Companies like to report one overall injury profile, which consolidates its operations. But several operating units may be performing very poorly, lurking within companywide figures.
What if the results among worksites are skewed so that a relative handful of poor performers cause the overall average injury rate to double? What if these relative few worksites demand virtually all the remedial attention?
The skewing phenomenon at work here is what has been called the "flaw of averages."
Anyone who has brought a boat into a harbor is very sensitive to this. The entrance may have an average depth of 15 feet at low tide. That average is a snare if there is a narrow sand bar four feet under water in the channel and your boat draws five feet. I learned that the hard way in a Connecticut harbor.
We should be asking not only what the consolidated injury figures say, but also what is the injury experience for the worst performing 20 percent of operating units. The lion's share of future improvement is going to be found in this small share of worksites.
Earlier this year, OSHA released 2007 injury data for 59,887 worksites. In the data for industries and individual companies, the skewing of injury risk is striking.
Twenty hotels in Las Vegas have an average lost-time injury rate of 0.9 per 100 workers. However, the worst performing 20 percent of these hotels have an average rate of 2.0 lost-time injuries per 100 workers.
United Airline's lost-time injury rates for 22 of its airports average 3.6 lost-time injuries per 100 workers. The average lost-time injury rate for its worst five locations is 7.5.
Get the picture? The performance of the worst fifth is double that of the entire population.
Then there are home improvement stores, over a thousand of which are in the OSHA report. In California, the stores other than Lowe's and Home Depot averaged a lost time injury rate of 2.3. Lowe's and Home Depot's rates were 0.78 and 1.6 respectively. No surprise: within each of the three groups, the average injury rate of the worst 20 percent of stores was at least double the average for all.
Among the 116 Home Depot stores reported in California, 16 of them had no lost time injuries at all. The total experience was so skewed that if all of the stores had the same number of employees, the worst 20 percent of the stores suffered just about the same number of lost-time injuries as did the other 80 percent.
This is yet one more confirmation that a corporate injury analysis is incomplete if it doesn't zero in on the worst performing fifth of its operating units. A worksite's relatively poor lost-time injury record can be due to a higher overall injury rate, or due to more of its injuries that do occur resulting in lost time. The OSHA report suggests the latter contributes more to a poor lost-time injury record.
Poorly performing operating units appear much less able or willing to keep injured workers at work, such as on modified duty. Curing that problem should be high on the remedial list.
A Home Depot spokesperson assured me that its injury recording system is very consistent across its outlets. The skewing in its injury rates is thus not a figment of varied injury recording practices.
Of course, it is useful to focus on the worst 20 percent within any population, as continuous improvement is a valuable mental discipline.
PETER ROUSMANIERE is an expert on the workers' compensation industry.
August 1, 2010
Copyright 2010© LRP Publications