In Depth March 11, 2010, 5:00PM EST

Caution: Stats May Be Slippery

Why claims of an almost miraculous decrease in U.S. worker injuries don't stand up to scrutiny

By Ben Elgin

Browse Issues

Browse the BusinessWeek Archive

Story Tools

http://images.businessweek.com/mz/10/12/370/1012_mz_57injury.jpg

Adkins battled AK Steel in court over a shoulder injury Billy Delfs

Peril permeates a steel mill. Molten metal explodes when exposed to water. Overhead cranes hoist steel coils weighing up to 80,000 pounds. Heavy vehicles roll within feet of high-voltage wires. "We have a work environment that presents every hazard known to man," says Alan H. McCoy, vice-president and spokesman at AK Steel in West Chester, Ohio.

In the face of these challenges, AK Steel (AKS) last year achieved what appears to be the nearly impossible. The company, which had sales of $4 billion, recorded only 15 injuries among 6,500 employees—its safest year ever. Between 1994 and 2009, it slashed annual injuries reported to the federal government by 96%. According to data compiled by the U.S. Labor Dept. it is more dangerous to serve coffee at Starbucks or count currency as a bank teller than it is to maneuver massive metal slabs at AK Steel.

Across industrial America, employers are making extraordinary assertions about safer conditions for the nation's 14 million factory workers. The overall rate of injuries has fallen 48% in a decade. "Employers are doing more to improve workplace safety," says Marc Freedman, executive director of labor law policy at the U.S. Chamber of Commerce in Washington.

Now, however, the Occupational Safety & Health Administration, which until recently also praised the claimed improvements, admits to having second thoughts. "The statistics we have on workplace injuries are incomplete and, in some cases, inaccurate," says David Michaels, the newly appointed Assistant Secretary of Labor who heads OSHA. Suspecting that safety achievements may have been exaggerated at some industrial companies, his agency is stepping up investigations to identify bogus injury numbers.

Michaels' suspicions were fueled by a report critical of OSHA issued in November by the Government Accountability Office, a nonpartisan congressional research agency. Based on interviews with physicians, occupational health experts, and others, the GAO found that some employers underreport injuries to reduce insurance premiums. Without identifying particular companies, the GAO also found that some workers fail to report broken arms and gashed legs because they fear being fired—a concern intensified by difficult job market conditions. As of last year, "OSHA's efforts to verify the accuracy of the data [were] not adequate," the GAO concluded.

Susan Baker, a professor at the Johns Hopkins Bloomberg School of Public Health and a scholar of workplace injuries, agrees that safety improvements appear to have been overblown. "It is extremely unlikely that injury rates would plummet like this," she says.

About 1.5 million U.S. companies with 10 or more employees must report work-related injuries and illnesses to the government if they require more than minimal first aid. Without accurate reporting, OSHA can't readily identify potential hazards, possibly leading to broader risks for more workers.

That's what some union leaders allege is going on at Caterpillar (CAT), the $32 billion maker of construction equipment based in Peoria, Ill. On paper, Caterpillar has trimmed its rate of workplace injuries by 81% since 2003. "We believe all accidents and injuries are preventable, therefore our goal for these is zero," the company says on its Web site.

The United Auto Workers, which has an acrimonious history with Caterpillar, counters that a lot of the improvement stems from the company's tendency to punish workers who say they were hurt on the job. "The extent of Caterpillar's blame-the-worker safety program is frightening," the UAW Local 974 newsletter asserted in its December 2008 issue. The local represents workers at Caterpillar's Peoria plant. "This unacceptable practice is carried out for one reason only—to intimidate workers and 'improve' safety statistics," the union newsletter added.