History of Health and Safety

Reading booksDo we really understand where we are in health and safety?

There is a not so famous Churchillian quote that “A nation that does not understand its history has no future”. Winston Churchill was first and foremost a consummate historian. He used his knowledge of the trend of events in the past to predict future events. He wasn’t perfect at it and he got some things wrong (Gallipoli, the independence of India and the UK’s National Health Service) but he correctly foretold the expansionist behaviours of Nazi Germany and Japan that led to World War II.  As a keen student of history I believe that safety should be examining its history for trends and at least trying to use that study to identify the challenges it may have to face in the future.

But few in safety stand back from their involvement in the here and now to recognise that things are changing. If they did they might be alarmed. They would see that the original theories of Heinrich and Bird do not match modern standards of research. They would see new theories being formed and then challenged by even newer theories: Heinrich and Bird have been overtaken by Rasmussen and Reason and they in turn are being challenged by Dekker and Hummerdal et alia. They would see a growing disillusionment by businesses big and small about the bureaucracy of safety. They would see that world-wide the regulators and service providers have been unable demonstrate just how much their efforts have reduced injury and fatality rates. They would see regulators around the world having their budgets cut. They would see a growing public and media disquiet about health and
safety. They would see a small number of adventurous thinkers exploring different ways of thinking about safety and trying to understand the masses of data that was previously not available.

Like Winston Churchill the health and safety communities may not be able to predict the future exactly. But at least Churchill was aware that the world of the 1920’s and 1930’s was undergoing a cataclysmic change. The world of health and safety is undergoing cataclysmic changes but most of those involved in the professions of safety are blissfully unaware of it.

It is time for H&S to wake up otherwise it will either have no future or the future that does happen will not be to its liking.



The image and text below relate to the comments after the post.

WA Mining Preliminary quantitative data from the WA government resources database shows a strong correlation between increased mining output and decreasing fatalities (the coefficient of correlation is 0.8 – anything above 0.9 is considered a perfect correlation) This is similar to in house studies done on construction and minerals extraction in the UK. It suggests that the reduction in fatality rates is business related. This conclusion requires further qualitative research to identify the drivers for increasing output. As a general principle regulators rarely become involved in driving commercial decisions.


  1. Andrew Townsend Reply

    Matt – The statistical naivety of the H&S profession astounds me. That graph is the same shape for every industry in the world – from shop keeeping to road transport to construction. The results correlate with economic indicators and educational achievement. They do not correlate with increasing regulation.

    If you want the definitive research report on the subject look up RR-913 on the UK’s HSE website plus look up cohort studies and control studies in any textbook on research!


  2. Matt Steele Reply

    Hi Andrew, I think the point of the article I posted (with the link) might have been misinterpreted, it wasn’t meant to draw attention to the graph but the content.
    You assert that regulators worldwide (in our case the WA and Qld Mine inspectorate)
    1) Are “unable demonstrate just how much their efforts have reduced injury and fatality rates”
    2) Are “having their budgets cut”
    3) And that we are seeing “growing public and media disquiet about health and safety”
    In our region this is defiantly not the case. The WA mining industry which employs 98,000+ people have (last year) had their first fatality free year since records began in 1896 (the graph was just showing this visually). This extremely important milestone is being celebrated by the public, and can be attributed largely to the terrific work our regulators are doing in engaging with all stakeholders (including unions and mining companies) to make our industry safer and more productive, (they can very clearly demonstrate that their efforts are reducing fatality rates). In the last 3 years alone the WA mines inspectorate have employed an additional 30 inspectors (Qld at least 11) and ramped up their site visits significantly (Their budgets are increasing). These men and women provide us with valuable on site assistance and we in the industry are proud to work with them.
    Are they perfect? No, but they are evolving to embrace new methods and research (funding a substantial amount of reasearch themselves) to ensure that last year’s achievements can be further improved upon through the more difficult job of maintaining a fatality free industry.
    I apologise if you took offence, I merely wish to point out that your assertions that our regulators do not contribute are under resourced and trapped by old methods are (in my opinion), incorrect.

    1. Andrew Townsend Reply

      Matt – I am probably the one to apologise if I gave the impression of taking offence. I have no doubt that the tremendous effort of WA mining and petroleum sector has had an effect but how big is that effect?

      I note that fatality rates were already declining steadily before the present interventions.

      Before any initiative can quantify its effect it has to establish a base line and eliminate the ‘confounding variables’ such as education, general improvements in health etc But above all you need to eliminate capital investment in modern extraction equipment and multiskilling. That takes a 5-10 years worth of data to be statistically significant. 2-3 years’ worth of data is too short a period to make any claims.

      Please would you do me a small favour by plotting the following data on an XY scatter chart…

      For the last 15-20 years on the X axis plot annual tonnes mined per employee against fatility rates on the Y axis – or send me the figures and I will do it for you. You will almost certainly find that as productivity goes up injury and fatility rates go down. Typical figures are that a 4% percent inprovement in productivity is associated with a 50% reduction in injury/fatality rates. You will probably get an ‘rsqared’ value of 0.6-0.7. In the studies I have done in high hazard industries in the UK the main cause (60%-70%) of reduced injury rates was investment in new equipment and training for multiskilling.

      I am fully prepared to be proved wrong in this case.

      However you will need another few years of data and be able to eliminate the alternative (co variables) explanations first.

      I remain astounded not offended that our H&S lords and masters have received so little training in research statistics. The UK’s HSE are similar – they have dropped some real stats clangers over the past decade. I should not really advertise but please read ‘Safety Can’t Be Measured’ when it is published in August


  3. Andrew Townsend Reply

    Matt – On second thoughts I should not condemn what I have not studied in depth. I would like to know more about the ways that H&S has improved productivity. If it results in a significant improvement in work output per man the WA operation should be studied by the rest of the world (the rest of the world usually uses avoidance of unplanned costs). Keep the dialogue going


  4. Alistair Camm Reply

    Interesting thread – I note however a contrary experience in Queensland where an increase in the number of regulators has in fact coincided with one of the worst safety performances on record for the mining industry.

    The research that Andrew noted (RR 913) also indicated that there was little correlation between regulatory interventions and improvement in safety outcomes. It reminds of the correlation between increased global temperatures and decrease in the number of pirates – see link below



  5. Hans Houtman Reply

    The “bureaucracy of safety” has led to a growth of regulations to a level where most people that work as front end operators are no longer aware of the totality of the regulations. In combination with changes in society, leaving the ideas of the risk culture behind us and embracing the precautionary culture, this is a dangerous situation. The precautionary culture is not only about taking precautions for the risk in an industry, but also about punishment for those who act against the regulations (this is more or less a shortcut, but it is the bottomline as well).

  6. Andrew Townsend Reply

    Daniel – I have just found some figures for WA annual output since 1990 and plotted them versus fataiities. I’ll send them as a jpeg file. Please add it to the post. rsquared value = 0.57 suggests that approx 57% of reduction in fatality rates is to do with improvements in core business and approx 43% is other. It needs a lot more work and data to track down what the ‘other’ is. So the jury is out on the safety contribution. Zero fatalities is a major achievement and the industry should be quite rightly proud. But most of it is not to do with safety. The HSE are a tad upset by us prodding their statistical assumptons of success; so the WA government and SWA might as well be the same.


  7. Suzanne Reply

    Having worked as a Safety Inspector in Canada, as well as a consultant (both sides of the fence) I would like to tell my story. I have seen firsthand how statistics are manipulated to please boards of governors and other stakeholders. The number of regulations and number of inspectors is not as important as what inspecting officers are focussing on. For example, when our province enacted environmental tobacco smoke H&S regulations, every single officer was tasked with enforcing this, and within a year there was widespread compliance, albeit a few penalties imposed. This regulation has stuck and now it is commonplace to have designated smoking areas outside of the workplace.

    Compare this with another set of regulations recently introduced that required employers to assess and control risk of musculoskeletal disorders (representing 50% of worker claims), and yet, it had no “teeth”, not a lot of enforcement (mostly education and consultation) and thus the MSD claim statistics remain relatively unchanged and compliance is lacking. So without knowing the details of the mine inspectorate – what questions are they asking, what actions are they requesting, the entire statistic is only a politically motivated number that is driven by other factors.

    Resources for a regulator are politically dependent – a change in government often signals a reduction in resources/staff, but my experience with two regulating authorities is that this is a wave that goes down but then up again. I would say Canada has seen regulatory downsizing but in some provinces it has grown again, and importantly the growth seems to in my opinion more on the “claims” side and not as much growth on the “prevention” side.

    In my consulting/private experience, a visit from a regulator gets attention at that site. Whatever the officer emphasizes becomes a golden rule at that workplace. If the officer has paid attention to the “right” things, the workplace will likely be “safer”. In private practice, I use the threat of a regulatory visit to get things I think are important prioritized at my workplace, and that makes me feel like I am making it “safer”.

    But here is another interesting occurrence. In British Columbia, our forest products industry has shrunk over the years but has been around a long time. Not too long ago, two different sawmills suddenly exploded and workers were killed. Wood dust. Seems obvious now in hindsight, but for decades previously there was not one explosion because we also have a lot of rain hence high moisture content in the wood. Over time the wood being processed has become drier and so the hazard crept unexpectedly into sawmills. The regulator missed the ball, so to speak because there is not ready evidence that inspecting officers were asking sawmills to control that risk. Another regulator – the Fire inspector – may have been delivering the message but the two organizations were not in good communication.

    My point is, it is not the number of regulations nor the number inspecting officers – it is about emphasis on certain hazards and luck. And even then an unexpected hazard that affects an entire industry can be missed. So back to the beginning, I think safety statistics are made-up artefacts to serve stakeholders, and not a reflection of safety and definitely not a predictor of future safety.

  8. Andrew Townsend Reply

    Suzanne – Thank you for that post.

    I agree with you about the cynical use of statistics and that they cannot be used to predict the future. The best that they can do is give clues about how to explain what went on in the past; and even that has to be cross checked by other qualitative means.

    Maybe I have been exposed to an unrepresentative sample of H&S professionals and politicians but the quoting of and dependency on single items of research again is alarming. While Australia was asleep I have been going through the Western Australia and Safe Work Australia published online research reports. They have been measuring perceptions not outcomes; there is no way of being able to link these perceptions to injury rates. Some of the survey questions SWA use are amateurish: they are phrased in such a way that ‘response bias’ is inevitable – the results are meaningless. The WA ones are better.

    The UK’s HSE construction people are similar: they grimly hang on to the results of a ‘Delphi’ study without realising that a Delphi study is a posh way of corralling expert opinion; it still does not relate to outcomes.

    I realise that, by criticising the work of others, I am holding myself hostage to fortune. The research program we will be carrying out on construction safety is bound to attract much criticism. Mercifully I have some really tough research tutors and the peer review process to provide the quality control on the work.

    There is probably the need for short sharp 800 word posts on quality assurance in research and its relevence to H&S.

    Please excuse the reply grasshoppering around various subjects. It has been a long frustrating day trying to follow the thread of the various research reports.


  9. Hans Houtman Reply

    Why do we use statistics? Isn’t it true that they only are about the numbers (and whether they go up or down) but not about the environment where people made errors and mistakes? The errors and mistakes are no longer seen in statistics: check the images above.
    Let me give an example about the use of statistics. On average the river was 50 cm deep. So on average it was possible to cross the river. The guy that stepped in to cross the river was never seen back alive….he drowned. The river was much deeper than the average 50 cm.

    So maybe it is better to concentrate on the real issues: human and environment and procedures.

  10. Andrew Townsend Reply

    Hans – I think it is a hangover from Newton, Babbage and Taylor. Everything can be calculated and management is a science.

    As someone who uses stats for the clues that they give about the human condition, I am painfully aware of the process that goes…

    Recording of the observations
    Turning the observations into numbers
    Averaging those numbers
    Turning the averages into a theory
    Which then becomes a formula

    My experience is that by the time it gets to the theory and formula stage the so called professionals have lost sight of the fact that the original observations were about people with their almost infinite variety of colour, character and strengths. Qualitative researchers talk about ‘contextual richness’; this almost entirely gets lost by turning things into numbers. It really gets lost when those numbers are then turned into averages.

    It appears that those numbers are then used to do calculations to which people are expected to conform. I believe we should try understand social phenomena qualitatively and only use stats as rather crude illustrations of the effects these phenomena.

    What really gets me is the assumption that people will respond with machinelike repeatability. People are people – they are not machines. They are capable of great feats of adaptation, innovation and compassion that no machine will ever do.

    By way of illustrating the illustration aspects of stats, I know from experience that if you treat your workforce with value and humour, productivity goes up and injury rates go down. But you can’t measure respect and humour so you can’t put a formula on it.

    Imagine the formula if you did

    productivity = k X ln( a1 x ‘respectals’ per square metre + a2 x ‘giggles’ per 200,000 hrs).

    They way society misuses its numeracy is ridiculous !


  11. Hans Houtman Reply

    Andrew, humour is so important, we need more of it in this society. The world is reduced to numbers and more numbers. It amazed me in the blog of the Harlem Shuffle that these guys were fired: it is having fun during work and these guys would hot have done the shuffle if it wasn’t safe to do.
    So for the rest of our life: “A day without laughing is a day without living.” It is my motto in life and everybody is free to use it. In the end: less statistics and less managers (these two keep themselves alive and are hardly of use in our society).
    Smile, tomorrow might be worse.

Leave a Reply

Your email address will not be published. Required fields are marked *