Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 41
Medical Professional Liability in Screening for Genetic Disorders and Birth Defects NEIL A. HOLTZMAN, M.D., M.P.H. In the spirit of Jacob Marley I am going to present glimpses of screening past (for phenylketonuria, or PKU), screening present (for fetal neural tube defects using maternal serum alpha-fetoprotein—MSAFP), and screening future (for a wide range of disorders using DNA-based tests). Physicians have been held liable for errors in PKU screening, and some will almost certainly be sued for mistakes in MSAFP testing. Despite its elegance, recombinant DNA technology, which is the basis of future screening, does not solve the problems of the past or present; exposure to liability will become greater as the magnitude of screening increases. Although this discus- sion is restricted to the problems of only one class of technological innovation, my concluding suggestions on how to reduce the chance of liability apply to many other innovations which, like screening, offer the promise of improving health outcomes. Screening, as I use the term, involves the testing of a healthy popula- tion to predict who is at increased risk of manifesting disease in the future or whose offspring are at increased risk. Because the pool of potential recipients of screening tests is very large (for example, all pregnant women or all newborns), the failure to screen even a small proportion could lead to malpractice suits when someone who was not screened manifests the disease. The less common the disease, the less likely are such suits. They have, nevertheless, occurred for PKU (inci- dence of about 1 in 12,000) and are more likely for neural tube defects (incidence of about 1 in 1,0001. 41

OCR for page 41
42 MEDICAL PROFESSIONAL [IABiLiTY: VOLUME I:! TABLE 1 Definition of Terms Used to Assess Screening Sensitivity The probability that a person (or his or her offspring) who will manifest the disease will be detected by screening, that is, will have a positive test result. Specificity The probability that a person (or his or her offspring) who will not manifest the disease will have a normal (negative) screening test result. Predictive value of a positive test result The probability that a person with a positive test result (or his or her offspring) will manifest the disease. Reliability (measures of test performance) Precision Repeated determinations yield the same result. Accuracy The determinations center around the true value. When screening is legally mandated (as is the case for PKU in most states), when physicians are legally required to offer it (as providers of obstetric care must do for MSAFP testing in California), when it be- comes the standard of care, or simply when reasonable people think screening should be done (and the capability to perform it is present), providers who fail to screen or offer screening will have a difficult time defending themselves from liability suits. The failure to screen is only the first problem. Some people destined to manifest a disease will be missed by a test because it is not perfectly sensitive (see Table 1 for definitions). Here, too, the less frequently the disease occurs, the less likely it is that many people will be missed by screening (false negatives). The Tow frequency of a disorder will not, however, reduce the chance that people who are falsely labeled at risk will bring suit for being needlessly exposed to potentially harmful interventions. When tests for different disorders but with the same sensitivity and specificity are compared, the one for the least common disease will have the greatest chance of being falsely positive. The predictive value of a positive test result depends not only on specificity and sensitivity but on the incidence of the disorder being tested. It may be easier to defend oneself against liability arising out of false negatives or false positives than against liability arising out of failure to screen at all. This is because the test may be biologically incapable of correctly labeling everyone who is screened. Although the mean concen- trations of substances such as phenylalanine or alpha-fetoprotein in maternal blood will be significantly different in those with and those without the respective disorder, considerable variation around the mean will result in some overlap between the two groups. Tests based on qualitative characteristics, such as mutations, may not be foolproof either. A mutation that is known to cause a disease in some people will not cause it in others. One of the difficulties in any individual case in which a mistake has been made is knowing whether it resulted from biological variation (in which case those responsible for the performance

OCR for page 41
SCREENING FOR GE~TiC DISORDERS AND BIRTH DEFECTS 43 of the test would be exonerated) or from faulty performance of the test. Frequently, by the time an error in screening is suspected, the specimen is no longer available for repeat testing. An indirect gauge of the chance that the laboratory made a mistake can be obtained by measuring the reliability of the laboratory on other specimens. When this assessment is done systematically, it is known as proficiency testing. SCREENING PAST: PHENYLKETONURIA Phenylketonuria (PKU) is an inherited disorder of amino acid metab- olism in which the accumulation of phenylalanine is almost invariably associated with severe mental retardation. In 1954 Bicke] and col- leagues demonstrated that the concentration of phenylalanine in the blood could be reduced by providing phenylketonurics with diets from which the phenylalanine had been largely removed.) The older infants and children in whom the diet was first tried failed to show any persis- tent, significant reversal of mental retardation, despite the decline in the concentration of the amino acid.2 The question remained whether administration of the special diet during or soon after the neonatal period could prevent retardation from appearing in infants with an inherited defect of phenylalanine metabo- lism. The approach seemed plausible because the placental circulation keeps the PKU fetus's concentration of phenylalanine at or close to normal in utero. We have since learned that early administration of the special diet can prevent retardation in the vast majority of children with PKU.3 In a few children who have different inherited causes for the increase of nhenvlalanine in their blood. the Tow-phenylalanine diet by ,~ ,` , itself is not effective.4 The proportion of infants detected by neonatal testing in whom devel- opmental delay would not be prevented by the low-phenylalanine diet was unknown when in 1965 most states passed laws that required the screening of all newborns for PKU. The laws were passed on the heels of a report by Guthrie and Susi of a simple, inexpensive test for detecting increases of phenylalanine concentrations in the blood.5 The test re- quired only a few drops of blood, which could be collected on filter paper from a prick in the heel of the newborn, and was therefore applicable to screening. Unfortunately, neither the effectiveness of the low-phe- nylalanine diet, nor the sensitivity and specificity of the screening test, nor the reliability of the laboratories performing it was established before newborns were routinely screened and those with positive results started on treatment.6 7 Without knowledge of the imperfections of the new technology, the probability of unfavorable outcomes and, conse- quently, of malpractice suits increased.

OCR for page 41
44 MEDICAL PROFESSIONAL LIABILITY: VOLUME II Problems did materialize. It soon became evident that the phe- nylalanine concentration that was being used as a criterion for diagnos- ing PKU was too Tow. The concentration must be at least five times the upper limit of normal before it is associated with retardation. Some children with lower abnormal concentrations were erroneously treated and suffered serious protein deficiencies as a result.8 (Phenylalanine is an essential amino acid; even phenylketonurics require it, although in much smaller amounts than do children without the condition.) Other children were treated before anyone realized that their elevations of phenylalanine were only transient. The problem of false negatives did not emerge quickly. The develop- mental delay produced by PKU often does not become evident until the second year of life. Even when examining 2-year-olds with developmen- tal delay, physicians are often slow to attribute the retardation to PKU because they place too much confidence in the validity of the screening test. In 1969 a resident at Johns Hopkins discovered that a 14-month-old infant who had been referred with developmental delay hac] phe- nylketonuria. The infant had been screened as a newborn, and the results had been reported as normal. This event, together with the finding that more boys with PKU were being detected by newborn screening than were girIs~espite the fact that the genetics of the disorder suggested that equal numbers should be detected prompted me and my colleagues at Hopkins to conduct a survey of PKU clinics and state health departments.9 We discovered 23 false negatives and found that the probability of PKU infants being missed was greater the earlier they were screened, particularly for girls. Further confirmation that age of screening was important came from comparing the sensitivity of screening in the United States, where most infants were screened at or before 4 days of age, to that in the United Kingdom, where most infants were screened (in their homes by health visitors) between 6 and 10 days. The sensitivity in the United Kingdom was 100 percent, compared with 93 percent in the United States.~° These findings suggested a biological basis for the less-than-perfect sensitivity of the screening test in the United States. (Since this study, milk feeding of newborns has started earlier; this practice has been associated with a lower probability of false negatives in early screening in the United States.) We also found evidence that U.S. screening laboratories Mitered markedly in the frequency with which they found elevated phe- nylalanine concentrations, a result that suggested variable quality. On average, there were about 10 false positives for every PKU infant detected. At about the same time, a report from the Centers for Disease Control (CDC) revealed the poor proficiency of several ofthe laboratories that were routinely performing screening tests. Exacerbating the

OCR for page 41
SCREENING FOR GE~TiC DISORDERS AND BIRTH DEFECTS 45 problem of quality, and making it difficult to control, was the large number of laboratories performing newborn screening tests in the United States.~3 In the United Kingdom and Ireland, where screening programs are centralized and where the communication of results fol- lows well-defined policies, not only was the sensitivity of the PKU test higher than in the United States but the interval between screening and follow-up was much shorter.~4 It became clear that, in view ofthe chance of test error, as well as the chance of biological false positives, any positive screening test result should be followed up with another deter- mination of blood phenylalanine before diagnosing PKU or beginning treatment. (Follow-up is also needed for other disorders for which new- borns are now routinely screened.) A more recent survey of state health departments by investigators at the CDC revealed 43 PKU infants who had been missed by screening, a minimum of 1.4 percent of all PKU infants screened.~5 (Health depart- ments are unlikely to know of all missed cases.) A few ofthe infants were missed because a specimen never reached the screening laboratory; a few others were missed because a positive test result was never followed up. Such problems might be obviated if parents were better informed about screening. (Because screening in most states does not require parental consent, parents often do not know their infant has been screened until they are told that the test result is positive.) Most of the missed cases were attributed to the laboratory determination, although it was not usually possible to pinpoint the error or to be certain that it did not result from the biological limitations of the test. Laboratories analyzing relatively small numbers of specimens were more likely to miss infants with PKU than were those with greater test volumes, which suggests that quality problems were important. Legal action was taken in 15 of the 26 cases of which the respondents had adequate knowledge. Many of the cases were still pending in 1986 when the survey was reported, but settlements of up to $3 million have been made. The problems of PKU screening continue. In 1985 the American Bar Foundation examined the problem of legal liability in newborn screen- ing.~6 It concluded that failure of quality control often increased the likelihood of screening errors and, consequently, of legal liability. SCREENING PRESENT: MSAFP TESTING FOR NEURAL TUBE DEFECTS Only after Congress passed amendments to the Food, Drug, and Cosmetic Act in 1976 were manufacturers required to demonstrate the safety and effectiveness of diagnostic test materials and other "medical devices" before they could be marketed. Had such a law been on the

OCR for page 41
46 MEDiC~ P~FESSiONAL CITY: VOILE ~ books in the mid-1960s, PKU screening might have been better vaTi- dated before it was incorporated into standard neonatal care. (Phe- nylalanine kits have since been put in a category by the Food and Drug Administration EFDA] that requires them to meet "performance stan- dards." The FDA, however, still has not established the standards.) As the still-unfolding story of MSAFP testing suggests, determination of the effectiveness of a screening test prior to its marketing does not ensure that it will be used appropriately. Anencephaly, which is not compatible with more than a few days of survival after birth, and open spine bifida, which almost always results in paralysis below the waist and occasionally in hydrocephalus and mental retardation, are the two most common neural tube defects de- tectable by MSAFP testing. Together they are found in about 1 in 1,000 live births in the United States. Although their occurrence is genet- ically influenced in at least some cases, more than 95 percent of affected infants are born to families without a previous history of anencephaly or open spine bifida. The association between elevated concentrations of AFP in the blood of women in the second trimester of pregnancy and open neural tube defects was discovered by Brock and his colleagues in Scotland in 1974.~7 Medical centers serving areas in the United Kingdom in which the frequency of open spine bifida was several times higher than it was in the United States soon began to screen. The principal reason was to offer women carrying fetuses with neural tube defects the opportunity to terminate their pregnancies. Considerable data on the sensitivity and specificity of MSAFP testing were amassed in the United Kingdom and in a few other European countries by the time the FDA received applications for premarket approval of MSAFP test kits in the United States. By then it was known that, to detect about 70 percent of fetuses with open spine bifida and 90 percent of those with anencephaly, the upper limit of normal MSAFP would have to be set at about the 95th percentile. This means that 50 out of every 1,000 pregnant women who are not carrying an affected fetus will have a positive test result. If only 1 in 1,000 is carrying the affected fetus detectable by the test, then there would be 50 false positives for every true positive, giving a predictive value of a positive result of 2 percent. It might seem that any test that has such a poor predictive value is not worth doing, particularly if the abortion of unaffected fetuses results. One should recall, however, that tests used in populations in which the disorder being sought has a low prevalence will have low predictive values of positive results, even when the tests have high sensitivity and specificity. The specificity of PKU screening is at least 99.9 percent, but the predictive value of a positive result when the incidence of PKU is 1 in

OCR for page 41
SCREENING FOR GE~TiC DISORDERS AND BIRTH DEFECTS 47 12,000 live births is only 8 percent. When the specificity is relatively low but the prevalence is high as is the case for serum cholesterol screen- ing for coronary artery disease in healthy middle-aged men- the predic- tive values will also be low; only about 30 percent in the case of choles- tero] screening. The likelihood of low predictive values of positive screening test re- sults emphasizes the importance of using follow-up tests to confirm or cancel the positive screening test results. Follow-up for MSAFP tests includes a repeat MSAFP determination; amniocentesis, with measure- ment of the AFP and characterization of the acety~cholinesterase in the amniotic fluid; and ultrasound examination of the spinal region of the fetus. If these studies are properly conducted and interpreted, the chance that an unaffected fetus will be aborted is less than 1 in 200.~9 When such high probabilities can be attained at costs that are low compared with the costs oftreating the disorder without early detection, which is the case for both PKU20 and neural tube defects, screening is economically justified, provided that most people will accept screening and its sequlae. The problem with MSAFP testing in the United States is that there is no assurance that providers of obstetrical care recognize the need for follow-up studies or will be able to obtain them in a timely fashion. Fewer than half the medical and pediatric residents at alohas Hopkins, as well as practicing physicians taking continuing education courses there, could correctly estimate the predictive value of a test for a disor- der that occurred in 1 of 1,000 people when the test was falsely positive in 5 percent of unaffected people.22 This is the situation with which a physician is confronted when an MSAFP test is reported positive. Most respondents greatly overestimated the predictive value. Among obste- tricians who had participated in educational programs on MSAFP test- ing, only 22 percent knew that the predictive value of a positive result was less than 5 percent, and only 45 percent knew how to proceed when MSAFP test results were positive.23 (The percentages answering cor- rectly were higher among obstetricians who subsequently adopted MSAFP screening.) In evaluating premarket approval applications, the FDA does not take into consideration practitioner preparedness to use a test. Frank Young, the commissioner of the FDA, stated recently: The FDA also cannot decide for practitioners when a test is appropriate, and under what circumstances any particular test should be used. These are judg- ments that must be made for individual cases. FDA does not have now or should not have a direct regulatory role in the practice of a physician.24 This view was not shared by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP)

OCR for page 41
48 MEDICAL PROFESSIONAL LIABILITY: VOLUME 1:1 when MSAFP test kits were being considered by the FDA in the late 1970s. Fearing the inappropriate use of the tests and attendant mal- practice suits, they urged the FDA to restrict the sale of the kits to laboratories that agreed to coordinate the follow-up of positive test results among referring physicians and centers at which additional tests could be reliably performed.25 The 1976 amendments gave the FDA the authority to do this. Initially, the FDA proposed the restrictions sug- gested by the ACOG and the AAP, but in 1983 the agency withdrew the proposal and subsequently approved the marketing of MSAFP kits with virtually no restrictions.26 The marketing of new medical devices becomes a potent force for their adoption, even when knowledge and capabilities for appropriate use lag behind. Soon after the FDA gave unrestricted premarket approval to MSAFP kits, the ACOG legal department urged obstetricians to advise all of their prenatal patients of the availability of MSAFP testing and to document in each patient's medical record her decision regarding perfor- mance of the test. The rationale was to give obstetricians "the best possible defense" when women who were not tested had babies with neural tube defects.27 Although the ACOG urged obstetricians to learn more about MSAFP screening and follow-up, it failed to recognize that until obstetricians knew more about the procedures, malpractice could arise out of misuse as well as nonuse of the test. At least one state, California, has regulated MSAFP testing. Patients' fees are paid to the state program, which then contracts on a competitive basis with eight private and health maintenance organization (HMO) laboratories to perform MSAFP testing, using supplies and protocols provided by the health department. (California has a similar arrange- ment for newborn screening.) The concentration of AFP in the serum specimens is reported to the health department, which determines whether the values are abnormal. The results are mailed to the refer- ring physician, and positive results are also sent by computer to 14 regional genetic centers. The center nearest the patient contacts her physician by telephone and arranges appropriate follow-up.28 Whether these restrictions will reduce malpractice claims and awards for MSAFP testing remains to be seen. The California system has also facilitated the collection of data on the association between low concentrations of AFP in the mother's blood and the occurrence of Down's syndrome in the fetus.29 Our finding that obstetricians who subsequently performed MSAFP testing had better knowledge of screening after participating in rudi- mentary education programs than those who did not adopt it30 suggests that physicians who will adopt a new technology are receptive to learn- ing more about it. Unfortunately, the opportunities to do so are not

OCR for page 41
SCREENING FOR GE~TIC DISORDERS AND BIRTH DEFECTS 49 always present. If federal or state agencies were to require such educa- tion, the chance of misuse might be reduced. Going one step further, third-party payers could require physicians to demonstrate an under- standing of how a new technology should be used before reimbursing them for providing it. Technologies that have the greatest potential for misuse could be singled out for this approach. Third-party payers are more likely to be interested when misuse will increase their costs of providing patient care. These costs do not usually include the costs of malpractice litigation. A more fundamental solution is to teach medical students and house officers to appreciate the dangers and difficulties posed by new technologies, including, in the area of screening, the probabilistic nature of test results. SCREENING FUTURE: DNA-BASED TESTING Until the advent of recombinant DNA technology, relatively few genetic disorders or birth defects were amenable to screening. There had to be some substance in a readily accessible body tissue, such as blood, whose quantity or quality indicated increased risk of future disease in the person being tested or in his or her offspring. For many disorders, such substances were unknown; for others, they could not be measured in readily accessible tissues. Recombinant DNA technology removes both of these constraints. Analysis of the DNA of white blood cells of children or adults, or easily accessible chorionic villi or amniotic fluid cells of fetuses, will reveal the presence of disease-causing or suscep- tibility-conferring genetic variants (alleles) even when the gene is not active at the time of testing or in the tissue used for testing. These variants arose, often several generations earlier, as a result of mutation. Before such analysis can be used for screening, scientists must iden- tify the disease-causing or susceptibility-conferring allele. This identi- fication can be accomplished with the new technology even when noth- ing is known about the normal function of the gene. (The following discussion is abridged from reference 31, in which citations can be found.) The first step is to localize the gene responsible for the disease of interest to a specific region of one of the 22 autosomes or the X or Y sex chromosomes. To accomplish this, blood is needed from affected and unaffected individuals in families in which there is strong evidence for Mendelian inheritance of the disease. The genes for several rare disor- ders have been localized, as have the genes for cystic fibrosis, Duchenne- type muscular dystrophy, adult polycystic disease, familial hyper- cholesterolemia, some forms of retinoblastoma, Alzheimer's disease, and bipolar affective (manic-depressive) disorder. In the next few years, genes that play a role in breast and lung cancer, hypertension, periph-

OCR for page 41
50 MEDICAL PROFESSIONAL LIABILITY: VOLUME II eral vascular disease, peptic ulcer, and schizophrenia are likely to be localized. Once the gene has been localized, it is possible to predict the risk of future disease in asymptomatic individuals, or in fetuses, who belong to families in which the disease has already occurred on an inherited basis. Localization does not permit population-based screening. Before screening is possible, the segment of DNA that constitutes the gene of interest must be identified and the DNA of the normal allele of the gene distinguished from the DNA of disease-causing or suscep- tibility-conferring alleles. These activities have already been carried out for Duchenne-type muscular dystrophy, familial hypercholes- terolemia, retinoblastoma, sickle-cell anemia, thalassemia, hemo- philia, and a few rare disorders, including PKU. A DNA sequence that increases the risk of insulin-dependent diabetes mellitus has recently been discovered. Because great strides have been made in simpli- fying the technology, it will not be long before companies that are currently developing DNA-based tests for genetic disorders will be ap- plying to the FDA for premarket approval.32 Tests may eventually be- come so easy to perform that physicians will use them in their office laboratories. In discussing screening past and screening present, ~ stressed the importance offollow-up. For many ofthe disorders for which DNA-based screening will be developed, follow-up is no less important, but confirm- atory tests, such as those available for PKU or neural tube defects, may not exist. It is true that DNA-based tests for disorders that are inherited in a straightforward Mendelian fashion (e.g., PKU or sickle-cell anemia) will be more specific than current tests for these disorders because they directly detect disease-causing mutations. Such disorders, however, are not the only ones for which tests are being developed. Searching for larger markets, the biotechnology companies working in this area are very interested in tests for common disorders—cardiovascular disease, cancer, and mental ilIness.33 In some families persons possessing alleles capable of causing these diseases will always manifest the disease, but this trend will not be the case in all families. Differences in alleles at other, modifying loci or differences in environmental factors will affect the expressivity of the gene. Until we can determine the presence of these other factory task in which success may prove elusive the predictive value of positive DNA-based tests may not be very high. Furthermore, until we understand more about the early, presymptoma- tic stages of these disorders, tests that confirm or cancel the positive screening test result will not be available.

OCR for page 41
SCREENING FOR GE~TiC DISORDERS ID BIRTH DEFECTS 51 It should be possible to determine the predictive value of DNA-based screening tests. This determination can be accomplished most quickly by performing the tests on a large number of unrelated people who are past the age at which the disease usually appears. The number of positive tests in people who remain free of symptoms, divided by the total number of positive tests (in those with and without manifestations of the disease), will approximate the predictive value of the test in younger individuals. The approach will also indicate the sensitivity of the test. For many diseases, more than one allele will be capable of causing the disease (an example of genetic heterogeneity). Tests that fad] to detect all of these alleles will not be perfectly sensitive. Data on sensitivity, specificity, and predictive value should be required as part of the premarket approval application and should be made known to the health providers using the tests. Genetic screening in the future will involve not only pregnant women and newborns but children and young adults. Test results that convey the risk of future disease in the person being tested will lead some people to modify their life styles or take medications to lower their risk of future disease. Test results indicating that the person being tested is the carrier of an allele that would place his or her offspring at increased risk will lead some people to avoid the conception or birth of such children. Much remains to be learned about how high risks have to be before people will act to reduce or avoid them and how people's tolerance for risk varies. Health providers also have a lot to learn about how to communicate risks objectively and effectively. Genetic counseling will be an important part of screening in the future, but the number of specially trained counselors is too few to meet the anticipated demand. Even if people do not want to know their own risks, insurance companies and employers will be interested in using genetic screening to identify people at risk for future disease or premature death. Insurance com- panies will not insure people at increased risk of some costly diseases— at least not at standard premiums—and employers could refuse to hire workers at increased risk to keep their health benefits costs down and reduce the chance of harmful reactions to the work place on the part of susceptible persons. As new tests are marketed, they will rapidly become the standard of care, as is now happening with MSAFP testing. The fear of liability is not deterring the development of new tests, and it will not deter the adoption of them once they are marketed. In several recent court deci- sions parents of affected children and the children themselves have been awarded damages because predictive genetic tests were not performed.34 Providers will test even when they have inadequate understanding of

OCR for page 41
52 MEDiC~ P~FESSiONAL CITY: VOLUME ~ the probabilistic nature of the results and are unable to counsel clients effectively. As DNA-based tests will be both falsely negative (owing to genetic heterogeneity) and falsely positive (owing to diminished expres- sivity), suits will arise unless there is widespread recognition that the tests cannot give definitive answers. The dangers of liability will be further increased by testing in laboratories whose proficiency has not been demonstrated. CONCLUSIONS From testing for one rare disorder 25 years ago, genetic screening has evolved to the point where screening for a wide range of disorder~some of them contributing significantly to total morbidity and mortality is technically feasible. For those genetic disorders whose manifestations can be prevented, delayed, or ameliorated only by presymptomatic inter- vention, genetic screening provides a unique opportunity to reduce the magnitude of disability. For genetic disorders for which effective inter- ventions have not been developed, screening can identify individuals or couples at risk of having affected offspring, giving them the option of avoiding conception (and having children through adoption, surrogate motherhood, or ovum, embryo, or sperm donation) or birth (by prenatal diagnosis and abortion). Although not everyone will view these options as benefits, there is little doubt that they will reduce the burden of disease. Yet in doing so there are dangers of misuse. My interest here is with misuse that increases the chance of professional liability and with what can be done to reduce that chance. The fear of malpractice if they do not screen will prompt many physi- cians to offer screening. This practice will reduce the number of suits brought because of failure to screen. When screening is offered in insti- tutional settings (e.g., hospitals or HMOs), systems that flag eligible patients who have not been screened could reduce the chance of uninten- tional failure to screen. The chance of incurring liability could soon be greater because of misuse oftechnology than because of nonuse. As very few, if any, popula- tion-based tests will be perfectly sensitive and specific, some people will have false-negative results and others false-positive results. Sensitivity and specificity of the test should be determined under screening condi- tions before tests are approved for marketing. The FDA has the author- ity to carry out this recommendation. Once a test is approved, the laboratories performing it should be monitored for proficiency. At present, the system of laboratory regula- tion is inadequate and varies from state to state. The role of the federal government, established under the Clinical Laboratory Improvement

OCR for page 41
SCREENING FOR GE~TIC DISORDERS AND BIRTH DEFECTS 53 Act of 1967, has diminished in recent years, although the Health Care Financing Administration performs some inspections. The College of American Pathologists organizes voluntary proficiency testing pro- grams. Because DNA-based tests represent a departure from other types of tests, the ability of laboratories to perform the tests reliably should be specifically examined. Limiting and centralizing the number of laboratories that will be reimbursed for performing a test—as California has done for newborn and MSAFP screening—will probably improve laboratory quality; it will certainly make it easier to monitor. The trend, however, is toward further decentralization of laboratories. More tests are being performed in physicians' office laboratories; as the technology is simplified, genetic tests could be performed there as well. Very few states regulate physi- cians' office laboratories. Health providers must be taught to recognize the probabilistic nature of screening test results. With proper understanding, they would not hesitate to screen again if they encountered a high-risk situation in a person with a negative test result, and they would confirm positive screening test results before taking potentially dangerous or irrevers- ible action. In offering screening and in communicating results, prop- erly trained providers would counsel their patients about the uncer- tainty attached to screening. This counseling is particularly important when no confirmatory tests are available. Until curriculum changes ensure that the vast majority of medical school graduates understand how to interpret screening test results, other measures are needed to reduce the chance of misuse and potential liability. Third-party payers should consider requiring some demonstra- tion of competence before they reimburse providers for tests. Alter- natively, states could require as California has done—the involve- ment of geneticists or other specialists in the follow-up of persons with positive results. The scope of genetic screening in the future is so large that it is likely to involve most people in making decisions about screening that conflict with their beliefs or attitudes. To prepare people to make these deci- sions, much more extensive education is needed about human genetics and the implications of genetic testing. As there is no assurance that everyone will either be taught or will learn the issues involved in testing, fully informed consent is essential to preserve individual auton- omy and to ensure that the individual understands the reasons for screening as well as the risks and uncertainties. My colleagues and have demonstrated that significant information about screening can be imparted in brief, easily understood disclosure statements.35 36 Greater understanding on the part of consumers will reduce the chances of

OCR for page 41
54 MEDICO PROFESSIONAL ARTY: VOLUME ~ malpractice suits, but, most important, it will increase the chances that testing will serve their best interests. REFERENCES 1. Bickel, H. J., W. Gerrard, and E. M. Hickman.1954. Influence of phenylalanine intake on the chemistry and behavior of a phenylketonuric child. Acta Paediat. 43:64-77. 2. Baumeister, A. A. 1967. The effects of dietary control on intelligence in phe- nylketonuria. Am. J. Ment. Defic. 71:840-847. 3. Holtzman, N. A., R. A. Kronmal, W. van Doorninck, C. Azen, and R. Koch.1986. Effect of age at loss of dietary control on intellectual performance and behavior of children with phenylketonuria. N. Eng. J. Med. 314:593-598. 4. Tourian, A., and J. B. Sidbury. 1983. Phenylketonuria and hyperphenylalaninemia. Pp. 270-286 in The Metabolic Basis of Inherited Disease, J. B. Stanbury, J. B. Wyngaarden, D. S. Fredrickson, J. L. Goldstein, and M. S. Brown, eds. New York: McGraw-Hill. 5. Guthrie, R., and A. Susi. 1963. A simple phenylalanine method for detecting phe- nylketonuria in large populations of newborn infants. Pediatrics 32:338-343. 6. Holtzman, N. A. 1977. Anatomy of a trial. Pediatrics 60:932-934. 7. Holtzman, N. A., E. D. Mellits, and A. G. Meek. 1974. Neonatal screening for phe- nylketonuria. I. Effectiveness. JAMA 229:667-670. 8. Holtzman, N. A. 1970. Dietary treatment of inborn errors of metabolism. Ann. Rev. Med. 21:335-356. 9. Holtzman et al. 1974; see note 7. 10. Starfield, B., and N. A. Holtzman.1975. A comparison of effectiveness of screening for phenylketonuria in the United States, United Kingdom and Ireland. N. Eng. J. Med. 293:118-121. 11. Holtzman et al. 1974; see note 7. 12. Ambrose, J. A.1973. Report on a cooperative study of various fluorometric procedures and the Guthrie bacterial inhibition assay in the determination of hyper- phenylalaninemia. Health Lab. Sci. 10:180-187. 13. Committee for the Study of Inborn Errors of Metabolism (CSIEM). 1975. Genetic Screening: Programs, Principles, and Research. Washington, D.C.: National Academy of Sciences. 14. Starfield and Holtzman. 1975; see note 10. 15. Holtzman, C., W. E. Slazyk, J. F. Cordero, and W. H. Hannon. 1986. Descriptive epidemiology of missed cases of phenylketonuria and congenital hypothyroidism. Pediatrics 1985. 78:553-558. 16. Andrews, L. B., ed.1985. Legal Liability and Quality Assurance in Newborn Screen- ing. Chicago: American Bar Foundation. 17. Brock, D. J. H., A. E. Bolton, and J. B. Scrimegeour.1974. Prenatal diagnosis of spine bifida and anencephaly through maternal plasma-alpha-fetoprotein measurement. Lancet 1:767-769. 18. Holtzman, N. A. In press. Genetic variation in nutrition requirements and suscep- tibility to disease: Policy implications. Am. J. Clin. Nutr. 19. Report of Collaborative Acetylcholinesterase Study. 1981. Amniotic fluid acetylcholinesterase electrophoresis as a secondary test in the diagnosis of anen- cephaly and open spine bifida in early pregnancy. Lancet 2:321-326.

OCR for page 41
SCREENING FOR GE~TiC DISORDERS AND BIRTH DEFECTS 55 20. Office of Technology Assessment, U.S. Congress. 1988. Healthy Children: Investing in the Future. OTA-t-345. Washington, D.C.: Government Printing Office, pp.93-116. 21. Meister, S. B., D. S. Shepard, and R. Zeckhauser.1987. Cost effectiveness of prenatal screening for neural tube defects. Pp. 66-93 in Prenatal Screening, Policies, and Values: The Example of Neural Tube Defect, E. O. Nightingale and S. B. Meister, eds. Cambridge, Mass.: Harvard University Press. 22. Holtzman, N. A., R. R. Faden, C. O. Leonard, G. A. Chase, A. J. Chwalow, and S. Richmond. Submitted for publication. Effect of education on physicians' knowledge of a new technology: The case of alpha-fetoprotein screening for fetal neural tube defects. Ibid. Young, F. E. 1987. DNA probes; fruits of the new biotechnology. JAMA 258:2404-2406. Holtzman, N. A. 1983. Prenatal screening for neural tube defects. Pediatrics 71:658-659. 26. Sun, M. 1983. FDA draws criticism on prenatal test. Science 221:440-442. 27. Annas, G. J., and S. Elias.1985. Maternal serum AFP: Educating physicians and the public. Am. J. Public Health 75:1374-1375. 28. Lustig, L., S. Clarke, G. Cunningham, R. Schonberg, and G. Tomkinson. In press. California experience with low MSAFP results. Am. J. Med. Genet. 29. DiMaio, M. S., A. Baumgarten, R. M. Greenstein, H. M. Sasi, and M. J. Mahoney. 1987. Screening for fetal Down's syndrome in pregnancy by measuring maternal serum alpha-fetoprotein levels. N. Eng. J. Med. 317:342-346. 30. Holtzman et al. Submitted for publication. 31. Holtzman, N. A. In press. Proceed with Caution: Predicting Genetic Risks in the Recombinant DNA Era. Baltimore, Md.: Johns Hopkins University Press. 32. Office of Technology Assessment, U.S. Congress. 1988. The commercial development of tests for human genetic disorders. Health Program staff paper. Washington, D.C. 33. Ibid. 34. Andrews, L. B. 1987. Medical Genetics: A Legal Frontier. Chicago: American Bar Foundation. 35. Holtzman, N. A., R. Faden, A. J. Chwalow, and S. D. Horn. 1983. Effect of informed parental consent on mothers' knowledge of newborn screening. Pediatrics 72:807-812. 36. Faden, R. R., A. J. Chwalow, E. Orel-Crosby, N. A. Holtzman, G. A. Chase, and C. O. Leonard.1985. What participants understand about a maternal serum alpha-fetopro- tein screening program. Am. J. Public Health 75:1381-1384.

OCR for page 41