • Losses and write-downs of brand image and goodwill
  • Lost and missed future contracts
  • Delays in delivery
  • Litigation expenses
  • Individual suffering of the injured and her/his coworkers

To make investigation as efficient and effective as possible, the author defines “Seven Arts of the accident investigation”. First one on the list, which may not be the first thing in mind, is to always prioritise which accident and incidents to investigate. For the most severe accidents such as fatalities and major environmental incidents will be investigated externally as well by regulators. These are the only ones you should investigate internally though as they happen so rarely (hopefully) and to prevent major accidents from happening, you should find ways to learn from the more minor ones. Besides prioritising the incidents to investigate, you should also decide how far to investigate. In some cases the latent errors and deficiencies in protection are easier to see but sometimes it takes an effort to reveal them. Just remember that resources are not infinite and not all errors are even detectable.

Third point on the list is about who should do the investigation in the first place. No investigator has time to investigate everything but if she can train and democratise investigations of more minor incidents to the field and foremen, more lessons can be revealed. Preparation of those who investigate helps a lot. Even doing one or few sample test investigations can instil some basic habits before the real deal happens.

Investigation requires access to information and that access requires authority from someone. Knowing who has this and being able to influence it is really important. To gather information also requires a good tool kit that should include at least: mobile phone with a good camera, notepad and voice recorder. The last of the seven arts is about writing the report. In the end the investigation is worthless unless useful results are recorded and put into action.

Remember that the more severe the accident or potential accident, the more detailed the investigation should be. All the aforementioned arts of accident investigation lead towards richer information and better end-results. Even though it’s almost impossible that two investigations are alike, there are common themes in many of them. Secure the area and inform those who need to know. Interview the witnesses objectively. Take photos and videos of the scene. Use notes and a voice recorder to mark your assumptions and thoughts down. Preserve evidence and take samples (of e.g. blood, hair, paint, soil, dirt). Put together a basic idea of the story of what you think happened. Additionally, CCTV and other electronic information may help.

The book also goes into detail about how to carry the interviews of the witnesses. They might often feel they will be blamed for providing information which should be clarified not to be the case as early as possible (as long as they haven’t violated the laws or policies). Questioning should be respectful and factual, not the “bad-cop-like” police interviewing. Mindset should be that both investigator and the interviewee are aiming to find out what truly happened. If possible and lawful, try to record the interviews. 

Good accident report should evolve at the same time as the investigation goes on. In the minimum the shareable part of it should include: introduction: what, when, where and “how much” (losses); team, evidence, what everyone agrees that had happened, what investigators think happened, conclusions and recommendations.

In the end, the point of the entire investigation is to reduce and prevent accidents and losses in the first place. Hence, the book digs deeper into how to put the recommendations into action. Safety Committee, which is something that many of our customers also have, is one way for that. The author also contributes to the Safety Committee idea by suggesting a separate Accident Investigation and Prevention Committee that can be a more focused group than a wider Safety Committee. Another way to go towards accident prevention is to increase the engagement in reporting of adverse events.

A final reminder from the author is that 2.3 Million people die every year because of work accidents and illnesses of which the author assumes saveable to be around 1.6 Million with better practices shown and explained in the book. It goes without saying that this is a disastrous number both in humanly and financially. What we have also noticed in today’s business environment is that companies with bad safety systems and records are not getting even invitations to participate in tenders.

This blog post was part of the safety book summaries blog series, that also includes the following books:

Is your organisation utilising safety metrics and KPIs to their full potential? You can check that amongst other things from our FREE guide:

New call-to-action

We are building the world's first operational involvement platform. Our mission is to make the process of finding, sharing, fixing and learning from issues and observations as easy as thinking about them and as rewarding as being remembered for them.‍

By doing this, we are making work more meaningful for all parties involved.

More information at

Related posts

16 Incidents every University and School should be reporting

University incident reporting, as well as school incident reporting, are critical for the...

Safety Management
8 min read

From Accidents to Zero by Andrew Sharman: Summary & Ideas

From Accidents to Zero by Andrew Sharman is a concise and easy to read from A to Z manual for any...

5 min read

6 ways to Engage Students in Incident Reporting

Student incident reporting can be an incredibly effective way of bringing about change in important...

Safety Management
4 min read

Involve your stakeholders to report

At Falcony, we create solutions that multiply the amount of observations and enable our customers to gain greater understanding of what’s going on in their organisations, areas of responsibility and processes.