The idea behind NEISS was all right, but the data were misleading. Chairman Terrence M. Scanlon says flatly that "the numbers were goofy." Suppose that a fellow arrives at an emergency room with a broken arm. How did it happen? He fell off a ladder. That is recorded as a ladder-related injury. The report goes into a computer with similar reports from 65 other hospitals. The batch of figures is cooked for a while, and subsequently it is announced that an estimated 10,000 ladder-related accidents occur every year.
The figure tells us something, but it doesn't tell much. Specifically, says Scanlon, the data won't tell us whether the ladder was at fault or the fellow was at fault. The guy may have been drunk and lost his balance. But the figures tend to scare the public, and the more the public is scared, the more the public will support the Consumer Product Safety Commission. Thus the
commission's practice has been to make the data public in an annual press release.
In an effort to clarify the goofy figures, Scanlon and two other commissioners, Anne Graham and Carol Dawson, voted on Feb. 6 to add two questions to the list of questions already asked. In addition to providing data on age, sex, nature of injury and type of product involved, emergency room patients would be asked: (1) Do you think you were using the product the way it was meant to be used? and (2) Do you think your accident happened
because the product you were using didn't work right?
The additional questions will be asked in the emergency rooms of eight hospitals, including Children's Hospital in Philadelphia, before the patient is treated. Scanlon says the experiment will cost only $1,500. It may point toward a more sensible understanding of actual product liability.
Armstrong entered an emphatic dissent. She found it ridiculous to subject injured patients to questioning "before medical treatment is administered." This pilot project, she said, will not improve the statistical validity of the NEISS data. It will not provide definite evidence of accident causality. The majority's decision runs counter to the recommendations of expert consultants. The supplemental data would lack reliability.
"The study methodology . . . is intrusive, inhumane, and irrelevant to the treatment of emergency room patients. In my opinion, it is highly objectionable to request injured patients, who are awaiting medical treatment, to take time to fill out a form which asks questions which have no medical relevance before the medical assistance which they require is administered."
To waste public funds on such unproductive research, she concluded, is ''unconscionable, irresponsible and inexcusable." Her three colleagues had shown "utter disregard" for their obligation to act responsibly. And she hoped the Office of Management and Budget would veto the experiment because of the President's renowned opposition to bureaucratic paper work.
Who's right? Armstrong and Stuart Statler have a point: It does seem ridiculous to question an accident victim before an intern can sew him up. Scanlon, Dawson and Graham are right in wanting to refine the misleading data produced by NEISS. If the admissions clerks at the eight hospitals will use common sense in asking the two extra questions, probably no harm will be done. And not much good, either.