Predictive analytics provides powerful tools that enable the rapid sifting of data to discern patterns that indicate impeding incidents. Safety is not alone in this quest to mine and analyze data to predict the future. Advanced and predictive analytics have revolutionized many industries, from biotechnology and mapping of the genome to banking and market research, and is the foundation of Internet search engines such as Google. When applied correctly, predictive analytics allows leaders to gain deep insight into their business and deploy their scarce resources in an optimal way.

In this article, we will explore how safety professionals can use data to predict and prevent workplace injuries.

Predicting incidents

Safety inspections are the cornerstone of every effective safety and risk management program. Hundreds of thousands of safety inspections are conducted every week and provide a rich and large source of safety data. What does this large data set tell us? Can it be used to predict future incidents?

Research1 has proven that workplace injuries and safety incidents can indeed be predicted before they happen with accuracy levels between 80 and 97 percent.

Four safety truths

Further study of safety inspection data yielded four universal safety truths that help us understand what the safety inspection data is telling us.

Safety truth #1: More inspections predict a safer worksite

At first this relationship may not appear to be clear. In fact, the correlation between the volume of inspections and incidents is low, however a careful analysis shows that the only worksites that have a low incident count are the ones that have high inspection volume. Conversely, the only worksites with high incidents are the ones with low inspection volume. The message is unmistakable: the worksites with poor safety outcomes have low inspection volumes.

If your company is experiencing high injury and incident rates, the data indicates the first step is to simply get onsite with your safety checklist and do more inspections.

Safety truth #2:  More inspectors, specifically more inspectors outside the safety function, predict a safer worksite

Once an organization starts doing more inspections, the next step is to get more people, and specifically more people outside the safety function, involved. There is a link between incidents and the degree of diversity among people involved in performing inspections. It shows that the probability of having an incident decreases as the number and diversity of the people performing inspections increases.

Sites that have a high level of participation in the inspection process have a better safety record than sites with a few professional inspectors, even if the total number of inspections performed by the two groups is similar.

In other words, having a large number of diverse inspectors doing a few inspections each is better than a few inspectors doing a large number of inspections, even if they are highly trained safety professionals.

If you have increased your number of inspections (Safety Truth #1), but are not seeing improvements in injury prevention, get more people, and people outside of safety, involved in your inspection program.

Safety truth #3:  Too many “100% safe” inspections predict an unsafe worksite

While at first it may seem counterintuitive, a high number of inspections that show very few, or no, unsafe or at-risk conditions invariably came from some of the most unsafe worksites that were studied.

While one could interpret the inspections at their face value and assume that the site is safe given low levels of unsafe conditions, this is rarely the case. It turns out that even the safest worksites (e.g. EMR well below 1.0) often have inspections that record a moderate level of unsafe observations.

Intuitively it may seem that as worksites improve their safety performance, the number of unsafe conditions reported by safety inspections would fall, but what happens in practice is quite different. The proportion of unsafe conditions found remains fairly steady as organizations continue to improve their safety performance.

Generally, as the work environment changes, due to new processes, procedures, equipment, employees, etc., new unsafe observations are found that were not evident in the old environment.

Or, what was once considered an acceptable condition or behavior is now deemed unsafe based on new information. Inspectors continually become more critical and discerning of conditions and behaviors in the workplace.

If most inspections are returning 100-percent safe information, your organization may be “flying blind,” meaning the worksite is at a higher risk of having an incident, but the inspectors are either not seeing or not reporting the leading indicator signs of those incidents. Research shows that the safest worksites continually find a certain level of unsafe conditions and behaviors, and then fix them before they become actual incidents.

If you’re still having issues with your injury prevention program, make sure your program not only rewards high levels of inspections (Safety Truth #1), by many and non-safety team members (Safety Truth #2), but also trains for and rewards the reporting of unsafe observations from your safety inspections. The more unsafe observations you get, the more you can resolve before they become actual safety incidents.

Safety truth #4:  Too many unsafe observations predicts an unsafe worksite

To state the obvious, a persistently high level of unsafe conditions is associated with a high level of incidents. Analysis of the data showed that companies in this group have nearly the same level of risk as those that find virtually no unsafe conditions (the “flying blind” stage outlined in Safety Truth #3).

What often occurs is that a lot of inspections are done (properly adhering to Safety Truth #1) by a large and diverse inspection group (properly adhering to Safety Truth #2) and they find a high level of  unsafe conditions and behaviors (thus positively avoiding Safety Truth #3). However, the levels of unsafe observations keep increasing because they are not being resolved. This can be referred to as the “inaction” stage. In this stage, the inspection program is strong, but the resulting injury prevention activities are not.

In order to move away from  inaction and move back to the area of least risk, a worksite must commit to resolving its unsafe conditions and behaviors, which the research shows should drive down the level of future unsafe observations to an acceptable level.

When the truths are employed

The research found that worksites that successfully incorporated all four of the Safety Truths had two to three times less incidents. By promoting high levels of inspections, across both safety and non-safety functions, where it was expected that unsafe observations would be continually found and addressed, world-class worksites were able to manage their risk of injuries and stay in the area of lowest risk.