Beltman Forrest Riley Sr., 54, was at the controls of a battery tractor at Consolidation Coal Co.'s Shoemaker Mine near Dallas, W.Va., two years ago when the tractor's left rear wheel ran over a pipe.

The wheel lifted one end of the 10-foot-long, 2.5-inch diameter pipe and sent it through a small opening in the operator's compartment, where it struck Mr. Riley on his left side, pinning him in the cab.

Co-workers had to bring in another tractor to move Mr. Riley's tractor before they could get him out. They administered first aid but could not save him. Mr. Riley, of Steubenville, Ohio, was pronounced dead at a hospital that Oct. 9 morning.

On Dec. 30, MSHA officials removed Mr. Riley from its list of people who died in mining accidents that year. After a medical examiner's report concluded he had suffered a heart attack, MSHA and West Virginia's Office of Miners' Health, Safety and Training both said his death "resulted from natural causes."

MSHA spokeswoman Amy Louviere said last week that "the medical examiner believed the heart attack probably precipitated the accident," although no mention of that appears in the report.

The report, however, does say that Mr. Riley's injuries "cannot be entirely excluded" as a contributory cause of his death.

Mr. Riley's is not the only death that doesn't appear in MSHA's annual count.

In the past four years, 153 deaths have occurred at U.S. coal mines, but another 72 deaths were labeled "nonchargeable" incidents, 55 of them attributed to heart attacks or other "natural causes." As of late October, another eight cases were "pending determination," according to an internal MSHA daily fatality report.

During the same period in mines that don't produce coal -- metal/nonmetal mines -- 154 deaths were listed as "chargeable" and 100 classified "nonchargeable" because the workers died of natural causes.

The distinction is important because, unlike a death from a roof fall, there is no public follow-up investigative report that spells out the circumstances or recommends steps that might prevent deaths.

And because public information on nonchargeable deaths is limited, cases such as Mr. Riley's have some wondering if MSHA is too narrowly defining what a chargeable death is.

"It's really a matter of what the agency's function is," said Dr. Anthony Robbins, of Tufts University, a former head of the National Institute for Occupational Safety and Health and former president of the American Public Health Association.

"If its function is prevention, which it should be, then to make this not chargeable misses the opportunity to focus on some aspects which could prevent deaths in the future."

Celeste Monforton, of George Washington University, who worked at MSHA for six years before leaving in 2001 and has assisted in West Virginia's investigations of mine accident deaths this year, said she was appalled that Mr. Riley's death would be attributed to natural causes.

"That sure sounds like it's mining-related to me. I'd have a heart attack, too, if I were pinned in a cab," said Ms. Monforton, who is an environmental and occupational health specialist.

Ms. Monforton's tenure at MSHA overlapped a change of leadership at the agency, from J. Davitt McAteer to David Lauriski, which, she said, coincided with a heightened emphasis on the yearly fatality count.

MSHA's new goals
In August 2001, Mr. Lauriski told a meeting of the Kentucky Mining Institute he had set an agency goal of reducing mining fatalities by 15 percent per year in the next four years, and cutting the nonfatal lost work days by 50 percent.

"This, of course, will require the commitment and help of all who work in the mining business," Mr. Lauriski told the mining industry crowd. "It can be achieved, and with your help, it will."

That year, 42 miners died, the highest yearly total since 1995, including 13 miners killed in an explosion at the Jim Walter Resources Mine No. 5 in Alabama one month after Mr. Lauriski's remarks.

MSHA documents show another 24 coal mine deaths were judged "nonchargeable" in 2001. All but four of those nonchargeable deaths were attributed to heart attacks, and one of the four was listed as "natural causes."

Since 2001, the number of chargeable coal mine deaths has held steady or gone down until this year, while nonchargeable deaths have increased from 14 in 2002 to 21 last year, nearly matching the number of charged deaths. Fifteen were attributed to "natural causes."

It is clear that both MSHA and the mining industry pay attention to the numbers.

On Jan. 5, 2005, MSHA sent out a news release heralding a "new record low" for fatalities in 2004. But before a news release could be prepared in January to mark another record low last year, 12 miners perished at West Virginia's Sago Mine.

In Sago's aftermath, the mining industry pointed to last year's death toll of 22 as evidence that mining still was safer than ever.

That record low in deaths has turned out to be a quirk of the calendar. Had Sago occurred two days before, the 34 deaths last year would have been in line with the 10-year average. And, not counting the Sago deaths, this year still would be on track to match or exceed that 33-34 average annual death toll since 1995.

Ms. Monforton believes the focus on the yearly fatality count and whether it has gone up or down has overtaken both good science and good sense.

"Certainly, people in the industry and in the agency want to see those [fatality] numbers go down, so there is an inherent desire not to count things," she said. "But because we're driven by some arbitrary number count, we're losing a whole subset of information that can be very helpful in terms of prevention."

Two other recent mine deaths reviewed by the Pittsburgh Post-Gazette raise further questions about the application and consistency of "nonchargeable" findings:

In July 2004, Mystic Energy security guard Brian Castle was killed on his way to work when his vehicle collided with a coal truck on mine property. MSHA did not include Mr. Castle in its yearly total for 2004 because "although the site of the accident was on a tract of land owned by Eastern Coal Corp., there were no mines operating on that property," Ms. Louviere said. "The victim was in his private vehicle, on his own time, and not at a mine."

In May 1995, Tory Davis, 5, and his father, Timothy Grace, who were inside a converted school bus serving as a lunchroom, were crushed under a collapsed highwall at a sand and gravel mine in Wenatchee, Wash. On its Web site, MSHA lists Tory and his father as the 17th and 18th deaths that year at metal/nonmetal operations.

Yet MSHA does not count the October 2003 death of Cindy Foglesong, who had accompanied her friend to a western Illinois mine where he worked. She was crushed to death when a limestone ceiling in the mine collapsed on her.

The death was not counted, Ms. Louviere said, because "the victim was not an employee of J.M. Huger Corp. Mine, was not performing mine work, and illegally trespassed onto mine property without approval from mine management."

'Byzantine definition'
The explanation does not convince Mr. McAteer, who headed MSHA during the Clinton years. Mr. McAteer once insisted the agency count as chargeable the early 1990's death of a small boy killed while riding in a car. The family was driving by a mining operation when the car was struck by flying rock from a planned explosion at the mine.

"The question is, did the operation of the mine cause the death? It doesn't have to be an employment nexus," Mr. McAteer said.

"To create this Byzantine definition of what counts and what doesn't count neither serves the miners nor the agency nor the government nor the companies because it creates a false database that doesn't reflect what's really going on."

Ellen Smith, owner and editor of Mine Safety and Health News, took MSHA to task in 2004 for not counting the deaths of two tree cutters killed on mine property. Also that year, she questioned the initial "nonchargeable" finding of a 54-year-old truck driver who died 16 days after his brakes went out on an access road at the Colony Bay Surface Mine in Wharton, W.Va., submerging him in a sediment pond for two to three minutes.

In each instance, she said, the "nonchargeable" determination went against precedents, sometimes going back three administrations. In all three instances, MSHA later reclassified them as work-related.

In fact, MSHA's reversal of a decision not to count a death as work-related -- called a chargeback -- have become much more common in recent years. Since 2000, seven coal deaths have been added to the annual total after further consideration. In the five years before that, it happened once.

Some believe all deaths should be counted, including heart attacks.

"If someone is working under conditions of extreme pressure and stress that has resulted in a heart attack, other workers are going to be faced with the same risks unless that information is recorded and dealt with," said Dr. Michael Silverstein, a professor of environmental and occupational health sciences at the University of Washington.

Dr. Silverstein sees an inherent conflict when enforcement agencies such as MSHA and the Occupational Safety and Health Administration decide which deaths count as work-related. "There are incentives with these government agencies to find ways to claim success, and that is to reduce work fatalities. Every administration does this."

According to MSHA's accident/illness investigation handbook, the district manager makes the initial decision on whether a death is reportable, and that decision is reviewed by the program administrator at MSHA headquarters in Arlington, Va.

The administrator might choose to send the case to the four-member Fatality Review Committee comprising the district manager, the chiefs of safety for coal and metal/nonmetal mines, and the chief of MSHA's Division of Mining Information Systems.

Only the head of MSHA, the assistant secretary of labor, can review the committee's decision.

None of these proceedings is open to the public, nor is there any public report if the committee decides the death is not related to the mining operation.

"There should be some hard, fast rules about what a 'chargeable' death is. It shouldn't be left to some secret society," said Ms. Monforton.