Lies, Damn Lies, and the BLS
Sep 28, 2017

This Journal has seldom met a statistic that it didn't cherish, but a recent article published on calls our devotion to data into question. The article looks at the 2015 compilation of the Survey of Occupational Injuries and Illnesses (SOII) published by the Bureau of Labor Statistics (BLS). Our author, Jim Thompson of, takes a look at some of the discussion which has swirled around the state reporting used by the BLS for several years now.

The author then adds, '[B]ut those numbers are widely acknowledged, even by the BLS, to be inaccurate. The BLS notes on its own website that the bureau and its research partners 'have conducted multiple studies which indicate that SOII failed to capture some cases... '." Egad. The BLS compilations are the baseline that most of us use when making sense of other data sources for workers' compensation claims and related stats. The BLS and the SOII appear to have a consistent bias towards undercounting on the job accidents.

Both employers and employees have reasons not to report all claims. A few rogue employers, of course, sometimes try to reduce premiums by keeping minor claims off the books or foisting them onto the group health plan. Employees sometimes fear retaliation or firing if they report claims, so they find various ways to cover up occupational incidents.

All this is true enough, but the real question is whether the undercount phenomenon is significant enough to call the BLS and similar data sources into serious question as benchmarking standards. What's important to note in reading this article is that the same motivations for underreporting have been in place for many years. In other words, the long term trends noted by the BLS and others towards fewer and different types of injuries in the workplace are probably correct. Additionally, the more severe incidents on the job are the least likely to not be reported, thus, if anything, the "traditional" underreporting probably makes comp injuries appear slightly more severe on average than they really are.

In sum, it's good to keep in mind that all statistical analyses are limited by the accuracy and completeness of the underlying data and the BLS and other government numbers, or, by extension, the NCCI and WCRI numbers, are no exception. Decimal dust should never be taken too seriously in our line of work. The big trends that we all talk about and try to deal with - declining injury rates, the aging workforce, the impact of opioids, etc. - remain just as real as ever. The fact that we don't have perfect numbers never excuses us from making the best use we can of the ones we do have.

Driving from Theory to Fact 

A new study from the Insurance Institute for Highway Safety (IIHS), nicely titled "Stay Within the Lines", shows that the early promise of automated safety systems in new cars, especially the lane keeping function, is actually coming true on the highways of America.

A review of actual crash data from 25 states for events between 2009 and 2015, when these systems were first becoming available, showed that the number and severity of crashes were materially lower for cars with lane keeping assistance. Most notably, the fatality rate for system equipped cars was 86% lower. Now an 86% improvement gets your attention. Even more, the IIHS study estimated that if all vehicles on the road in 2015 had had these systems in place, some 85,000 accidents would have been avoided entirely. This was a small study, but the folks at the IIHS know their crash stats, so these outcomes should be considered seriously.

A parallel study which looked only at blind spot detection systems showed similar results. What these studies tell us is that, although the self-driving car isn't quite ready for the dealer's showroom today, some of the technologies involved are already proving effective, preventing crashes, and saving lives. 

While flying cars are still a long way off (darn), some future tech is here today and working. Are you, as a risk professional, taking full advantage of these advances for your G/L and A/L programs and even for parallel types of loss engineering in other areas? Look out the window. The future has just arrived and it's parked in the company lot.

Straws in the Wind: Busy Feet and Clean Teeth 

The August issue of the Journal of Occupational and Environmental Medicine (JOEM) includes a couple of small scale studies that remind us that little things often point the way to larger issues. One study looked at the correlation of what the authors call "active commuting" (that is, walking or cycling to work, as opposed to sitting in a steel and glass box, alone or with several hundred strangers) with above average productivity at work. The study focused on HPQ (an index to worker productivity developed by the World Health Organization) as an outcome, but it is a fair assumption that high HPQ/ highly productive employees are less likely to be involved in on the job accidents or to drive other risk costs.

The study looked at a group of male office workers in Kobe, Japan, but these folks were statistically similar to white collar workers in any OECD country. The important point is that those employees who had an active commute of 20 minutes or more showed materially better productivity scores than those who had shorter commutes or more conventional commuting routines. The researchers associated the physical stimulation of 20+ minutes of walking or cycling each way with heightened alertness and superior job performance.

The second study, by a separate research group, looked at the correlation between productivity and dental health. Yep - dental health. This group used dental benefit records to look at employees' dental status and then compared this to various measures of overall health and on the job productivity. To ruthlessly summarize a complex set of findings, those employees who kept their teeth in the best shape were the healthiest in other measures and, more important from a risk management point of view, the most productive.

Is there a takeaway here? Well, it appears that despite our constant search for the one Silver Bullet that solves all our risk problems, it's actually the small stuff that counts in the real world. More to the point, it's the people who make the difference. The employees who take the daily grind of getting to work as an occasion for a brisk walk or a bike ride, the people who bother to take good care of their teeth and use their dental benefits intelligently are the secret to superior work productivity and almost certainly better risk performance.

We love our systems, our analytics and our best practices as we all try to be the next Risk Manager of the Year at RIMS, but first and foremost we have to attract, hire and retain the right people - the people who brush their teeth vigorously before they hop on their bicycles to pedal off to work. Who knew? Well, now you do.

The Traffic in Hurricane Alley

The Economist, that last bastion of hard data in a world of loose talk, took a look a couple of weeks ago at the question: is the US having more hurricanes? Much idle chatter has collected around this question recently, but a session with NOAA numbers going back to 1850 (subscription required) tells us that, indeed, the frequency of hurricanes is actually going down. There is a tiny hint that average severity may be rising, but this is mere decimal dust thus far. As we noted in the first item, numbers have to be used carefully and often skeptically, but they are still the best path to managing risk effectively. Sound opinions are best built on solid numbers.


Share This
* Required Fields