McGarry had been wrong, and not for the first time. In fact, a review of medical records, court documents and legal transcripts shows McGarry has made errors and oversights in autopsy after autopsy.
In three instances since 2005, his findings in cases in which people died in the custody of police officers have been challenged by doctors brought in to perform second autopsies. In each case, McGarry's findings cleared officers of wrongdoing. The other specialists concluded the deaths were homicides.
Contacted by phone, mail and in person, McGarry repeatedly declined to comment for this article or related radio and television stories.
Some in the field champion McGarry, praising his track record. "I have the utmost respect for Dr. McGarry and he taught me much when I was in a forensic fellowship program," said Dr. James Traylor in an e-mail. Traylor was trained by McGarry and worked alongside him in the New Orleans morgue. "I am unaware of any ‘mistakes' that he may have made."
Second autopsies are a rarity in most jurisdictions, but New Orleans civil rights attorney Mary Howell said she often taps forensic pathologists to perform follow-ups when she knows McGarry has handled a case. The degree to which their findings have differed from McGarry's is "shocking," Howell said. In some cases, they discovered, McGarry's work was so incomplete that bodies were "half-autopsied."
Gerald Arthur, a 45-year-old construction worker with a history of drug arrests, died after a struggle with police on a New Orleans street in 2006. Based on McGarry's findings, coroner Minyard ruled the death an accident, but a forensic pathologist with the Georgia Bureau of Investigation brought in by Arthur's family to perform a second autopsy found four broken ribs that McGarry had not noted. In his report, the GBI pathologist also stated that McGarry "failed to dissect" key neck muscles, causing him to miss hemorrhages that, in his view, suggested Arthur had been strangled.
In a deposition, McGarry disputed that assertion, saying he had dissected the neck muscles but had come to a different conclusion. "I don't have any evidence that this man had a death due to neck strangulation," he said.
No criminal charges have been brought in connection to Arthur's death. His family settled a lawsuit against the police department last year for $50,000.
Also in 2006, McGarry autopsied Lee Demond Smith, a 21-year-old man who died in jail in Gulfport, Miss. McGarry decided that Smith had been killed by a pulmonary embolism, a blood clot in the lungs, based on evidence of internal bleeding. Again, another specialist brought in to do a second autopsy found injuries that McGarry had not: abrasions on Smith's forehead and chest, as well as a half-dozen bruises on his legs and hands. The doctor concluded that Smith, like Arthur, had been strangled.
No criminal charges have been filed in Smith's death either.
Raymond Robair, a 48-year-old handyman who died shortly after an encounter with police, was autopsied by McGarry in 2005. Based on McGarry's examination, coroner Minyard declared Robair's death an accident.
But McGarry had not noted the wounds covering Robair's legs and arms. A second forensic pathologist hired by Robair's family documented 23 separate bruises, including a thigh contusion more than a foot long. The fatal injury was a severe laceration of Robair's spleen that caused extensive internal bleeding, according to the second autopsy, which was performed by another GBI doctor, Kris Sperry. Robair "was the victim of a beating," his autopsy report states.