Sunday, August 26, 2007

Diagnostic Errors
By Internists Often
Go Unrecognized


In one case that was included in the study of
diagnostic errors, an elderly man with endstage
chronic obstructive pulmonary disease
(COPD) was admitted to the emergency department
at 1:00 a.m. for midabdominal pain
and a hematocrit that had declined from a
baseline of 36% to 29%. The patient was on 60
mg prednisone for the COPD.
Although he was delirious from pain, the
man related a history of a bleeding peptic ulcer
some months earlier with similar pain.
Upon examination he was confused, with mild
abdominal tenderness. His stool was negative
for occult blood, and a nasogastric tube could
not be passed.
“The clinicians who saw him in the emergency
room weren’t quite sure what was going
on, but their No. 1 impression was that the patient
had a recurrence of his peptic ulcer,” Dr.
Graber said.
When the patient was seen 6 hours later by
the ICU attending physician, there was no evidence
of melena or hematemesis, but the man
was now in shock. The attending physician
considered the possibility that the patient had
dissection of an aortic aneurysm. A 6-cm dissecting
aneurysm was confirmed by CT scan,
but the patient died in the radiology suite.
“An investigation found that the team had
not read through the patient’s old chart—nor
was it available—and had not contacted the patient’s
primary care physician,” Dr. Graber
said. “Either one of these sources would have
quickly told the story. The patient had a
known abdominal aneurysm.”
The data collection by the medical team
“was grossly incomplete,” he said. “The synthesis
of information was faulty. They had the
wrong context. They had never considered the
possibility of aneurysm. They were thinking
GI causes.”
As for the system errors in this case, “clearly
the lack of medical records was a problem,” Dr.
Graber said. “So was the culture of the organization.
How could someone in the emergency
department feel comfortable taking care of this
patient without seeing his old records or without
talking to his primary care provider?”

No comments: