Newborn Screening Technology: Proceed With Caution

Abstract

The American College of Medical Genetics (ACMG) recommends a significant expansion in the number of conditions targeted by newborn screening (NBS) programs.1 In this commentary we advocate a more cautious approach. NBS dates to the early 1960s, when the technology developed to conduct large-scale testing on dried blood spots for phenylketonuria (PKU).2 PKU remains the paradigm condition for NBS because of features of the disease and its treatment, which are particularly advantageous to population screening. It is a condition that silently causes neurologic devastation but is amenable to early detection and effective prevention with a diet of moderate burden and complexity.3 Many children affected with PKU and their families have benefited from state screening programs over the past 4 decades because of collaboration between health departments, families, primary care providers, and metabolic specialists.However, PKU screening is not an unmitigated success.4,5 There was initial uncertainty about whether children with variant forms of hyperphenylalaninemia required treatment and about whether affected children require life-long dietary management.6 Indeed, some children with benign conditions were seriously harmed from unnecessary restrictions in their diets.5 In addition, long-term studies demonstrate decrements in cognitive function for affected children and adolescents who are not fully adherent to the diet,7,8 yet adherence to the diet is challenging because of its poor palatability, high cost, and limits on insurance coverage in many policies. Affected women who are off the diet are at high risk of bearing severely neurologically impaired children.9 Only recently have many programs begun tracking affected women to enable notification, education, and management. These difficulties by no means negate the value of NBS for PKU, but they highlight the problems with the successful implementation of a population-based screening program even when a model condition is targeted.NBS is a system with many elements from blood-spot acquisition to long-term management in the medical home. Any weak links in this chain will impair the efficacy of the program. State health departments continue to struggle to garner adequate resources and expertise to maintain current programs.2 Other aspects of the NBS system, like insurance coverage for ongoing specialty care and special diets, are beyond health department control. It remains the case in medicine more broadly that population screening of asymptomatic individuals is rarely an effective approach to uncommon diseases.10–12 Therefore, the onus of responsibility should fall on those who propose population-screening programs for such disorders to marshal the data about benefits and risks and to justify the efforts and expense that these programs entail.Despite limited data about screening effectiveness in improving health outcomes, NBS programs across the country have added a wide variety of additional conditions over the years, leading to a marked variability in screening panels. 2,13 The trend toward addition of new tests is due, in part, to changes in test technology and to advocacy groups, often supported by family members of affected children who have sought to expand screening in their state for their particular condition. More recently, a number of advocacy groups have collaborated to promote expanded screening at the national level.14 In 2000, pediatric professional organizations, public health programs, and the federal government evaluated NBS and concluded that a uniform national panel should be developed and reflect the best available data and expert opinion.2 The ACMG was commissioned by the Health Resources Services Administration (HRSA) to conduct an analysis of the literature and gather expert opinion to provide recommendations for a uniform NBS panel. The final ACMG report was supported promptly by the March of Dimes,15 the American Academy of Pediatrics (AAP),16 and the Secretary’s Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children,17 and many states are adopting its recommendations to promote expanded NBS.In this commentary we raise 2 sets of concerns with the recommendations of the ACMG report.18 The first set of concerns focuses on the limitations of the ACMG process. These limitations lead us to conclude that the ACMG report should not be considered definitive and that short-term state and federal NBS policy decisions should be based on a variety of additional considerations. However, we also recognize that new and potentially valuable technologies are available for NBS, such as tandem mass spectrometry (MS/MS), and should not be shelved pending years of research. Rather, we suggest that the new technology be introduced within a research paradigm so that data on efficacy and cost-benefit analyses can inform policy decisions in the foreseeable future. Therefore, our second set of concerns is that programs are expanding rapidly, partly in response to the ACMG report, without the infrastructure in place to determine if the technology is bringing benefits or harms to children.The recommendations for expanded newborn testing by the ACMG are based primarily on a survey designed to ascertain professional and lay opinions about tests and conditions potentially amenable to NBS. A subsequent assessment of the literature was conducted as a secondary component, in part to validate the results of the survey. Professional and lay opinions can provide valuable information for policy decisions, but without adequate data, opinions provide a weak foundation for new initiatives. Scoring methods for screening and preventive programs, such as those used by the US Preventive Services Task Force19 and the AAP, identify professional opinion as the weakest form of evidence for policy decisions. This form of evidence is even less persuasive if the relevant opinions are not ascertained in a rigorous fashion.In the ACMG project, the major tool for ascertainment of opinion was a survey. A survey’s value depends heavily on the quality of the analytic framework that guides its design and implementation: the topics addressed in the survey, the wording of the questions, the definition of the population to be surveyed, the selection of a sample from that population, the way the data are analyzed after collection, and the way the data are used in making policy recommendations. In each of these methodologic elements, the process used by the ACMG raises concern. The ACMG survey has no coherent analytic framework. For each of 84 conditions, there are 19 distinct questions. The questions touch on topics relevant to standard criteria for test selection but in an arbitrary way. Some questions ask about objective facts such as the incidence of the condition. Other questions mix fact and value. For example, one question asks the respondent to rank the burden of the untreated condition on a 5-point scale from profound to minimal; to answer, the respondent must combine assumptions about the consequences for the infant with his or her normative judgment about how serious a burden those consequences represent. Not surprisingly, the raw data from the survey demonstrate a high degree of variability even for familiar conditions such as PKU.The lack of an analytic framework is also evident in the construction of the weights used in aggregating the survey responses to rank conditions. No justification is provided for the weights assigned to different responses on the survey. It is notable that aspects of test methodology were given equal weight with the potential for benefit for affected infants, a methodologic assumption that was not explained in the report and which runs counter to the screening principles that have traditionally justified NBS. For example, 200 points were awarded to a test if it was known to be associated with a life-saving treatment, but a test could also be awarded 200 points if it could be performed on a multiplex machine regardless of whether it had any proven benefits.In addition, the decision to heavily weigh ancillary benefits of testing such as information to inform future reproductive decisions and elimination of the “diagnostic odyssey” for parents goes beyond the traditional criteria for NBS tests. A large number of points could be scored on the survey for such benefits to families in the absence of any perceived benefit to affected children. By contrast, traditional criteria have primarily emphasized benefits to affected children with the recognition that benefits to families are important but secondary. These expanded criteria are not explicitly justifiedThe report does not specify clearly for what population the survey was intended and why. It states vaguely that “input and opinion were sought from a wide array of child health professionals, subspecialty care experts and individuals interested in newborn screening.” The survey was sent to selected individuals and posted on the Web, meaning that a response rate is impossible to determine. The survey itself listed 8 nonexclusive categories of potential respondents with instructions to check off all that apply. Four were categories of providers of screening services, and 3 were providers of health services (diagnostic, primary care, and specialty care). The last category was “consumer.” What this means is not clear, because no definition is provided (the actual consumers of NBS tests are newborns). Because the population is not specified, the report’s later discussion of the extent to which the sample is “broadly representative” is meaningless.Respondents were biased toward individuals actively involved in NBS services and lay advocacy groups. In contrast, only 10 primary care providers submitted responses. In addition, the responses per condition were highly variable, with an average of 7 of 84 conditions scored per respondent and 47 ± 20 responses per condition. Therefore, there is no assurance of consistency in the assessments across conditions. Indeed, it is possible for different conditions to have been scored by an entirely different set of respondents.The literature review lacked the standard methods of analytic framework, key questions, a literature-search strategy, inclusion/exclusion criteria, or systematic assessments of quality of evidence or gaps in evidence. The use of these techniques is now standard professional practice for comprehensive literature reviews. In addition, the literature reviews apparently were conducted by individuals knowledgeable about the draft recommendations derived from the survey. In some cases, the reviewers were members of the ACMG Working Group. This approach does not permit an independent assessment of whether survey results were supported by the literature.The report does not fully address a number of key issues including false-positive results, secondary targets, ethical issues, and service-delivery issues. There is inadequate discussion of the anticipated impacts of false-positive results and results of uncertain clinical significance. These outcomes will be common when using a large panel targeting poorly understood conditions.20 The positive predictive value for MS/MS is estimated to be ∼10%,21 meaning that there are 9 false-positive results for every true positive. False-positive results lead to significant program costs, but, more importantly, harms to unaffected children may result from unnecessary implementation of medications or severely restricted diets. Harm may also come to children with test results of unknown significance who are treated but who ultimately will be found not to have needed intervention. In addition, the literature consistently demonstrates that a subset of parents experience distress and long-term concerns over the health of their child after false-positive NBS results.22–25 False-positive results and results of unknown significance are expected and tolerable in programs that bring clear benefits to affected children. However, for conditions that are only marginally treatable or untreatable, the negative impacts of NBS programs may outweigh the benefits. That is, NBS programs for some conditions will produce more harm than benefit for children. It is clear that the potential negative impacts of screening must be explicitly balanced with potential benefits on a condition-by-condition basis. The ACMG report does not present these analyses.The concept of “secondary targets” in the ACMG report is insufficiently developed and justified. Secondary targets are defined as conditions that “are part of the differential diagnosis of a condition in the core panel or are clinically significant and revealed by the screening technology but lack an efficacious treatment (as with some identified through MS/MS technology) or because there are incidental findings for which there is potential clinical significance (hemoglobinopathies).”1 The report recommends that these results be reported to families but does not advocate that programs provide follow-up services to families. These conditions are not familiar to the vast majority of primary care providers. Management of this information and the conditions themselves (25 in number) will be very challenging for state programs, parents, and primary care providers.The report reflects no discussion of the ethical and legal issues relevant to a significant expansion of NBS programs. National committees addressing NBS have consistently included detailed discussions of the ethical and social implications of this technology, including 2 National Academy of Sciences reports,26,27 the AAP/HRSA Newborn Screening Task Force,2 and the President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research.28 More specifically, with respect to the ACMG Report, there is no discussion of the social, ethical, and legal implications of mandating a large number of new tests within current state programs, including many tests with uncertain benefits or lack of evidence for health benefit. The silence on these issues is particularly striking given the divergence of this report from the conclusions of such groups as the National Academy of Sciences and the President’s Commission.The limited availability of metabolic specialists in many areas of the United States will be a high barrier to the effective implementation of the ACMG recommendations. It is unclear how programs, families, and primary care providers will manage information generated on a large number of rare and complex conditions without readily available expertise. The ACMG report does not address this concern adequately.The report also does not adequately address issues of cost-effectiveness. It states that a basic cost-effectiveness study of NBS was conducted and “demonstrated that newborn screening is when with medical but the report itself does not provide any evidence to this A to a commissioned but the study has not been only the of a cost-effectiveness study without any analysis a of In this there is to question whether the study can be rigorous to the A cost-effectiveness analysis data, and the report itself to the limited data available on many that are to an ACMG on cost-effectiveness in to those of a report in the United and of Screening for of A a clearly the were to data to demonstrate cost-effectiveness of MS/MS technology for conditions than PKU and chain within the The lack of data to policy decisions on expanded NBS is by recommendations for new programs should from a of and In there is no discussion of to clinical detection of conditions as an to population screening. In the report reflects no significant to the on a condition-by-condition limitations of the ACMG process the value of the report for public policy not that the ACMG report a case for expanded screening for a large number of new conditions for every child in the The results and recommendations in this are important to the national but they should not be considered the definitive for major program decisions at the state or federal in the over expanded NBS is the lack of information on the efficacy of current NBS programs and the lack of research about the treatment of the conditions and the of screening to clearly justify policy decisions about which tests should be performed to This will be by new A of multiplex technologies such as MS/MS is the to for a wide variety of conditions from a In the technology will permit analysis for an even number of conditions and using dried blood these a is whether to manage tests for different conditions on the multiplex panel as tests or as a test programs only to results on conditions that clearly criteria, results on the conditions, and the potential benefits from those results, will be from families and care providers. the results on a large number of conditions for which there is limited or no evidence of benefit to affected children may harm to some children and families and is to be a poor use of This is a for which we can no the ACMG report not address this The expert panel an ethical to all results, including those for secondary conditions. The report emphasized the benefits to families even in the absence of benefits to affected children and the of multiplex These assumptions to their recommendations to the of NBS programs if the will lead to expansion of testing as test come such as multiplex to expand test because we can new tests with potential value does not a approach to public propose the of with the expansion and of NBS programs a research paradigm that will the questions about the benefits and within a is needed to the benefits and risks of early detection by NBS programs. The AAP/HRSA Newborn Screening Task explicitly for research to NBS technology and that the decisions about which tests to use in newborn screening programs often have been in an on recommendations from professional groups, state and newborn screening and research are to provide the evidence needed for decisions and The ACMG also long-term data and to NBS this was not has it been emphasized in subsequent policy in this should the of the clinical and of screening on affected individuals and with false-positive the only clinical of an NBS program in the United States is the of NBS can be justified in some in when the of a condition is understood and there are serious consequences that can be by early research may be the research foundation for the treatment of rare conditions is This is the result of the difficulties in data on the treatment of rare conditions. The fact remains that affected children through NBS programs are not or in research in which treatment can be or in which short-term or long-term outcomes can be Affected children are treated to the of the or providers. Therefore, many studies are based on of children in The of that these children are than experience is important but the of when using are A literature review by found that the clinical approach to metabolic conditions by NBS is not of of has been based to a large degree on and of disease with clinical A that treatment for has been based on and as they were with a based on clinical judgment and a evidence in needed in the expanded NBS implementation in state programs, and will about the benefits and of programs and over the best way to care for affected children.The way to public policy be to the questions addressed by the ACMG but with a methodology and to a set of issues. However, the ACMG has been supported by professional and lay organizations, and many states are actively expanded NBS programs using MS/MS given the in the methodology used to the ACMG they should not be considered the of for state programs. The ACMG report does not provide or data to conclude that the panel is for child or the effective use of does this state programs and federal toward policy 4 recommendations to with NBS programs addressing the lack of adequate implementation of expanded NBS should be conducted within a research could be or national in an approach or require in case of initial treatment and must be given to medical and social outcomes for with rare disorders who are identified with and without screening of in state testing and for with false-positive tests. on conditions that beyond the ACMG recommendations should also be screening in screening in and By data on a or national information could be on program and with a number of affected children to inform NBS policy decisions. A variety of research could be but all should to and of research as as quality of This approach the use of but to screening not programs to the use of or for conditions targeted by NBS programs should be and evaluated on a or national through The an have been in pediatric care because of the of national for the treatment of rare in children. in treatment can be by and evaluated in a rigorous using current as the data in a for these rare conditions is to making in treatment In addition, or national experience in or conditions of unknown clinical significance is in whether such conditions or or the of the of should be to the public health services, and ethical issues to multiplex The of a NBS program is to determine if the uniform panel is clinically and part of the one to and these A on MS/MS and technologies could result in recommendations on key should technology be evaluated from a public health and and how should this the expansion or of conditions on the uniform What is the responsibility of programs to results on conditions that not criteria for screening What is the responsibility of programs to or not results of uncertain clinical What are the of the medical and NBS program in the of results to parents for secondary conditions and conditions of unknown clinical it to screening for conditions that not traditional What are the legal for decisions not to selected results through multiplex A analysis of these and concerns provide to NBS programs as they struggle with new legal and public the data and ethical and policy analysis from the an independent and should provide a for comprehensive policy recommendations about NBS. this final is it may be valuable to a long-term process so that the initial research and policy can with the in these recommendations will require an degree of collaboration between states and a for the federal government in NBS than has been the case It is clear that of this research will require a significant new federal as However, NBS programs are important to the of and they should not be expanded without in place to their and to at the for discussion of these issues through their National National of Health for Newborn Genetic