You Don’t Need a Crystal Ball. Stop Accidents Before They Happen
Using predictive modeling and leading indicators to improve your risk management outcomes
Why is it that we are so willing to accept weather forecasts, polls for election outcomes, and other predictors we see almost every day, but struggle when similar predictions are suggested to control losses?
The first thing to understand is that no predictions are 100 percent accurate. They are merely indicators of results that are expected to occur when predetermined indicators, or factors, exist. For example, using certain statistics from the Bureau of Labor as an indicator, we can predict that, for every 100 employees working in the construction industry, we should expect 5.4 of them to be injured during the year. Of course, we know that some companies will perform better than average while others will do worse. Is this a result of luck or chance? We know that the companies that enjoy better than predicted injury outcomes do so because they apply sound safety practices to keep their employees protected.
Those safety practices are indicators that we are trying to evaluate. If a company holds regular toolbox meetings, does the expected outcome of future losses change? Would it be easier to sell a new safety idea to management if you could quantify this impact?
Making Your Own Predictions
Most companies collect loss information and look for causes of loss one accident at a time. This approach is normal and will yield results typical of your industry. However, if you instead create a historic look at several loss variables and then overlay your loss history over the general industry data, it might yield some very helpful information for your business. Of course, some of these results may be intuitive. Notwithstanding, they may help you identify anomalies or frequent losses so you can then take the steps to take to avoid them in the future.
A Practical Example:
Fact: New employees get injured more often than seasoned employees.
Reliable national statistics tell us that a new employee is seven times more likely to get injured in the first year of employment.
When analyzing your loss history, you might discover that your injuries happen closer to their initial hiring date and could be indicating an improvement opportunity related to the hiring and training process. By identifying this possible issue, you can focus your attention on possible solutions to reduce workplace injuries.
In automobile accidents, most of us realize that many outside factors influence accidents, including weather, population, traffic, fatigue, vehicle type, etc. Internal data can be used to identify corollaries between vehicle accidents and particular factors. For example, factors such as trip distances, net payroll, percent of driver training, idle time, average monthly mileage, zip codes, total miles driven over the last six months, and others can be analyzed to determine if they are a contributor to losses.
Change your thinking from the old model to the new model:
Predictive modeling isn’t really a new concept, but it may require a change in the way we approach information to be truly useful. If you can steer away from the Old questions below and toward the New, you may find yourself with data rich with indicators to help you find creative ways to prevent losses.
OLD: What happened? To NEW: Why is this happening?
OLD: How many accidents happened? To NEW: What is the data saying?
OLD: What do you think the problem is? To NEW: What do you think will happen next?
OLD: What should we do about it? To NEW: How do we stop them from happening?
OLD: What is the worst that can happen? To NEW: What is the best we can make happen?
Don’t miss out on using these tips to improve safety for your business and your employees!
Acadia is pleased to share this material for the benefit of its customers. Please note, however, that nothing herein should be construed as either legal advice or the provision of professional consulting services. This material is for informational purposes only, and while reasonable care has been utilized in compiling this information, no warranty or representation is made as to accuracy or completeness. Recipients of this material must utilize their own individual professional judgment in implementing sound risk management practices and procedures.