Today, we are going to bring you our multi-Part article series on what your fleet can do to predict and prevent severe accidents. And while you may initially think, “But that’s impossible,” in fact it can be done. Of course, every fleet wants to increase safety and mitigate crash risk, but they just don’t know how. By properly using data, sleep education and effective truck driver coaching, fleets can improve their overall level of safety and prevent horrible accidents.
Of course, the most significant aspect of a severe accident is the human toll. People can become injured or even die. Lives can be ruined. Beyond how it impacts people and their families, accidents can be substantial and leave little room for unbudgeted costs, whether they be from insurance claims, litigation, repairs, or service level damages. Consider that a single severe collision could cost your fleet millions of dollars, and you can see the problem. Numbers like that could cripple a small fleet.
One of the major problems in dealing with severe accidents lies in the fact that they are typically infrequent and largely happen at random… or do they? Could it be that, contrary to popular belief, large truck accidents are not random at all? Indeed, they may very well be a natural culmination of information, a set of subtle data points that can be isolated and analyzed. With the right information, trucking companies may very well be able to detect or prevent an accident before it ever occurs.
The key is understanding what issues the data points to. One of the most common causes of road accidents involving large trucks is that of fatigue. And while most conventional safety programs deal with specific truck diver behaviors, such as not checking mirrors or proper speed control, something like loss of control is generally a physiological problem. When a truck driver is suffering from fatigue or sleep abnormalities, they may feel awake even when their mind is asleep.
Consider this scary fact: A truck driver technically could be 100% in compliant with Hours of Service (HOS) regulations while still being asleep at the wheel. When truck drivers are tired and become distracted from operating the vehicle, accidents occur. Most severe accidents occur when truck drivers lose control of the vehicle and are not responsive at the point of contact.
When a truck driver has been exposed to:
- Disrupted sleep
- Truncated sleep
- Sleeping during the day
- Cumulative fatigue
- New sleep patterns and times
They may be at risk for a severe accident. Additionally, there are six major accident types that fall into the “severe” category and can be attributed to fatigue and loss of control:
- Roll-overs
- Run-off-road
- Head-on
- Jack-knife
- Side-swipe
- Rear-end
Each of these accident types could be potentially fatal for anyone else on the road as well as the truck operator. These types of loss of control accidents happen when the operator is disconnected or distracted from the truck driving task at hand. In these situations, they may not take any action, but had they been awake, would have seen the point of impact at least five to seven seconds before the accident occurs.
This is where the data gleaned from electronic logs can be used to the benefit of the truck driver and the fleet. Although the ELD rollout has not been without its fair share of confusion and complaint, it does provide a rich data set that can be used to do more than ensure compliance, it can save lives.
Data Deficiencies
There are essentially two different safety technologies that utilize safety audits to help fleets keep their safety margin high. The first type has been around for some time, and it essentially utilizes audits, perception surveys, and scorecards that judge a truck driver’s behavior. These factors help give fleet managers some insight as to where the higher levels of risk come from.
All told, most fleet managers conduct ongoing reporting and analysis, whether it be looking at data to determined what happened and why, after it already happened, how they can use information to prevent negative events from occurring in the first place. As a result, scorecards and dashboards are becoming more and more popular as a way for fleet managers to take information and develop correlations to predict what might happen. The problem with that approach is that correlation is not causation.
Scorecards are valuable, but they are best for telling you how well you will execute on strategy. These methods rely on big data but, fleet managers can only learn something from the data if they are using the right strategy in the first place. One of the most common perceptions about utilizing big data is that if you look hard enough and have the proper analytical tools in place, you can find the answer. But what if you are looking in the wrong place?
Scorecards and dashboards merely measure the symptoms and outcomes of a problem, whereas fleet managers should really be concerned with finding the root cause of accidents. Instead of fixating on specific events, such as hard breaking or unsafe lane changes, fleets must instead seek to figure out the root cause of a problem. In many cases, this could be something as simple as a lack of sleep. Data mining cannot address physiological problems.
The true solution lies in method two, which draws upon predictive analytics to gather a large sample of data and find patterns in it. Once a pattern is isolated, active truck drivers can be measured against the predictive patterns. Here is how that is done.
Predictive Analytics
It is possible to identify risk using predictive analytics. The entire point of predictive analytics is to identify truck drivers who may be at risk for having a severe accident. This way the fleet manager can intervene. An effective predictive model will use data such as prior employment, number of trips, loaded miles, and empty miles to determine an appropriate level of risk, per truck driver.
Custom data sets can be built using the fleets own systems to make these predictions. Many times, automated fleet management systems can arrange the data with very little input from fleet management. This frees up a lot of time and effort. A comprehensive system will rely on objective HOS data to measure true levels of performance. These models don’t rely on a truck driver’s perception of their condition and alertness. Humans are fallible, and as a result are prone to making the wrong decision when it comes to judging their effectiveness at a particular task.
When a predictive analytics system pulls data from a dedicated application monitoring HOS data, custom-tailored safety messages can be sent to a dashboard monitored by dispatch. An intervention can be staged immediately, with timely remediation of high-risk behaviors happening before a preventable accident occurs.
Many wonder how this data can be used, and if it represents an unfair way to target certain truck drivers. To be clear, predictive analytics is not a “gotcha” solution. Motor carriers use the model to put high-risk truck drivers on a program of remediation. Rather than taking punitive action, they instead take predictive action. The ability to predict and prevent severe accidents can save fleets hundreds of thousands, or even millions, of dollars, not to mention severe death and injury prevented.
Why Fatigue Prevention is So Important
The fact is, every fleet should have policies, training, and procedures in place to prevent accidents. Whether it be through continuous sleep education, truck driver coaching, or other methods, acting is key to ensuring safe operation while your truck drivers are out on the road. With fatigue and sleep deprivation being the number one cause of accidents involving large commercial motor vehicles, your fleet should be doing everything possible to mitigate the problem.
If your predictive analytics system has isolated truck drivers that may be at risk of a severe accident due to sleep deprivation or fatigue, you need to know what interventions work and what don’t. Consider sleep class as just one example. Truck drivers who don’t attend sleep class:
- Incur an average accident cost that is 7.2 times higher than those who attend a sleep class.
- Have twice as many loss of control accidents.
- Experience five times as many “run-off-road” accidents.
Now compare that to truck drivers who do attend sleep class, where they are:
- 30% less likely to voluntarily terminate their employment.
- 6.75 times less likely to have a service failure.
Truck drivers understand how important it is for breaks when it comes to their productivity and safety. Well-rested truck drivers drive better and suffer far less accidents. A sleep education program that includes sleep science education for front-line fleet managers, truck drivers, and even family members, goes a long way to training operators on how to be better rested and improve their sleep quality and quantity. Predictive analytics can help fleets achieve these measures. The most important thing is to take action now, before a severe accident occurs.