| Index | Last Week | Contact DQ News | Latest Issue |
| Featured Articles | Data & Information | Science & Technology | Education |
| Economics | Business | Law | Medicine |
Autopsy Data - How Useful?
A paper and an editorial in the October 14th issue of the Journal of the American Medical Association examine the number of autopsies performed throughout the United States each year (5% - 10% of U.S. deaths each year, depending on the hospital) and at the Medical Center of Louisana at New Orleans (24% of patients who died there between 1986 and 1995).
According to the JAMA paper, until recently only two large scale studies have examined the discordance between clinical and autopsy diagnoses of malignant neoplasms - one in 1923 and the other in 1972. To compare more recent data with data of the past, a 10 year retrospective study was conducted at the MCLNO between 1986 and 1995. Autopsies were perfomed by residents under the direct supervision of attending pathologists. All neoplasms were coded using the SNOMED international coding system. Both a morphological and topological code were assigned to each neoplasm. The researchers established a database with patient demographic data, and autopsy data using SNOMED codes. All demographic data were obtained from autopsy protocols.
Of 1105 cases reviewed (654 males and 451 females) the mean age was 48.3 years and age at death ranged from 1-98 years. A total of 433 neoplasms were diagnosed, 250 of which were malignant. One hundred eleven malignant neoplasms in 100 patients had been either undiagnosed or misdiagnosed, and in 57 patients, the immediate cause of death was directly related to the malignancy (in the remaining 43% of the cases the most immediate cause of death was cardiovascular disease). In the 111 undiagnosed or misdiagnosed malignant neoplasms found, the most common sites of occurrence were the respiratory tract, the gastrointestinal tract and the genitourinary tract. These 3 systems account for approximately 83% of the undiagnosed or misdiagnosed tumors in the study set. The data are consistent with data from two other large studies of malignant neoplasms diagnosed at autopsy.
An editorial in the same issue of JAMA argues for an increased number of autopsies (25%) as a condition for hospital accreditation. The JAMA paper was written by Elizabeth Burton, Dana Troxclair and William Newman II, of the LSU Medical Center, New Orleans. The editorial was written by George Lundberg.
From a survey research and quality perspective, the Burton, Troxclair, and Newman paper raises several issues. The number of annual deaths in the United States has averaged about 1.9 million for the past 30 years, largely because the increase in population has been offset by a decrease in the death rate. Perhaps 200,000 autopsies are performed in American hospitals each year. Perhaps another 100,000 are performed by medical examiners. The autopsy rate was much higher several decades ago. Highly detailed autopsies - including those with microscopic slides of tissue - have been performed for over 100 years. As many as 1 million autopsies may have been performed in 1965 alone. Where are the data from these autopsies? Second, do 200,000 to 300,000 autopsies per year consitute a sample of sufficient size? Third, would doubling or tripling the number of autopsies performed annually solve longstanding medical quality assurance problems?
GPS Clock 'Rollover' - Major Data Quality Problem?
According to an article in the October 12th issue of Business Week, military and civilian users of the U.S. military's Global Positioning System may experience serious problems when GPS satellite clocks reset on August 22, 1999. When the GPS systems was designed in the 1970s, cost pressures and bandwidth limitations forced programmers designing the system to limit the cycle to 1,024 weeks. That way, the software that kept the GPS calendar could be squeezed into a small block of code.
Unfortunately, it is not clear how many of the estimated 3 million military and comercial GPS receivers will be able to handle the rollover. Receivers manufactured after 1995 should be immune to "rollover" problems. Certain corporations (utilities, manufacturing firms, and financial services corporations) utilize GPS signals to monitor, time, and synchronize operations - ranging from time-stamping financial transactions to synchronizing power allotments to customers. The U.S. Armed Forces face similar problems. They increasingly rely on GPS for everything from aircraft navigation to cruise missile and "smart bomb" targeting, using both commercial and military receivers. For now, civilian and military "authorities" recommend consulting GPS manufacturers and taking advantages of the fixes and upgrades the manufacturers offer. Even so, there's no guarantee GPS will be glitch-free next August 22nd. The article was written by Marcia Stepanek and appeared on page 87.
F.B.I.'s New DNA Database - How Predictive?
A front page report in the October 12th issue of The New York Times details the U.S. Federal Bureau of Investigation's plan to open a national DNA database on October 13th. The database, encompassing a new generation of forensic DNA techniques, promises to be so efficient (according to the Times) that some civil libertarians fear it will expand from people convicted of crimes to include almost everyone, giving the Government inordinate investigative powers over citizens.
The national DNA database consists of 50 databases run by the states but is unified by common test procedures and software designed by the F.B.I. Using the F.B.I.'s DNA testing protocols, it will be possible to compare a DNA sample from a suspect or crime scene in one state with all others in the system. Most states now require persons charged with serious crimes to provide a DNA sample. The F.B.I.'s DNA profiling system relies on STR's (short tandem repeats). A person's DNA is tested at 13 specific sites at which a short length of DNA is repeated. The number of repeats is highly variable from person to person. According to the Times, "measuring the number of repeats at the 13 designated sites gives a way of identifying each individual with a probability [of error] of one in several billion." [The Times didn't report who computed this probability]. The blood or other samples "donated" by individuals are retained by the states in collections known as DNA banks.
It's not known to what lengths federal and state authorities will go to develop DNA testing, databanks, and profiling. In England, the "authorities" can compel everyone who inhabits a building or village to give DNA samples. Some civil libertarians in the United States predict that DNA testing will eventually include everyone.
From an information quality and data quality perspective, the F.B.I.'s national DNA database poses serious problems. Unlike fingerprints, anyone can collect and transport human hair, blood, skin cells, and saliva from place to place. Moreover, DNA testing can "prove" that dozens or hundreds of people were at a crime scene....including one or more who the police would like to arrest. Nor does the Times article discuss how the police and their forensic laboratories would respond to situations where criminals deliberately attempt to deceive the police by dropping an item like a hat or glove belonging to a nearby resident at a crime scene. The article was written by Nicholas Wade.
Go to: Data Quality Home Page
Comments: dqemail@aol.com (10/11/98)