« AnteriorContinuar »
years from now, or 15 years from now or 20 years from now, we will see libraries of genetic information combined with epidemiological studies which converge into a catalog of probability figures which will tell employers what the percentage of occupational disease is likely to be at variance from the norm for a specific individual applying for a job.
Now if 5 percent is not fair but 10 percent is getting into the fair range, if that approach is transplanted into genetic screening, then you're going to have real chaos in hiring procedures it seems to me.
Let me ask you, Mr. Mitchell, under your reading of the FELA, if your company-and this is a hypothetical question—if your company had an X-ray which established that there was a 5-percent greater chance that the individual in question would incur an occupational injury and you decided that the person ought to be hired and the person later files suit after getting an occupational injury, would the preexistence of that 5 percent predictive X-ray be sufficient to send the case to the jury?
Mr. MITCHELL. In my opinion, there's no question but that it would, but on the premise that we were aware of the potential danger and we knowingly permitted the man to go to work. We would argue very strongly the point of view that I think you put in your question is true, and that is that this is not something for which we are responsible. The ultimate decision, however, would be made by the jury, as is the case in almost every FELA case.
Mr. GORE. But your company has an incentive to keep the decision away from the jury and with the court-with the judge-to keep the decision away from the jury rather.
Mr. MITCHELL. Not necessarily.
Mr. GORE. Well, let me ask the question in a different way. Would you advise legally the denial of employment in that job category to someone who had a 5-percent predictive X-ray or other indicator?
Mr. MITCHELL. If you're going to confine the arena of my advice solely to the Federal Employers Liability Act and eliminate all the other discrimination in employment acts to which we are subjected, my advice would be yes. Mr. GORE. What about a 1 percent?
Mr. MITCHELL. I think you get to a point where those of us who have studied law would say it's de minimis.
Mr. GORE. Two percent?
Mr. MITCHELL. If you came to a point where there is what I would consider to be any appreciable risk of injury, my advice would be if you have an opportunity to avoid it, avoid it. I do not believe that either medical science or what I perceive to be the future of genetics is sufficiently precise to draw the line between 1 and 2 percent. I'm not certain that there's anything that medically certain. But when you get up into the 5- and 10-percent range, as I think Dr. Rogers mentioned, you're beyond the realm of probable error and into the realm of where I think an injury an ordinary citizen is likely to say this company deliberately took a risk and we're going to make them pay for it.
Mr. GORE. Yes. Well, my point is, quite apart from a purely ethical determination of whether it's right or wrong to exclude someone from employment if there's an enhanced risk, there are preexisting ethical judgments embodied in statutes which will make it very difficult for companies to reach any apriori decisions.
Mr. MITCHELL. You're absolutely correct.
Mr. GORE. What other statutes might be in this category along with FELA?
Mr. MITCHELL. You mean in terms of the imposition of liability on employers?
Mr. GORE. Yes.
Mr. MITCHELL. Well, there are many other different kinds of compensation acts, but most of the industry—almost all of organized industry today is covered by one form or another of a workman's compensation act. Federal employees are covered by the Federal Employees Compensation Act. The longshoremen by the Harborworkers' Act and almost all State employees are covered by individual acts. The Jones Act and the FELA are parallel acts that as far as I'm aware are the only two remaining industrial compensation acts that predicate the ability to recover and the extent of the recovery on negligence and that's the heart of this problem.
Mr. GORE. Let me read to you, Dr. Rogers, a quote that was attributed to you at the 1973 "Conference on Low Back X-Rays” and ask you if this is an accurate quote and then explore with you some of the ramifications of it.
First of all, the quote: It behooves the railroad industry to ensure through its medical departments that its employees are in top-notch physical condition and also that the medical departments provide the railroad industry with applicants who are as near perfect physical specimens as is possible for us to find. In the face of staggering claims and settlements, we feel we must do everything possible to exclude the potentially serious back injury.
First of all, is that accurate? Dr. ROGERS. That is a correct quote, yes. Mr. GORE. And do you stand by that statement today? Dr. ROGERS. Yes, I do. I think it is inherent upon us as medical directors to furnish to the railroad people who are physically and mentally qualified to do their job.
Mr. GORE. The implications which are troubling to me, and not just within the context of your industry, are that if every industry involving strenuous physical labor insists upon perfect or near perfect physical specimens and science gives us increasingly accurate means of determining and defining perfect and near perfect, then we create some very difficult problems for possibly the majority of our citizens who will not fall in that category.
Dr. ROGERS. Here again, I think that this will depend on the expertise of the medical profession in arriving at a medical opinion of that back versus that job. We don't expect perfect backs. We'll never get them. But we do feel we should balance our findings on X-ray with what that back must be required to do.
Dr. MEYERS. May I add one comment there?
Dr. MEYERS. You will find that industries look for people to perform particular jobs with specific capabilities. For an example,
when I was working with the United States Steel Corp., we put out a search for deaf individuals. In this particular job we only wanted deaf individuals to work in that occupation. This was an occupation inside of a 45-inch pipe where the decibel levels were extremely high and despite ear protectors it was impossible to protect their hearing. So we found a number of completely deaf individuals and trained them to do welding operations inside of a 45-inch pipe.
So it's not that every industry in every job category is looking for a perfect specimen. I think I summarized in my paper and I think Dr. Rogers would agree that we have to insure, however, that the individuals who are going to be able to perform the job can do it safely and do it effectively.
Mr. GORE. I assume that both of you provide gonadal shielding for male applicants?
Dr. ROGERS. This is recommended by all of our radiologists.
Dr. MEYERS. We don't provide it, but it's routine for the department of radiology on all X-rays.
Mr. GORE. What about female applicants? I know it's difficult to provide similar shielding. Do you provide any such protection?
Dr. ROGERS. In the female applicant, we cannot get a satisfactory X-ray if we use gonadal shielding. Therefore, we have the requirement that the X-rays shall only be done at a certain time during the menstrual period.
Mr. GORE. When is that?
Dr. ROGERS. We like for this to be done within 10 days after the cessation of the menses.
Mr. GORE. What precautions do you take to protect people who are denied employment on the basis of back X-rays to prevent the adverse effect of labeling which make it difficult for them to find employment elsewhere?
Dr. ROGERS. In this regard, we advise the applicant that if he is further interested in the reason for his medical disqualification to have his personal physician write to me. In turn, I furnish this man's personal physician with all of the findings of our medical examination and give the reasons therefor.
Now if inadvertently we find something-say a tumor or a disease process in the spine such as tuberculosis of the spine or a tumor of the spine which the employee is not aware of, then we go directly to the employee and advise him that he should be seen by his personal physician immediately.
Mr. GORE. OK. Thank you very much. I appreciate very sincerely both of you—all of you coming to help us in this endeavor. I truly believe that the problems that we are going to face with regard to genetic screening are going to be very, very troublesome and we appreciate your assistance in helping us to understand the framework within which the proper questions have to be asked and I thank you very much.
Mr. MITCHELL. Mr. Chairman, could I make one final comment on the point you were making a moment ago? We, as a company policy, do not release medical findings without the consent of the person who's being examined, but I think it would be wise, if the committee is interested in exploring that particular concern, to consider it in the context of what the privacy acts of both the State and Federal Governments impose on companies in terms of their ability to disseminate the kind of information that seems to be concerning you. It is very unlikely that any company with competent legal advice would make any wholesale distribution or any indiscriminate distribution of this kind of information.
Mr. GORE. If laws on the books already provide adequate protection in that area, that's some comfort and we can direct our attention more vigorously elsewhere. Thank you very much for appearing.
Our final panel today-well, actually we'll take these witnesses one by one. Our next to last witness is Dr. Paul Rockey, Medical Director of the U.S. Public Health Service in Seattle, Wash. Dr. Rockey, we are pleased to have you with us today and I understand you have a slide presentation which we are looking forward to. Without objection, the entire text of your statement will be put in the record at this point and we invite you to proceed with all of it or a summary, as you see fit.
STATEMENT OF PAUL ROCKEY, U.S. PUBLIC HEALTH SERVICE,
SEATTLE, WASH. Dr. ROCKEY. Thank you, Mr. Chairman. I'm Dr. Paul Rockey. I'm Associate Director for Clinical Affairs at the U.S. Public Health Service Hospital in Seattle, Wash., and I'm assistant professor of medicine and health services at the University of Washington.
As we have heard this morning, preemployment screening of job applicants for health factors that might place them at increased risk for occupational disease or injury is a widespread practice. Findings from a health history, from a physical examination, and from specialized diagnostic tests often are used to assign employees to less hazardous duties, or to bar them from employment altogeth
A major factor that clearly must be taken into account is the validity or accuracy of the test used in screening. How well does that particular screening test select out those individuals who have a high probability of having or developing the disease or condition?
The predictive value of a screening test for a particular condition depends upon three factors: (1) the test's sensitivity, (2) the test's specificity, and (3) the actual frequency of the condition in the population being screened. [Slide.)
This table illustrates a common way of displaying such data. This is a simple 2-by-2 table. The left column indicates that the disease is present and the right column indicates that the disease is absent. The top row indicates that the test is positive and the bottom row indicates that the screening test is negative.
The sensitivity of a test is a measure of the test's accuracy in correctly identifying persons with the condition. It is commonly expressed as the fraction or as the percent of all persons with the condition who will have a positive test. In other words, sensitivity would be the true positives over the true positives plus the false negatives.
The specificity of a test is a measure of the test's accuracy in correctly identifying persons free of the condition. It is commonly expressed as the fraction or as the percent of all persons without the condition who will have a negative test. In other words, the true negatives over all those patients without the disease.
The positive predictive value of a test refers to the value of a positive test in predicting the presence of the condition. This is the percent of all those with a positive test who truly have the condition. In other words, it's a row percent along the top rather than a column percent, and it is expressed as the true positives over the true positives plus the false positives.
A factor that is not commonly recognized as influencing the predictive value of a test, in addition to sensitivity and specificity, is the frequency of the condition in the population in which the test is being applied.
A brief example will illustrate how the predictive value depends upon the actual frequency of the condition in the population. [Slide.]
An electrocardiogram taken during exercise is relatively accurate, as medical tests go. It is about 95 percent sensitive and about 95 percent specific for coronary artery disease. If this test is given to 1,000 patients with angina pectoris (chest pain), a group in which the frequency of coronary artry disease is 80 percent, of the 800 persons with disease, 95 percent or 760 will have positive exercise tests. Of the 200 persons without disease 95 percent or 190 will have negative exercise tests, but 10 will have positive tests. Therefore, along the top row, the positive predictive value of the test is 760 over 760 plus 10, or 98.7 percent.
However, what if the test were applied to healthy 18 to 30 year old job applicants, a group in which the frequency of coronary artery disease is about 2 percent? [Slide.)
Again, for a hypothetical pool of 1,000 applicants, now 20 of them will have coronary artery disease-19 of those 20 will be found by the exercise test. However, with only 95 percent specificity, there will be 931 negative tests among the 980 without the disease but there will be 49 false positive tests among those people without any disease. So along the top column, if we look at the predictive value of this relatively specific test, it's 19 over 19 plus 49 or 28 percent.
This example illustrates a major problem with using diagnostic tests as screening tools. A test which may have high predictive value in a medical diagnostic situation, such as patients with chest pain, retains little predictive value if it is applied to populations where the disease frequency is low and where a significant number of individuals may be misclassified. Individuals without risk of the disease are falsely labeled as being at risk. In this exmple, although 95 percent of those at risk for coronary disease are identified correctly, if such a test were used as a prerequisite for employment, 72 percent of those denied employment because of a positive test would have been misclassified.
Should industry be allowed to discriminate on the basis of such screening programs? The answer may depend upon attempts to balance positive and negative social values.
The Boeing Co. might convincingly argue for using the exercise cardiogram as a prerequisite for employment of its test pilots (and I'm sure it was used for the astronauts pictured in this room). However, should Boeing be able to use the test as a prerequisite for employment of its machinists or engineers?