Part II: PUBLIC HEALTH ASSESSMENT Chapter 7

Surveillance for Birth Defects and Genetic Diseases

Lorenzo D Botto¹and Pierpaolo Mastroiacovo²

¹Division of Birth Defects and Pediatric Genetics, Centers for Disease Control and Prevention, 4770 Buford Highway, MS F45, Atlanta, GA 30341
²International Center for Birth Defects, Institute of Pediatrics, Catholic University Rome, Italy

Birth defects, genetic diseases, and public health: historical perspective and scope of the chapter

About 1 in 20 live-born infants is expected to have a single-gene disorder or a condition with an important genetic component by age 25 years (1, 2), about 1 in 33 will have a major birth defect (3, 4), and a similar proportion will have a significant developmental disability (5). These infants will account for a disproportionate fraction of premature deaths (6, 7), pediatric hospitalizations (8-11), and health care costs (12, 13). Many of these conditions or their complications can be reduced by timely and effective primary prevention practices, and public health surveillance has the potential to significantly contribute to this goal.

Birth defects surveillance was born in part as a public health response to an international tragedy, the thalidomide epidemic of the late 1950s and early 1960s. Thalidomide, an anti-nausea drug then widely prescribed during pregnancy, caused thousands of infants to be affected by severe birth defects between 1958 and 1962 (14). Within a few years, birth defects monitoring programs were being created in Europe (e.g., in Norway, Sweden, and Hungary in 1964-1967), the Americas (e.g. in Atlanta, U.S.A, 1968; in Latin America [ECLAMC], 1968), and elsewhere in the world (15). Since then, public health surveillance for birth defects, while maintaining its original mandate to provide an early warning system for birth defect epidemics, has expanded its scope to include other public health functions such as assessing disease impact on morbidity and mortality, identifying causes, and evaluating policies. At the same time, because of the diverse and changing social, political, and economic environment, birth defects surveillance has continued to face new challenges, such as the increasing impact of prenatal diagnosis and pregnancy termination, and the changes in the way health care is managed in many countries.

The role of genetics in public health dates back to the same years that witnessed the initial growth of birth defects surveillance. Historically, the public health mandate of such surveillance was mainly driven, at least initially, by the development in 1962 of an easy and inexpensive test to identify infants at risk for phenylketonuria (the “Guthrie” assay). This test provided the health care community with a way to prevent the sequelae of metabolic disorders through newborn screening. Newborn screening has been expanded over the years to include several more disorders, and remains a powerful example of successful integration of genetics into public health practice. Newborn screening also exemplifies a major objective of such integration, namely, preventing the disease phenotype by identifying people at risk for disease and providing them with prevention opportunities.

Although very different at the surface, public health genetic programs and birth defects surveillance share much common ground. Similar populations may be involved (e.g., stillborn infants, newborns, infants, children), similar data sources may be used (e.g., birthing hospitals, pediatric wards, cytogenetic laboratories, vital statistics departments), many of the same medical and public health professionals may be involved (e.g., geneticists, pediatricians, perinatal and pediatric epidemiologists), and some of the same issues are likely to arise (e.g., risk for discrimination and stigmatization, use of confidential and private data). At the same time, however, each activity has its own peculiarities and faces unique challenges. Public health genetic programs, for examples, must navigate the complexities of genetic testing, particularly as new tests move from the research laboratory to the clinical and public health world, and they must be sensitive to the ongoing debate on the ethical, social, and legal issues associated with the use (and potential misuse) of genetic information.

In this chapter, we describe selected surveillance issues that are likely to be of increasing relevance as birth defects surveillance moves into the future. A more general discussion of birth defects surveillance can be found elsewhere (16-19). We build on these considerations to discuss the emerging needs of genetic disease surveillance, with major, though not exclusive emphasis on high-penetrance, single-gene conditions that usually manifest in the pediatric age group.

Public Health Surveillance and Birth Defects

Public health surveillance has been described as the collection, analysis, and dissemination of outcome-specific data to describe and monitor health events, with the explicit provision that these activities be ongoing, systematic and timely, and, most importantly, that they be linked to public health practices such as intervention and prevention programs (20, 21). In birth defects surveillance, data can be used to study the prevalence of birth defects, the morbidity, disability, and mortality associated with birth defects, their distribution (e.g., geographic and ethnic variation), and their evolution (e.g., natural history) in a population. Surveillance must also be timely, ongoing, and linked to public health practice, if surveillance data are to help meet public health needs such as detecting epidemics, defining and monitoring problems, generating hypotheses, stimulating research, evaluating measures of disease control and prevention, detecting changes in health practices, and facilitating planning.

Uses of birth defects surveillance

As already mentioned, monitoring prevalence and trends has traditionally been a major focus of birth defect surveillance. The data gathered from these surveillance activities can also used to evaluate the significance of apparent birth defect clusters. For instance, following a report of a suspected cluster of limb deficiencies in a coastal area of the United Kingdom, surveillance systems in other countries were able to quickly show that no such cluster had occurred at other coastal sites (22-24), thus providing cost-effective and timely reassurance to the general public and public health professionals. Surveillance data can also be used to independently verify trends reported by other programs or to suggest potential etiologic studies. For instance, a long-term increase of hypospadias was reported by two monitoring systems in the United States (25). The cause of this increase is still unknown, but it parallels similar increases reported in Europe and Japan in the last decades which leveled off in the mid-1980s (26). These trends are of particular interest in light of the current debate on the potential effects of endocrine disruption from natural and synthetic chemicals present in the environment (27-29). Although the study of trends alone are not likely to determine whether a cause-effect relationship does indeed exist, reports of trends may provide suggestive evidence that environmental factors are contributing to a particular health effect, and serve as a starting point for etiologic studies. Similarly, international differences in the prevalence of neural tube defects and in their time trends (30) suggest that a study which includes different worldwide populations may provide clues to the relative role of genetic factors (e.g., mutations in certain folate genes) and environmental factors (e.g., consumption of certain vitamins) in the etiology of neural tube defects. Data generated by surveillance systems may help monitor not only birth defects occurrence but also assess the mortality, morbidity, and disability associated with birth defects (11), and track them over time (7). Adding geographic information into disease-specific data can help identify areas where prevention efforts need to be intensified. For example, mortality among children with Down syndrome was found in one study (31) to be correlated with their region of residence, independent of presence of a heart defect or low birth weight, suggesting that variations in healthcare practices may account for at least part of the preventable mortality associated with the syndrome.

Birth defect surveillance systems, when integrated into disease registries, may also provide a convenient mechanism for enrolling patients in etiologic studies. For example, the Atlanta birth defect case-control study, which was conducted by the Centers for Disease Control and Prevention in the early 1980s and which formed the basis of many etiologic studies (32), was based on an existing population-based monitoring program. The discovery of valproic acid as a cause of spina bifida in humans further underscores the usefulness of monitoring programs in etiologic research. Shortly after the first report of an excess of valproic acid exposures among mothers of infants with spina bifida (33, 34), an international study of birth defect registries confirmed the association (35). In an analogous sequence, the report of a cluster of limb anomalies among a small clinical case-series of infants whose mothers had undergone chorionic villus sampling at one clinic (36) was soon followed by case-control studies based on birth defects monitoring systems that confirmed the association (37, 38)). Later studies further added to the evidence (39).

Data from surveillance systems can also be used for methodologic studies. For example, surveillance data were used to evaluate the ability of birth defects monitoring systems to identify new teratogens under different circumstances of exposure frequency, teratogen potencies, and etiologic heterogeneity of outcomes (17). These results were then applied to known teratogens such as thalidomide (40) and retinoic acid (41), and highlighted the strengths and limitations of surveillance under realistic scenarios. They also underscored, for example, how classification and coding can influence significantly a system’s ability to detect birth defect epidemics.

Data for surveillance: approaches and challenges

Data for birth defects surveillance can be collected from many different sources, including some that were not developed originally for surveillance. In many countries, the main data sources are birth defect registries, vital records, and administrative data systems, sometimes supplemented by provider surveys, sentinel networks, and other information systems. Data collection itself occurs through a variety of channels and under different sets of regulations. In some instances, data collection is mandated by statute or regulation, and the information is reported to public health authorities. In other cases, surveillance is promoted and coordinated by academic or research institutions, often on a voluntary basis. Different systems may coexist in the same country or region. In England and Wales, for example, the monitoring system coordinated by the governmental Office of National Statistics, which is based on notifications from physicians or midwives, coexists with other, smaller registries (e.g., the North Thames registry) that use additional sources of ascertainment (42).

Population-based and hospital based: beyond the labels

The coverage of the population in a surveillance system may vary, and traditionally a distinction is made between population-based and hospital-based systems. Population-based systems use for birth defects surveillance record data relating to all births to mothers resident within a defined area, irrespective of where the birth takes place. Hospital-based systems record data relating to babies born in one or more hospitals, irrespective of where the mother lives. Birth defect ascertainment from hospital-based systems may be more liable to bias than ascertainment from population-based systems. Pregnant women in whose pregnancy a birth defect was diagnosed prenatally, for example, may be transferred to a hospital of the registry network to give birth, thus artificially inflating the birth defect prevalence in the hospital-based registry. However, hospital-based systems may also provide several advantages: they may be more cost-effective, and they may provide high quality data, since hospitals in a hospital-based registry are often self selected on the basis of the interest and commitment of hospital staff and management. Population-based registries are also subject to limitations. Such systems require vast resources to cover a population completely, irrespective of place of birth. Thus these registries tend to be small, unless sustained by a national mandate as in some Scandinavian countries and Britain. However, even if reporting of birth defects and genetic diseases is legislated and funded by a national government, case ascertainment in large registries often relies on reports from health care providers or hospital staff (passive case ascertainment) rather than on record review by trained registry staff visiting the health facilities, and consistently high data quality may be difficult to achieve.

The traditional distinctions between hospital-based and population-based systems should probably be re-evaluated in light of the impact that selective pregnancy termination has had on the birth prevalence of birth defects. For instance, truly population-based systems should monitor pregnancy terminations in addition to stillbirths and livebirths, and these data should be collected on all resident women, irrespective of where the pregnancy termination takes place. Under these more stringent criteria, few birth defects surveillance systems, if any, are currently truly population-based. From a practical standpoint, the challenge for most current birth defects surveillance systems, regardless of their label, is to incorporate data sources on pregnancy terminations in a way that also deals satisfactorily with the ethical and legal issues associated with these confidential data.

Complementing traditional data sources

Data sources other than disease registries are also commonly used to collect information on birth defects, and their strengths and limitations have been discussed in detail elsewhere (e.g. (19)). Morbidity surveys, for example, can provide data on a disease’s natural history, its burden in the population, and on the distribution of a disease’s risk factors. In the United States, the National Hospital Discharge Survey (43) provides data on short-stay hospitalization for specific conditions; Medicare and Medicare claims data (44) provides morbidity estimates among specific segments of the population (e.g., the poor and the elderly); and the National Health and Nutritional Examination Survey (NHANES) (45), a survey on a nationally representative sample of people conducted through interviews, physical measurements and biologic samples, provides information on the distribution of common conditions and risk factors. Hospital discharge data are available in many countries and have been used, for instance, to assess the contribution of birth defects to pediatric hospitalization (11). These information systems can be particularly powerful when unique identifiers allow to reconstruct the morbidity experience of a well-defined cohort of people with specific birth defects or genetic diseases.

Mortality data are almost universally available from vital statistics records and provide population-based information on survival and cause-specific mortality (7). Limitations of these data include their reliance on coding systems that are not optimized for birth defects and genetic diseases (e.g., International Classification of Diseases – ICD codes), the often long delay between death and availability of the data, and limited information on risk factors.

Provider surveys are a potential source of data with which to track prevention and screening practices. These surveys are particularly useful if they are conducted on a representative sample and on an ongoing or recurrent basis. One example of such provider surveys is the National Ambulatory Medical Care Survey (46), an ongoing nationwide survey based on a probability sample of ambulatory visits in the United States.

Except for birth defect registries and vital records, most of these data sources are still infrequently used in birth defects surveillance. Moreover, data linkages across multiple data sources, though potentially valuable, are also infrequently done. Linked data sets have been used to assess the prevalence of heart defects in Sweden (47), and linked identifiers have allowed researchers to explore trans-generational risks for congenital anomalies in Norway (48). Although the practice of data linkage must take into consideration issues of confidentiality and privacy (see below for a brief discussion), it would undoubtedly increase the usefulness of many data sets and represent an efficient use of the limited resources available for birth defect surveillance.

Special Considerations in Birth Defects Surveillance

Birth defects surveillance poses specific challenges that stem from the nature of the outcomes under study. Below, we briefly discuss selected issues related to pregnancy termination, data confidentiality and privacy, targeting of specific populations at increased risk for specific birth defects, and birth defect classification and coding.

Prenatal diagnosis of birth defects and genetic diseases represents a potential opportunity for early treatment, as exemplified by in-utero surgery or early postnatal treatment of conditions such as spina bifida or diaphragmatic hernia. However, pregnancy termination of affected pregnancies seriously challenges many functions of birth defects surveillance, as many pregnancy terminations occur in settings such as clinics or outpatient surgeries that currently are not among the traditional sources of case ascertainment for many surveillance systems. Failing to incorporate these cases into the surveillance systems may have several consequences: birth defect epidemics may be missed, the impact of birth defects may be underestimated, and the effectiveness of primary prevention practices (e.g., the use of folic acid supplements to prevent spina bifida and anencephaly) may become more difficult to assess. Currently the impact of selective pregnancy termination appears to be particularly noticeable for some defects such as anencephaly and spina bifida. In several countries, at least half to nearly all known cases of anencephaly are prenatally diagnosed and the pregnancy is terminated (49). With the increasing availability of ultrasonography, it not unreasonable to expect that other severe birth defects will be affected by selective pregnancy terminations. In some areas, for example, such an impact is already being seen for severe heart defects such as hypoplastic left heart and pulmonary atresia (50-52). Incorporating data on pregnancy terminations into many current birth defect surveillance systems may not be easy, however. In some cases, this would entail dealing with a wide array of specialty clinics and physicians’ offices rather than with a few major hospitals. In addition, issues of confidentiality and privacy are likely to play a major role in determining whether these data are shared with the surveillance system. These considerations are compounded in countries where pregnancy termination is not legal.

The debate on privacy and confidentiality also affects in general the collection and use of birth defects data that include individually identifiably information. These same issues are also raised in connection with specific practices that could generate useful data for public health purposes, including data linkages across information systems and the adoption of national unique identifiers. The ethical, social and legal ramifications of surveillance are complex and include an individual as well as a societal dimension. It has been suggested (53) that it is useful to frame the debate about such surveillance in terms of social risk (54), defined as the danger that an individual will be economically or socially penalized should he become identified with a disfavored medical condition. It has been further noted (53) that social risk involves both the threat of risk (attitudes that threaten social harm) and the perception of risk (as it exists among those with the condition), and that both play a role in the acceptance of a surveillance systems. From a practical perspective, the complex determination of social risk, whether real or perceived, ensures that protection from social risk will have to involve not only legal measures (e.g., laws against discrimination) but also societal changes aimed at reducing or reversing socially harmful behavior as well as a systematic move to reform or improve health care access and coverage.

It has long been recognized that particular subgroups of the population may be at higher risk for birth defects and single-gene disorders than the general population. Some of these at-risk groups may be defined by socioeconomic or income level (e.g., an increased risk for spina bifida in low socioeconomic classes) or by race or ethnicity (e.g., increased risk for spina bifida in hispanic populations in the United States (55) and elsewhere(30, 56)). Specific efforts aimed at monitoring these populations and enhancing prevention opportunities are often warranted. Some of these groups, however, may be particularly difficult to trace and monitor over time. Poorer people, for example, may be less likely to seek or obtain health care, particularly in the absence of universal health coverage, and, if they are also part of a linguistic minority, they may be overlooked in mainstream surveys that do not accommodate linguistic diversity.

Once affected infants are identified and information is collected, the surveillance system is faced with the complexities of coding and classification. Ideally, codes must be specific and unique, so that a one-to-one correspondence is generated between the essential features of a child with birth defects and a set of codes. The coding scheme should reflect a developmentally meaningful classification to maximize the etiologic and pathogenetic homogeneity of the case groups on which surveillance and research are conducted. Such a coding system should also be widely used to facilitate data sharing and collaborative studies. No current systems, however, has all three attributes of unique and specific codes, mechanistic classification, and universal use. Although most monitoring programs and registries use the World Health Organization’s International Classification of Diseases (ICD), there are many variations. Because ICD has to deal with a wide range of human diseases, it is not sufficiently detailed for many specialized purposes, including birth defects surveillance. For example, omphalocele and gastroschisis, two abdominal wall defects that differ by etiology and pathogenesis, are assigned the same code (757.79). Furthermore , relatively few syndromes have a specific code, and even well-known disorders such as Beckwith-Wiedemann syndrome are assigned a generic code (759.89, “other specified anomalies”). In addition, many anomalies that are often termed “minor”, but whose presence is often of diagnostic importance in many teratogen-induced conditions, cannot be easily coded in ICD, as exemplified by the nail and digital anomalies seen in the fetal hydantoin syndrome (57, 58). To overcome some of these limitations, several extensions of the ICD codes for birth defects have been developed by some organizations, including the British Paediatric Organization (BPA), the U.S. Centers for Disease Control and Prevention (CDC), and EUROCAT. A few programs also use “home-made” coding systems in preference to ICD, usually for specific activities. Surveillance of multiple congenital anomalies, for example, is conducted in the International Clearinghouse for Birth Defects Monitoring Systems using one such “home-made” coding system (18). Some of the limitations of individual codes could be overcome quickly, at least in theory, by relatively minor revisions of the current system. Limitations of the classification system, however, reflect in part our still incomplete knowledge of normal and abnormal development, and thus improvements in defect classification await further progress in these areas. The current challenge to classifying and coding birth defects is to devise and disseminate a system that is sufficiently detailed to capture the basic differences among the outcomes that it monitors yet flexible enough to incorporate new findings provided by future research.

Making the best use of birth defects surveillance data

The public health value of surveillance stems in large part from how the information is analyzed and disseminated. Particular care should be taken to ensure that surveillance data are timely, easily accessible, and readily understandable.

Timeliness is central to a surveillance system’s ability to sustain disease control and prevention. In this context, it may be useful to distinguish among data that are needed immediately, recurrently (e.g., once a year), or only on a long-term basis (archival) (21). Data that are needed immediately, or as close to “real time” as possible, may include data on the occurrence of birth defects due to known teratogens (e.g., rubella or thalidomide embryopathy) or on otherwise preventable conditions (e.g. the birth of a malformed infant from a mother with metabolically uncontrolled phenylketonuria), or data showing a marked increase in specific birth defect rates. In such cases, prompt intervention may improve the outcome of infants with the condition, identify other people or pregnancies at risk, detect possible teratogens, and provide insight into which links of the prevention process failed (e.g., why a pregnancy occurred during treatment with thalidomide, or what factors, perhaps cost related, contributed to the failure of the phenylalanine-free diet). This level of timeliness implies that one must accept provisional data and the possibility of false alarms. Other types of data, on the other hand, may need to be evaluated and distributed on a less immediate basis, typically annually or semiannually. Examples of such data may include the birth prevalence of birth defects, the contribution of birth defects to infant mortality, hospitalization rates for selected birth defects, and results of surveys on folic acid supplement use. Although these data would also help investigators evaluating previously identified alarms or birth defect clusters, they would be of particular value in planning and evaluating health policies and practices. Finally, accumulated and revised data should be appropriately stored for long-term analysis of trends and for research.

Ease of access to data from birth defects surveillance is also an important consideration. In part, access is limited because such data are infrequently published in widely circulated medical journals, although exceptions occasionally occur (59). In fact, most surveillance reports are published in limited copies (e.g., as annual reports), if at all. Access is made somewhat easier by joint publication of surveillance data, as is done annually by the International Clearinghouse (60), though limited distribution often remains a problem. The Internet provides a powerful means of increasing access to publicly available surveillance data: for example, it reduces the time and cost associated with both distributing and accessing information, and it allows data from multiple sites to be organized through Internet links. Data from birth defects surveillance systems are beginning to become available on the Internet (see for example the list of links at the Web site of the International Clearinghouse at www.icbd.org), and this trend is likely to continue.

Current challenges of birth defects surveillance: the case of spina bifida

In many ways, spina bifida provides a good example of how surveillance data can be used to evaluate the impact of a major birth defect and reduce its burden of morbidity and mortality. Spina bifida, a congenital anomaly of the central nervous system, can cause death and severe disability (61). However, many of the complications of spina bifida that now occur are preventable (62). Furthermore, many cases of spina bifida could be prevented if women consumed sufficient amounts of folic acid before conception and in early pregnancy (63), and, in smaller measure, by strict metabolic control of maternal diabetes (64) and by avoiding in pregnancy certain anti-epileptic drugs such as valproic acid. Our increasing knowledge about the genetic basis for spina bifida is providing insights into the role played by common polymorphisms of folate genes, such as the thermolabile variant of MTHFR (65), in modifying the risk for spina bifida. Surveillance of spina bifida has traditionally relied on birth defects registries with access to birthing hospitals, and sometimes, in the absence of such registries, on alternative data sources such as vital records. As already mentioned, this traditional approach is seriously undermined by the significant proportion of cases of spina bifida that are prenatally diagnosed and by the termination of such pregnancies. In addition, genetic and environmental determinants of spina bifida are not commonly monitored, and if they are, these data sources are not well integrated with the traditional surveillance activities. Finally, surveillance of morbidity and mortality, especially beyond infancy, is rarely done. Thus a comprehensive approach to spina bifida surveillance would encompass multiple components, including a network of population-based birth defect registries, expanded to include data sources on prenatal diagnosis and pregnancy terminations. These birth defect registries would also collect storable genetic specimens such as blood spots or cheek swabs from the infant and parents for use in studies of the genetic determinants of spina bifida. They would also record information on potentially preventable exposures, that could be used for counseling, to reduce the risk for a further affected pregnancy in the same family, and to track the effectiveness of policies and campaigns aimed at reducing the number of preventable cases. Vital statistics data linked to the registries would provide information on survival, while linked hospital discharge data would enhance surveillance of co-morbid conditions. Furthermore, provider and population surveys would assess and track, in appropriate population groups, the attitudes, knowledge, and use of prevention measures, such as daily use of folic acid, preconceptional control of diabetes, or avoidance of certain drugs around the time of conception and in early pregnancy.

Public health genetic programs: emerging needs for population data

As mentioned previously, the modern origins of public health genetic programs may be traced to the early 1960s with the introduction of newborn screening for single-gene metabolic disorders. Although for many years the scope of these programs was generally limited to newborn screening, the explosive growth of genetic knowledge in recent years has given rise to new questions that require timely and accurate answers. For instance, what data are needed before a new genetic test (e.g., for MCADD – medium-chain acyl dehydrogenase deficiency) is considered for newborn screening, and what is the evidence that the new test is safe, effective, and cost-efficient? What is the genotype frequency of known and newly discovered mutations in different populations? Most importantly, what are the crucial social, legal, and ethical issues that should be considered when collecting and using genetic information for public health purposes? With the increasing interest in evaluating prevention effectiveness, even traditional activities have come under renewed scrutiny. For example, what impact does current genetic screening (e.g., for sickle cell disease) have on rates of disease, complications, disability, and death in the population? Are known prevention measures (e.g., penicillin prophylaxis for sickle cell disease) applied effectively throughout the population?

Public health surveillance can supply crucial information about single-gene disorders. These data can be used to define and monitor problems, evaluate prevention measures, detect changes in screening practices, and facilitate planning. Many of the issues discussed in connection with birth defects surveillance are also relevant to the debate on integrating genetics into public health, including data sources and disease registries; the complexities of coding and classification; the ethical, social, and legal ramifications; and the need for timely and easily accessible data. Certain aspects of single gene-disorders, however, demand specific considerations. For example, because the genotype frequency of a single-gene disorder is relatively stable (compared to the potential variability of birth defects rates), ongoing monitoring should specifically encompass the disease phenotype (as opposed to genotype), as well as the rates of complication, disability, and mortality. Moreover, the specific ethical, social and legal ramifications of using genetic information in public health surveillance must be carefully examined (53). Genetic testing also becomes a central issue, as advances in gene-based research lead to the development of new genetic tests. An expert panel from academia, industry, and government has proposed that genetic tests be evaluated according to their analytic validity (the ability of the test to classify the genotype), clinical validity (the ability of the test to predict the disease phenotype), and clinical utility (the modification of the clinical phenotype associated with the use or non use of the test)(66). Such rigorous evaluation would provide data that will allow researchers to evaluate whether genetic tests, as they move from a research setting to the clinical world and beyond, are indeed safe and effective. Such information should also be shared with the public and medical professionals.

Integrating genetics into public health: two case studies

Two single-gene conditions, phenylketonuria and MCADD, help to outline further the emerging data needs of public health programs that use genetic information. In many ways, phenylketonuria exemplifies some issues associated with single-gene disorders that are currently part of the newborn screening programs and for which prevention measures are available (e.g., a low phenylalanine diet). Key needs for these conditions include evaluating and monitoring morbidity, disability, and mortality in people with the disorder, as well as the population-wide effectiveness of prevention measures. For phenylketonuria, such monitoring would involve tracking infants identified through newborn screening programs and integrating information on neurologic, psychologic, and social outcomes from a wide range of sources, including pediatricians, hospitals, developmental disability registries, and school programs that follow infants at risk for developmental delays. In addition, affected women should also be followed, as it is known that they are at increased risk of having infants with major birth defects and developmental disabilities, if their phenylalanine levels are not carefully controlled. Timeliness is also crucial, to ensure that specific events (e.g., the birth of a child with birth defects born of a mother with phenylketonuria) generate effective public health action, such as providing appropriate referrals for treatment and counseling and, by evaluating what aspect of PKU prevention failed, identifying the measures that would prevent further cases from occurring (e.g., providing dietary advice and free distribution of foods low in phenylalanine). From a systemic perspective, the example of phenylketonuria also suggests that it may be useful to reevaluate the role of the newborn screening in public health programs. Although traditionally not viewed as monitoring systems, newborn screening programs possess some of the features commonly associated with population-based surveillance, such as complete population coverage, with the added advantage of having been in operation for many years. Because of this, the public health usefulness of newborn screening programs could be enhanced, either by expanding the functions of such programs, or by integrating them with other public health information systems. In either case, however, newborn screening programs would benefit from a formal surveillance evaluation, that would include assessing parameters such as their sensitivity, flexibility, acceptability, and timeliness (67, 68).

Medium-chain acyl dehydrogenase deficiency (MCADD), a recently identified disorder of fatty acid metabolism, has been associated with an increased risk for infant deaths that in some cases might be decreased by instituting appropriate feeding practices (69). Currently there is a debate on whether MCADD should be added to the newborn screening program. This debate highlights some public health issues that are typically associated with genetic testing of newly identified conditions and that become particularly important when these tests are evaluated for inclusion in newborn screening programs. Although a comprehensive analysis cannot be even attempted here, it should be at least noted that for MCADD and other newly identified conditions, the lack of data is even greater than in the case of phenylketonuria. In many cases, in fact, basic epidemiologic data are lacking, including the frequency of the major mutations in different populations; the association of specific genotypes with premature mortality and other health outcomes; the interactions between environmental and genetic factors in determining disease risks; and the natural history of the conditions.

Concluding remarks

Although public health surveillance has occurred in different forms for several decades, changing scientific and social conditions currently challenge public health’s ability to monitor and prevent disease. These same changing conditions, however, also offer new opportunities. For birth defects, for example, prenatal diagnosis can lead either to earlier treatment or to pregnancy termination. Failing to incorporate the latter in disease surveillance may seriously limit many public health activities that are based on our knowledge of birth defect rates. For single-gene disorders, the development of new genetic tests may contribute to earlier detection of potentially harmful mutations and to earlier treatment or prevention of the disease. However, indiscriminate use of genetic tests could be harmful. In both settings, the challenge for modern public health programs is to generate useful, timely, and accessible data, and make them available to the community of public health officials, medical professionals, policy makers, and the public. Such data may contribute to a constructive and informative debate which should ensure that advances in birth defects and genetic research translate into a real benefit to individuals and the community, within the boundaries of ethics and the law.

References

1. Baird PA, Anderson TW, Newcombe HB, Lowry RB. Genetic disorders in children and young adults: a population study. Am J Hum Genet 1988;42:677-93.
2. Czeizel A, Sankaranarayanan K. The load of genetic and partially genetic disorders in man. I. Congenital anomalies: estimates of detriment in terms of years of life lost and years of impaired life. Mutat Res 1984;128:73-103.
3. Chavez GF, Cordero JF, Becerra JE. Leading major congenital malformations among minority groups in the United States, 1981-1986. MMWR CDC Surveill Summ 1988;37:17-24.
4. Kalter H, Warkany J. Medical progress. Congenital malformations: etiologic factors and their role in prevention (first of two parts). N Engl J Med 1983;308:424-31.
5. Boyle CA, Yeargin-Allsopp M, Doernberg NS, Holmgreen P, Murphy CC, Schendel DE. Prevalence of selected developmental disabilities in children 3-10 years of age: the Metropolitan Atlanta Developmental Disabilities Surveillance Program, 1991. MMWR CDC Surveill Summ 1996;45:1-14.
6. Infant mortality–United States, 1993. MMWR Morb Mortal Wkly Rep 1996;45:211-5.
7. Yang Q, Khoury MJ, Mannino D. Trends and patterns of mortality associated with birth defects and genetic diseases in the United States, 1979-1992: an analysis of multiple-cause mortality data. Genet Epidemiol 1997;14:493-505.
8. Cunniff C, Carmack JL, Kirby RS, Fiser DH. Contribution of heritable disorders to mortality in the pediatric intensive care unit. Pediatrics 1995;95:678-81.
9. Hudome SM, Kirby RS, Senner JW, Cunniff C. Contribution of genetic disorders to neonatal mortality in a regional intensive care setting. Am J Perinatol 1994;11:100-3.
10. Ling EW, Sosuan LC, Hall JC. Congenital anomalies: an increasingly important cause of mortality and workload in a neonatal intensive care unit. Am J Perinatol 1991;8:164-9.
11. Yoon PW, Olney RS, Khoury MJ, Sappenfield WM, Chavez GF, Taylor D. Contribution of birth defects and genetic diseases to pediatric hospitalizations. A population-based study. Arch Pediatr Adolesc Med 1997;151:1096-103.
12. Economic costs of birth defects and cerebral palsy–United States, 1992. MMWR Morb Mortal Wkly Rep 1995;44:694-9.
13. Waitzman NJ, Romano PS, Scheffler RM. Estimates of the economic costs of birth defects. Inquiry 1994;31:188-205.
14. Lenz W. Malformations caused by drugs in pregnancy. Am J Dis Child 1966;112:99-106.
15. Källén B, Hay S, Klingberg M. Birth defects monitoring systems accomplishments and goals. In: Kalter H, ed. Issues and reviews in teratology. New York/London: Plenum Publishing Corp., 1984:1-22.
16. Källén B. Epidemiology of human reproduction. Boca Raton: CRC Press, 1988.
17. Khoury MJ, Holtzman NA. On the ability of birth defects monitoring to detect new teratogens. Am J Epidemiol 1987;126:136-43.
18. Khoury MJ, Botto L, Mastroiacovo P, Skjaerven R, Castilla E, Erickson JD. Monitoring for multiple congenital anomalies: an international perspective. Epidemiol Rev 1994;16:335-50.
19. Lynberg MC, Edmonds LD. Surveillance of birth defects. In: Haperin W, Baker EL, Monson RR, eds. Public health surveillance. New York: Van Nostrand Reinhold, 1992:157-172.
20. Thacker SB, Berkelman RL. History of public health surveillance. In: Haperin W, Baker EL, Monson RR, eds. Public health surveillance. New York: Van Nostrand Reinhold, 1992:1-15.
21. Thacker SB, Stroup DF. Future directions for comprehensive public health surveillance and health information systems in the United States. Am J Epidemiol 1994;140:383-97.
22. Castilla EE, da Graca Dutra M. Limb reduction defects and coastal areas [letter]. Lancet 1994;343:1034.
23. Mastroiacovo P, Botto L, Fusco D, Rosano A, Scarano G. Limb reduction defects and coastal areas [letter]. Lancet 1994;343:1034-5.
24. Botting BJ. Limb reduction defects and coastal areas [letter]. Lancet 1994;343:1033-4.
25. Paulozzi LJ, Erickson JD, Jackson RJ. Hypospadias trends in two US surveillance systems. Pediatrics 1997;100:831-4.
26. Paulozzi LJ. International trends in rates of hypospadias and cryptorchidism. Environ Health Perspect 1999;107:297-302.
27. Carlsen E, Giwercman A, Keiding N, Skakkebaek NE. Declining semen quality and increasing incidence of testicular cancer: is there a common cause? Environ Health Perspect 1995;103 Suppl 7:137-9.
28. Carlsen E, Giwercman A, Skakkebaek NE. Declining sperm counts and increasing incidence of testicular cancer and other gonadal disorders: is there a connection? Ir Med J 1993;86:85-6.
29. Giwercman A, Carlsen E, Keiding N, Skakkebaek NE. Evidence for increasing incidence of abnormalities of the human testis: a review. Environ Health Perspect 1993;101 Suppl 2:65-71.
30. International Clearinghouse for Birth Defects Monitoring Systems. Time trends 1974-1996: anencephaly, spina bifida, and Down syndrome. In: Botting B, Castilla E, Mastroiacovo P, Siffel C, eds. Annual report 1998. Rome: International Centre for Birth Defects, 1998:100-101.
31. Mastroiacovo P, Bertollini R, Corchia C. Survival of children with Down syndrome in Italy. Am J Med Genet 1992;42:208-12.
32. Erickson JD. Risk factors for birth defects: data from the Atlanta Birth Defects Case-Control Study. Teratology 1991;43:41-51.
33. Valproic acid and spina bifida: a preliminary report–France. MMWR Morb Mortal Wkly Rep 1982;31:565-6.
34. Robert E, Guibaud P. Maternal valproic acid and congenital neural tube defects [letter]. Lancet 1982;2:937.
35. Bjerkedal T, Czeizel A, Goujard J, et al. Valproic acid and spina bifida [letter]. Lancet 1982;2:1096.
36. Firth HV, Boyd PA, Chamberlain P, MacKenzie IZ, Lindenbaum RH, Huson SM. Severe limb abnormalities after chorion villus sampling at 56-66 days’ gestation. Lancet 1991;337:762-3.
37. Mastroiacovo P, Botto LD, Cavalcanti DP, et al. Limb anomalies following chorionic villus sampling: a registry based case-control study. Am J Med Genet 1992;44:856-64.
38. Olney RS, Khoury MJ, Alo CJ, et al. Increased risk for transverse digital deficiency after chorionic villus sampling: results of the United States Multistate Case-Control Study, 1988-1992. Teratology 1995;51:20-9.
39. Firth H. Chorion villus sampling and limb deficiency–cause or coincidence? Prenat Diagn 1997;17:1313-30.
40. Yang Q, Khoury MJ, James LM, Olney RS, Paulozzi LJ, Erickson JD. The return of thalidomide: are birth defects surveillance systems ready? Am J Med Genet 1997;73:251-8.
41. Lynberg MC, Khoury MJ, Lammer EJ, Waller KO, Cordero JF, Erickson JD. Sensitivity, specificity, and positive predictive value of multiple malformations in isotretinoin embryopathy surveillance. Teratology 1990;42:513-9.
42. International Clearinghouse for Birth Defects Monitoring Systems. Contributing monitoring systems. In: Botting B, Castilla E, Mastroiacovo P, Siffel C, eds. Annual report 1998. Rome: International Centre for Birth Defects, 1998:25-39.
43. Graves EJ, Gillum BS. National Hospital Discharge Survey: annual summary , 1994. Hyattsville, MD: National Center for Health Statistics, 1996.
44. Bright RA, Avorn J, Everitt DE. Medicaid data as a resource for epidemiologic studies: strengths and limitations. J Clin Epidemiol 1989;42:937-45.
45. Ezzati TM, Massey JT, Waksberg J, Chu A, Maurer KR. Sample design: third National Health and Nutrition Examination Survey. Vital Health Stat 2 1992:1-35.
46. Woodwell DA. National ambulatory health care survey: 1995 summary. Hyattsville, MD: National Center for Health Statistics, 1997.
47. Pradat P. Epidemiology of major congenital heart defects in Sweden, 1981-1986. J Epidemiol Community Health 1992;46:211-5.
48. Skjaerven R, Wilcox AJ, Lie RT. A population-based study of survival and childbearing among female subjects with birth defects and the risk of recurrence in their children. N Engl J Med 1999;340:1057-62.
49. International Center for Birth Defects, World Health Organization, EUROCAT. World atlas of birth defects. Geneva: WHO, 1998.
50. Smythe JF, Copel JA, Kleinman CS. Outcome of prenatally detected cardiac malformations. Am J Cardiol 1992;69:1471-4.
51. Allan LD, Cook A, Sullivan I, Sharland GK. Hypoplastic left heart syndrome: effects of fetal echocardiography on birth prevalence. Lancet 1991;337:959-61.
52. Daubeney PE, Sharland GK, Cook AC, Keeton BR, Anderson RH, Webber SA. Pulmonary atresia with intact ventricular septum: impact of fetal echocardiography on incidence at birth and postnatal outcome. UK and Eire collaborative study of pulmonary atresia with intact ventricular septum. Circulation 1998;98:562-6.
53. Burris S, Gostin LO, Tress D. Public health surveillance of genetic information: ethical and legal responses to social risk. In: Khoury MJ, Burke W, Thomson E, eds. Genetics and public health. Oxford: Oxford University Press, 1999:xx-xx.
54. Burris S. Law and social risk of health care: lessons from HIV testing. Albany L Rev 1998;61:831-895.
55. Prevalence of spina bifida at birth–United States, 1983-1990: a comparison of two surveillance systems. MMWR Morb Mortal Wkly Rep 1996:15-26.
56. Little J, Elwood M. Ethnic origin and migration. In: Elwood JM, Little J, Elwood JH, eds. Epidemiology and control of neural tube defects. New York: Oxford University Press, 1992:146-167.
57. Hill RM, Verniaud WM, Horning MG, McCulley LB, Morgan NF. Infants exposed in utero to antiepileptic drugs. A prospective study. Am J Dis Child 1974;127:645-53.
58. Speidel BD, Meadow SR. Maternal epilepsy and abnormalities of the fetus and newborn. Lancet 1972;2:839-43.
59. Birth defects surveillance data from selected states. Teratology 1997;56:115-175.
60. International Clearinghouse for Birth Defects Monitoring Systems. Annual report 1998. Rome: International Centre for Birth Defects, 1998.
61. Lynberg MC, Khoury MJ. Contribution of birth defects to infant mortality among racial/ethnic minority groups, United States, 1983. Mor Mortal Wkly Rep CDC Surveill Summ 1990;39:1-12.
62. Kinsman SL, Doehring MC. The cost of preventable conditions in adults with spina bifida. Eur J Pediatr Surg 1996;199:199.
63. Czeizel AE, Dudas I. Prevention of the first occurrence of neural-tube defects by periconceptional vitamin supplementation. N Engl J Med 1992;327:1832-1835.
64. Reece EA, Sivan E, Francis G, Homko CJ. Pregnancy outcomes among women with and without diabetic microvascular disease (White’s classes B to FR) versus non-diabetic controls. Am J Perinatol 1998;15:549-555.
65. Shields DC, Kirke PN, Mills JL, et al. The “thermolabile” variant of methylenetetrahydrofolate reductase and neural tube defects: an evaluation of genetic risk and the relative importance of the genotypes of the embryo and the mother. Am J Hum Genet 1999;64:1045-1055.
66. Holtzman NA, Watson MS. Promoting safe and effective genetic testing in the United States: final report of the task force on genetic testing. http://www.nhgri.nih.gov/Elsi/TFGT_final/ 1998.
67. Klaucke DN. Evaluating public health surveillance. In: Teutsch SM, Churchill RE, eds. Principles and practice of public health surveillance. New York: Oxford University Press, 1994:158-174.
68. Klaucke DN. Evaluating public health surveillance systems. Public health surveillance. New York: Van Nostrand Reinhold, 1992:26-41.
69. Wang SS, Fernhoff PM, Hannon WH, Khoury MJ. Medium chain acyl-coA dehydrogenase deficiency (MCADD). HuGE Reviews 1999