Newborn Screening — Setting Evidence-Based Policy for Protection
http://www.100md.com
《新英格兰医药杂志》
In the early 1960s, a Massachusetts program for testing neonates for phenylketonuria became the first organized effort to screen newborns for genetic or metabolic disease in order to identify treatable disorders before they became symptomatic. Since that time, newborn-screening programs have expanded to include additional genetic and nongenetic conditions and have been implemented in all U.S. states, as well as in other countries. Although the importance and clinical successes of such screening are well recognized, many issues in newborn-screening policy and practice remain controversial.1,2
Newborn screening in the United States is mandated and regulated by the states, with little direction or authority at the federal level. Consequently, there is marked state-to-state variation in the number and types of disorders that are screened for and, therefore, substantial disparities in rates of diagnosis and in the services provided to babies born in different states.1,2
In the past decade, after adverse outcomes (including death) occurred in infants born in states that did not test for certain detectable, treatable disorders, critics decrying the unfairness of the system mobilized to change policy. They have urged public health departments to expand the number of tests used and to develop uniform policies and practices for all states. Pressure for change has come from lawsuits brought by some parents of children with genetic disorders, from patient-advocacy groups, and from some clinical and public health professionals, among others.
Some experts in public health policy and medical ethics, however, worry about the premature adoption of screening for disorders whose natural histories are insufficiently understood and for which the available treatments have variable or uncertain efficacy — cases in which the limitations of and risks associated with screening are therefore uncertain. Other experts ask what the cost of additional testing would be and, in a world of limited resources, how its cost-effectiveness would compare with that of other public health programs.3,4 It is against this backdrop that a task force of the American College of Medical Genetics (ACMG), commissioned by the Maternal and Child Health Bureau of the Health Resources and Services Administration, recently issued its recommendations for a uniform newborn-screening panel and system.5
The ACMG task force evaluated candidate disorders through a two-step process: first, a survey assessing the clinical characteristics of 84 disorders, the analytic qualities of the tests for them, and follow-up and management of the resultant clinical conditions; and second, the establishment of an evidence base with regard to each condition through review of the literature and expert opinion and the assessment of the survey data in the context of the evidence. On the basis of this work, the task force recommended that states require screening for 29 core conditions and report 25 "secondary" conditions that may be detected incidentally through screening for the core conditions (see table).
Disorders Recommended by the ACMG Task Force for Inclusion in Newborn Screening.
The assessments made in the development of newborn-screening policy are critically dependent on the composition of the group that makes them; the product of a group that does not meaningfully represent all relevant stakeholders will probably be compromised. It is arguable whether the ACMG task force had appropriately balanced and diverse representation. There was, for example, minimal representation of some important areas of clinical expertise, and whereas some militant proponents of expanded screening were included, I think there was inadequate representation of scholars who take an opposing stand.
The criteria developed by the task force for evaluating candidate diseases differ in several important ways from most others.1,2 First, the task force minimized the importance of the relative incidence of a given condition; other expert groups have required that a condition be "an important health problem that occurs frequently enough to justify screening an entire population."2 This change of emphasis results from the use of screening technology (e.g., tandem mass spectrometry) that permits the simultaneous detection of multiple disorders, some of which are rare, with little incremental technical cost (although substantial additional nontechnical costs might ensue). It is also indicative of the unprecedented premium that this task force placed on the use of specific types of technology — namely, multiplex technology.
Second, the group's evaluation of the suitability of a disorder for inclusion involved a calculation of the benefits to the family and to society of early intervention — not just the benefit for the baby. This analysis may have important ramifications for policy formulation, since it shifts the emphasis of newborn screening away from the medical interest of the child alone. It is likely that these revised criteria will provoke much discussion.
Policies for newborn screening should be made after careful weighing of the associated benefits, risks, and costs. For each disorder under consideration, this requires analysis of its frequency and natural history, the quality of the screening test, and the safety, effectiveness, and cost-effectivenes of treatment. Unfortunately, the relevant evidence is often limited because of the rarity of disorders: 12 of the 29 disorders recommended for the core panel are thought to have an incidence of less than 1 in 100,000 live births. All the conditions have clinical forms that are treatable, but the quality of the evidence of improved health outcomes attributable to newborn screening is variable — largely because it is difficult to conduct well-designed randomized, controlled studies of rare disorders.
In addition, for most metabolic disorders, there is considerable heterogeneity of clinical presentation and natural history, and in many cases, we have an incomplete understanding of the extent and meaning of this heterogeneity. For some disorders, we do not know what a "true" positive screening test means at the time of diagnosis, and in some instances, an abnormal biochemical screening result will have no clinical correlate at all. In such cases, the provision of treatment or even a "diagnosis" might have undesired consequences. Overall, the rarity of many of the disorders meant that the decisions made by the ACMG task force were often based on small nonrandomized studies and expert opinion — low-order forms of medical evidence.
Other groups have also recently assessed conditions for newborn screening. The Health Technology Assessment Programme of the National Health Service of the United Kingdom, the Human Genetics Society of Australasia, and the Massachusetts Newborn Screening Advisory Committee all arrived at different lists of disorders for which to screen, all shorter than that recommended by the ACMG.
In addition to disorders meeting its criteria for mandatory screening, the Massachusetts committee identified a number that merited serious consideration but for which there was insufficient evidence to justify mandatory inclusion. To address such cases, a pilot program was developed, whereby Massachusetts offers voluntary testing, with an informed-consent process, for 19 metabolic disorders detectable by tandem mass spectrometry, as well as for cystic fibrosis. This system, used by more than 98 percent of people who have babies in the state, is required to undergo periodic review; disorders that are not included in the mandatory panel today may be incorporated later, as additional data clarify areas of ignorance or contention. The program also provides an ethical means of prospectively collecting comprehensive population-based data with regard to these disorders.
This model of a panel of disorders that meet acceptable screening criteria and for which the evidence base is strong, combined with a standardized ancillary research program, would appear to be a reasonable way to meet the need for both a sound, evidence-based public policy that protects newborns and the collection of vital information in areas where it is currently lacking. As with the model proposed by the ACMG task force, political will would be required to change the current system. And, of course, even these changes would represent only an incremental step in protecting the health of all children born in the United States; ultimately, unfairness will remain as long as there are inadequacies and unevenness in clinical follow-up and care.(Marvin Natowicz, M.D., Ph)
Newborn screening in the United States is mandated and regulated by the states, with little direction or authority at the federal level. Consequently, there is marked state-to-state variation in the number and types of disorders that are screened for and, therefore, substantial disparities in rates of diagnosis and in the services provided to babies born in different states.1,2
In the past decade, after adverse outcomes (including death) occurred in infants born in states that did not test for certain detectable, treatable disorders, critics decrying the unfairness of the system mobilized to change policy. They have urged public health departments to expand the number of tests used and to develop uniform policies and practices for all states. Pressure for change has come from lawsuits brought by some parents of children with genetic disorders, from patient-advocacy groups, and from some clinical and public health professionals, among others.
Some experts in public health policy and medical ethics, however, worry about the premature adoption of screening for disorders whose natural histories are insufficiently understood and for which the available treatments have variable or uncertain efficacy — cases in which the limitations of and risks associated with screening are therefore uncertain. Other experts ask what the cost of additional testing would be and, in a world of limited resources, how its cost-effectiveness would compare with that of other public health programs.3,4 It is against this backdrop that a task force of the American College of Medical Genetics (ACMG), commissioned by the Maternal and Child Health Bureau of the Health Resources and Services Administration, recently issued its recommendations for a uniform newborn-screening panel and system.5
The ACMG task force evaluated candidate disorders through a two-step process: first, a survey assessing the clinical characteristics of 84 disorders, the analytic qualities of the tests for them, and follow-up and management of the resultant clinical conditions; and second, the establishment of an evidence base with regard to each condition through review of the literature and expert opinion and the assessment of the survey data in the context of the evidence. On the basis of this work, the task force recommended that states require screening for 29 core conditions and report 25 "secondary" conditions that may be detected incidentally through screening for the core conditions (see table).
Disorders Recommended by the ACMG Task Force for Inclusion in Newborn Screening.
The assessments made in the development of newborn-screening policy are critically dependent on the composition of the group that makes them; the product of a group that does not meaningfully represent all relevant stakeholders will probably be compromised. It is arguable whether the ACMG task force had appropriately balanced and diverse representation. There was, for example, minimal representation of some important areas of clinical expertise, and whereas some militant proponents of expanded screening were included, I think there was inadequate representation of scholars who take an opposing stand.
The criteria developed by the task force for evaluating candidate diseases differ in several important ways from most others.1,2 First, the task force minimized the importance of the relative incidence of a given condition; other expert groups have required that a condition be "an important health problem that occurs frequently enough to justify screening an entire population."2 This change of emphasis results from the use of screening technology (e.g., tandem mass spectrometry) that permits the simultaneous detection of multiple disorders, some of which are rare, with little incremental technical cost (although substantial additional nontechnical costs might ensue). It is also indicative of the unprecedented premium that this task force placed on the use of specific types of technology — namely, multiplex technology.
Second, the group's evaluation of the suitability of a disorder for inclusion involved a calculation of the benefits to the family and to society of early intervention — not just the benefit for the baby. This analysis may have important ramifications for policy formulation, since it shifts the emphasis of newborn screening away from the medical interest of the child alone. It is likely that these revised criteria will provoke much discussion.
Policies for newborn screening should be made after careful weighing of the associated benefits, risks, and costs. For each disorder under consideration, this requires analysis of its frequency and natural history, the quality of the screening test, and the safety, effectiveness, and cost-effectivenes of treatment. Unfortunately, the relevant evidence is often limited because of the rarity of disorders: 12 of the 29 disorders recommended for the core panel are thought to have an incidence of less than 1 in 100,000 live births. All the conditions have clinical forms that are treatable, but the quality of the evidence of improved health outcomes attributable to newborn screening is variable — largely because it is difficult to conduct well-designed randomized, controlled studies of rare disorders.
In addition, for most metabolic disorders, there is considerable heterogeneity of clinical presentation and natural history, and in many cases, we have an incomplete understanding of the extent and meaning of this heterogeneity. For some disorders, we do not know what a "true" positive screening test means at the time of diagnosis, and in some instances, an abnormal biochemical screening result will have no clinical correlate at all. In such cases, the provision of treatment or even a "diagnosis" might have undesired consequences. Overall, the rarity of many of the disorders meant that the decisions made by the ACMG task force were often based on small nonrandomized studies and expert opinion — low-order forms of medical evidence.
Other groups have also recently assessed conditions for newborn screening. The Health Technology Assessment Programme of the National Health Service of the United Kingdom, the Human Genetics Society of Australasia, and the Massachusetts Newborn Screening Advisory Committee all arrived at different lists of disorders for which to screen, all shorter than that recommended by the ACMG.
In addition to disorders meeting its criteria for mandatory screening, the Massachusetts committee identified a number that merited serious consideration but for which there was insufficient evidence to justify mandatory inclusion. To address such cases, a pilot program was developed, whereby Massachusetts offers voluntary testing, with an informed-consent process, for 19 metabolic disorders detectable by tandem mass spectrometry, as well as for cystic fibrosis. This system, used by more than 98 percent of people who have babies in the state, is required to undergo periodic review; disorders that are not included in the mandatory panel today may be incorporated later, as additional data clarify areas of ignorance or contention. The program also provides an ethical means of prospectively collecting comprehensive population-based data with regard to these disorders.
This model of a panel of disorders that meet acceptable screening criteria and for which the evidence base is strong, combined with a standardized ancillary research program, would appear to be a reasonable way to meet the need for both a sound, evidence-based public policy that protects newborns and the collection of vital information in areas where it is currently lacking. As with the model proposed by the ACMG task force, political will would be required to change the current system. And, of course, even these changes would represent only an incremental step in protecting the health of all children born in the United States; ultimately, unfairness will remain as long as there are inadequacies and unevenness in clinical follow-up and care.(Marvin Natowicz, M.D., Ph)