John Beale/Post-Gazette
	A rash of deaths prompted Dr. Gordon Christensen to push for an inquiry into the activities of a nurse at Harry S. Truman Memorial Veterans Hospital in Columbia, Mo.

COLUMBIA, Mo. -- That September morning in 1992, acting chief of staff Dr. Edward Adelstein passed along what he thought would be a routine request to colleague Dr. Gordon Christensen.

Some nurses on Ward 4 East at Harry S. Truman Memorial Veterans Hospital near the University of Missouri campus had asked Adelstein to investigate the high number of recent emergencies and deaths there. The nurses said a certain nurse seemed to be on duty when nearly all the deaths occurred. Adelstein wanted Christensen, the hospital's epidemiologist and the associate chief of staff for research and development, to review the data.

"I really was convinced there wasn't anything there," Adelstein said.

In less than a week, Christensen and researcher Andy Simpson had completed a study that mapped every death on 4 East against the whereabouts of every nurse -- up to 60 in all -- during the previous year, using work records, nursing notes and patients' charts. Christensen assigned a code name for each nurse to avoid bias. He knew that proving anything would be difficult.

"You have to understand, none of these were people who died with a knife in their chest. They were just people who died." But the data nearly leaped off Christensen's computer screen. One particular nurse, dubbed Nurse H, was on duty when 45 of the 55 deaths occurred on 4 East between March 8 and Aug. 22.

On average, one patient under Nurse H's care died for every three of his overnight shifts. For three of Nurse H's shifts, more than one patient died, the only nurse who had that occur. A later reanalysis by the VA's Office of Healthcare Inspections confirmed the findings, and that reanalysis was validated by a Penn State biostatistician.

According to Christensen, the statistical probability of that happening by chance was less than 1 in 1,000,000,000,000,000,000,000.

"It doesn't get any more abnormal than that," he said. The two went to the hospital director, Joseph Kurzejeski, and urged him to call the police.

What happened next took them by surprise.

A few days later, Christensen was told he could not present his data to an internal investigating board. During the ensuing months, he was told not to contact law enforcement officers and was warned that his analysis was considered part of the hospital's "quality assurance," which had to be kept private.

When he later reported the nurse to the state licensing board, he was threatened with sanctions for violating hospital confidentiality.

After years of "outstanding" work performance evaluations, Christensen, 55, ultimately found himself before a national VA panel investigating him for poor management of his department. "They were trying to discredit me," Christensen said.

Lost research grants

Others saw their professional lives take a downward turn, too.

	John Beale/Post-Gazette
	Hospital researcher Andy Simpson, left, helped Dr. Gordon Christensen, right, correlate death rates to nurses at Harry S. Truman Memorial Veterans Hospital.

Researcher Simpson, 54, a 15-year VA veteran with a doctorate in microbiology, failed to get renewal of three separate grants in the next three years. In 1995, with no money to support his research, Simpson was let go.

In 1995, Christensen was forced to give up his position as chief of infectious diseases and was excluded from planning and management task forces. His appointment to the VA's Disciplinary Appeals Board was blocked at the last minute, without explanation.

"We're simply not players here," said Adelstein, 63, an assistant professor of pathology and chief of laboratories at Truman.

The chief of staff, Dr. Earl Dick, also believed homicides had occurred and, when he pushed the issue, "Mr. Kurzejeski's relationship with me rapidly deteriorated as he became increasingly sarcastic and demeaning," he said later.

The next year, Dick was told he'd been rated "unacceptable" in his job evaluation. Dick was relieved of his duties in 1994.

No one may ever know if the patients at Harry S. Truman Hospital were murdered.

Nearly a year after the deaths, 13 bodies were exhumed, but after so much time, the FBI could not determine with certainty that murders had occurred. Because the deaths were originally listed as being due to natural causes, no autopsies were done.

In 1998, the widow of one patient, Elzie Havrum, won a $450,000 judgment against the VA in federal court after the judge found, based largely on Christensen's analysis, that the hospital negligently failed to protect patients from the nurse, identified as Richard A. Williams. Two years later, the U.S. 8th Circuit Court of Appeals upheld the decision.

That ruling helped renew interest in the deaths. After more tests, officials found evidence of a paralyzing drug in the exhumed patients' bodies and Williams was arrested.

But two months ago, homicide charges against Williams, 37, were suddenly dropped after Boone County Prosecuting Attorney Kevin Crane said the tissue testing was flawed and could not be used to prosecute him. Crane added that the VA "continues to consider this case under investigation."

After spending a year in jail, Williams is now a free man.

Williams, who has denied any role in the deaths, declined to be interviewed, but his public defender, lawyer Don Catlett, said his client did not hold a current license and was not interested in going back to nursing.

Christensen remains convinced the deaths were the work of a serial murderer, but said "so many things about this have been screwed up, no one may be held accountable for these deaths."

Christensen can't help but wonder if the result would have been different if the VA system had looked into the deaths more rigorously, instead of reacting defensively.

Told to stay quiet

After he made an initial call to the VA inspector general's office, hospital officials told him to have no further contact with the inspector general or the FBI. Later, when Williams went to work for a local nursing home, Christensen was threatened with punishment for contacting the state nursing board.

Not long after that, a new hospital director asked a national panel to come in and review Christensen's performance and, in June 1997, it recommended Christensen's removal, citing "concerns about [his] leadership and management skills."

While the hospital never followed through on the recommendation to dismiss him, Christensen still felt persecuted. "I am convinced the VA intended not just to eliminate an inconvenient employee, but to destroy the credibility of my accusations by destroying my professional credibility," he said.

A strong sign of that, he felt, was the fact that his job evaluations gave him consistently high marks for competence and leadership until 1993, the year after he had reported information on the patient deaths. Even then, he was downgraded only on administration, not his clinical abilities. As he persisted in pushing for further investigation into the deaths, his job evaluations continued to decline, and in 1996, he received an overall "unsatisfactory" rating.

Kurzejeski, who retired in 1994, did not return phone messages left at his home in recent weeks. In a 1997 affidavit, he said he was not trying to interfere with the investigations but wanted Christensen "to essentially stop pursuing those efforts and to work at his officially assigned duties as a VA researcher."

Christensen believed that stopping a possible serial murderer took precedence over those duties, and he believes he paid a high price for following his beliefs.

Adelstein felt the hospital's wrath, too.

One month after testifying on behalf of Elzie Havrum's family in its civil suit against the VA, Adelstein was accused of improper removal of a drug used to euthanize pets, an incident that had happened two years earlier.

John Beale/Post-Gazette
Dr. Edward Adelstein, former acting chief of staff at Harry S. Truman Memorial Veterans Hospital, asked epidemiologist Dr. Gordon Christensen to conduct what was thought to be a routine investigation correlating death rates and nurses.

Adelstein, trained as a veterinarian, said he used a small amount of the drug to kill a neighbor's pet that had been suffering from cancer and seizures.

Adelstein was told to undergo verbal counseling and had a report filed in his permanent record.

What's still puzzling to Christensen and Adelstein is that they believed they were simply reporting data when they came forward with the troubling findings, and that their motivation to protect patients would be viewed favorably. "I remember saying, 'We're fine. We've just done our job,' " Adelstein said.

Even when hospital administrators began questioning their findings, "we just thought they didn't understand, [and] that when they did understand, of course they would do the right thing."

Instead, their careers were thwarted. "The stress of 11 years of this -- the wear and tear, the whole attack to your character -- it's just beyond imagination," Christensen said.

At the end of July, Christensen decided to retire from the VA after 21 years, "in large part because of this," he said. He'd stayed as long as he did "because I wanted to leave with dignity." For personal reasons, neither Christensen nor Adelstein left Columbia, where both have appointments at the University of Missouri's medical school and where Adelstein is deputy medical examiner for Boone County. Nor did they pursue legal redress, believing it would be too lengthy and costly.

But the ordeal apparently has not hurt Christensen's standing with his colleagues; he has been elected president of the University of Missouri's faculty senate.

Said Christensen with a smile: "People thought I would stand up and say some things."

The Cost of Courage: Day Three