The advantages of screening for diseases and disorders in children seem obvious. Ideally, tests catch problems early and increase opportunities for treatment and recovery. However, Jennifer Kwon and Richard H. Dees note that screening programs can have a number of complications, including ambiguous benefits, the need to educate families and the public, results that land in a gray area between normal and certain disorder, blurred lines between screening and research, and competition for scarce funding. Kwon and Dees urge caution and careful consideration of potential costs alongside potential advantages.
At first glance, the idea of screening infants and young children for diseases is a sure winner. Early detection of diseases can help families begin treatments that can prevent irreversible disability or death, so mass screening programs are a public health boon. By allowing society to treat illnesses before they manifest themselves, we keep patients from much harm and pain and save money by creating a healthier population and reducing long-term care needs. However, to evaluate these screening programs, we must understand the specifics of their costs and benefits, in terms of both the monetary implications and the effect on the quality of people’s lives. Just as important, we must look at the potential effects on everyone who might be affected by these programs.
The classic screening success story is the development of the nationwide program of newborn screening for phenylketonuria (PKU). PKU is an inherited disorder affecting 1 of every 15,000 infants in the United States. In PKU, the essential amino acid phenylalanine accumulates in the brain, leading to early brain injury and eventually to severe seizures and mental retardation. Dietary restriction of phenylalanine can prevent neurologic damage, but only if instituted in the first months of life. Therefore, early detection allows children with PKU to live productive lives. The test is easy to administer, accurate, low-cost even when administered to all newborns, treatment is readily available, and the payoff is extraordinarily high, since it prevents severe mental retardation. As a result, the PKU testing program is a paradigm for why and how we screen for diseases.
On the flip side, the classic anti-screening story is the disastrous attempt to screen for carriers of sickle-cell anemia among African American children in the 1970s. Carrier screening identifies individuals who are unaffected themselves but who may nevertheless pass a genetic disease on to their offspring. At the time, there was no cure for sickle-cell disease, so carrier screening might have been a useful means to help people make reproductive decisions to reduce the incidence of the disease.1 Indeed, a carrier screening program that began about the same time among young adults in the Ashkenazi Jewish population reduced the incidence of Tay-Sachs Disease, an always fatal neurodegenerative disease found almost exclusively in this population, by 90 percent. However, the sickle-cell test was aimed at African American children, rather than at the adults who could have used carrier information to make reproductive decisions, and the test was understood so poorly that even the U.S. Air Force refused to allow mere carriers to become pilots because it was assumed that they had the disease and were in danger of passing out in the cockpit. The program thus benefited no one, and it was charged that it led directly to racial discrimination.
Despite such cautionary tales, screening programs for infants and children have proliferated as new technologies have made it possible to test for many rare diseases from a single blood sample. For that reason, the American College of Medical Genetics (ACMG) now recommends mandatory newborn screening for 29 diseases.2 The vast majority of these disorders can cause irreversible neurologic injury over time if not identified and treated promptly. Some disorders, such as PKU and hypothyroidism, are effectively treated by dietary management or medications or both. In other cases, such as cystic fibrosis, the disorders are not presently curable, but early and meticulous medical management can lead to longer and more productive lives. In addition to screening newborns, new programs have been developed for other diseases that emerge during childhood. Recently published guidelines from the American Academy of Pediatrics, for example, suggest that children be screened for autism at regular intervals, beginning at 9 months, so that early treatment can begin and thus “minimize the core features and associated deficits, maximize functional independence and quality of life and alleviate family distress.”3
Advocacy groups call for more widespread pediatric screening, recommending testing for more than 50 disorders, most as part of the groups of tests done on newborns. This higher number is in part a reflection of improvements in technologies for diagnosing and treating rare disorders since the college’s 2005 recommendations, but it primarily reflects the different criteria that the advocacy groups use to judge the effectiveness and the value of the tests for particular diseases. These criteria vary among the different constituencies affected by screening: public health officials, families who have experienced the pain of lengthy diagnosis and failed treatments in rare disorders, and even biotechnology companies that stand to gain by prompt treatment of certain rare disorders.
Sensitivity versus Specificity
The accuracy of the tests and of the laboratories that analyze the results affect the value and usefulness of screening. In general, screening programs are designed to identify the people who have a disease (true positives) and to exclude those who do not (true negatives). But no matter how good the screening test, there will always be those incorrectly identified as having the disease (false positives) and those incorrectly identified as not having it (false negatives). A perfect test would identify all children who have the disease and only those children, but in the real world, a tension often exists between identifying everyone who has the disease (the test’s sensitivity) and excluding all those who do not have it (the test's specificity). In other words, to minimize the number of false negatives, we often accept more false positives. To assess the costs and benefits of any screening program, then, we must consider these four categories:
True negatives. The vast number of children screened would fall into this category, since most diseases do not have a high prevalence in children. After the test, the parents of these children know that their child does not have the disease for which he or she has been screened, and so they feel confident that they do not need to worry about it. With screening they gain some peace of mind—a small benefit, albeit one that affects a large number of people.
False negatives. For most screened diseases, the costs of a false negative are grave: where screening is justified, early treatment is essential to avoid severe disability or death, and the programs are explicitly designed to have no false negatives. If there are any such “missed” diagnoses, the consequences could be devastating—even if the infant being screened is in the same condition he or she would have been in without the program. Parents would not know the infant needed treatment. Worse yet, the test result may give families and even physicians a false sense of security that makes them slower to diagnose an affected infant.
False positives. Most of the otherwise avoidable adverse impacts of screenings fall on people who do not have a disease but who are identified by a test as having it. There is some risk for normal individuals to be subjected to treatments that are unnecessary or harmful. Therefore, it is important, though time- and resource-consuming, to perform confirmatory tests and counsel worried parents about the disease. These costs are not trivial. In any rare condition, finding the true cases of disease will result in a large number of false positives. So even for PKU, the number of false positives swamps the number of true positives. In New York, for example, 273 of the 240,000 newborns tested in 2007 had a positive result for PKU, but only 13 actually had the disease.
True positives. These patients obviously benefit the most from testing. In 2007 in New York alone, 655 newborns were discovered to have serious diseases for which immediate treatment was provided. Often money is saved as well, since the costs of early treatment are much lower than the costs of care for severely disabled patients.
We must also factor in the effects of screening programs on the public at large. The most obvious effect is monetary, since these programs cost taxpayers money—money that could be used for other purposes.
Given this framework, the basic argument for screening is usually that the benefits of early detection and early treatment for the true positives outweigh the anxiety and the occasional mistaken treatment of false positives and the costs of setting up and maintaining the program itself. In the cases in which early treatment has a clear impact on the children affected, such as PKU, such a calculation easily favors screening.
Few cases are so simple, though, for one of at least five reasons. First, the benefits of early detection may not be so clear. One of the factors that motivated the AAP recommendations for routine autism screening, for example, was the increasingly convincing evidence that early intervention—before age 4—significantly improves outcomes. However, the recommendations assume that routine screening—costly for such a large group of children—will discover a significant number of children with autism that normal checkups would miss. These otherwise-undiagnosed children will then receive services at a time when they are likely to be most effective. But the costs of treatment may be high, since early interventions usually involve intensive and very expensive "teaching" or "behavioral training." Moreover, for some, these interventions are unsuccessful, and still others will receive this expensive therapy because they screened positive, but they never would have developed the more severe symptoms. Because increasing the functional abilities of autistic children even by a small amount can significantly improve their lives and those of their families, autism screening seems justified on balance. But coming to that conclusion is much more complicated than it is for PKU.
Second, the need to educate both parents and the public about the testing and the disease complicates the basic argument for screening. Every parent should have some basic understanding of what tests their children undergo, and of course the parents of children who test positive require careful explanation of the results, the disease their child might have, and the options for future courses of action. In addition, the general public must be informed through various media about the tests and what they mean. Without such education, the screening can actively cause harm; indeed, if a diagnosis carries a stigma, as sickle-cell anemia did in the 1970s and as autism still does among some, then an extensive public education program is needed to overcome that misinformation and the resulting resistance to testing. Such education, then, is essential for success, but it also increases program costs.
Third, testing often reveals patients who have “abnormal” results but who do not clearly have the condition in question. For example, autism screening is likely to find children who lack important social skills but who do not clearly fall on the autism spectrum. We expect such children to benefit from increased services—but then, almost any child is likely to learn better with increased one-on-one attention and a careful scrutiny of his or her educational needs. Such “marginal” situations might also subject children to whatever stigma might be associated with a diagnosis. To take another example, consider the novel newborn screening program in New York State for Krabbe disease, a highly debilitating neurodegenerative disorder that leads to death in early childhood; it is caused by a deficiency in galactocerebrosidase, one of the enzymes needed for normal neuronal development. New York state, via the test, has identified newborns with low enzyme levels but without a severe enzyme deficiency, raising concerns that they may develop the disease later in life. No one knows exactly what these results mean in terms of prognosis, however. By monitoring the infants, we may discover whether they ever develop the disease and may be able to determine what enzyme levels correlate with eventual disease development, but efforts to do so have turned the screening program into a statewide clinical research project.
Indeed, in some cases, advocates for children with rare disorders may promote screening programs to invigorate research as a means of finding a cure. This is a fourth complication in the argument for screening. Krabbe screening was implemented in New York after intense lobbying from former Buffalo Bills quarterback Jim Kelly, whose son Hunter died from it. Currently, the only hope for a cure lies in the early use of umbilical cord stem cells. This treatment, however, involves chemotherapy and stem-cell transplantation in infants as young as one month old, since the treatment is effective only before neurologic symptoms appear. Because the experimental treatment is new, the long-term prognosis for those receiving the treatment is unclear, and the transplant itself is risky at such a young age. In the first full year of the screening, four children were found to have been at high risk for developing the infantile form of Krabbe disease; of those, two had transplants and one has died from transplant-related complications. Since Krabbe is fatal, we might think such experimental therapies are worth pursuing, and of course parents who disagree are free to refuse the treatment on behalf of their children. But the screening program now serves as a state-sponsored recruiting opportunity for Krabbe disease research.
At a minimum, the situation has strayed far from the ideal of a screening program such as that for PKU. We are now using legal mandates that compel parents to decide whether or not to allow their newly diagnosed children to participate in experimental treatments whose outcomes are still unknown.4 Such research programs may be legitimate, but they require a different kind of ethical justification. Essentially, the advocates of widespread screening are leveraging general public support for screening of curable diseases to advance their own research agendas. Their desire to do so is certainly understandable. For diseases such as Krabbe, a screening program may be the only means by which participants for clinical trials can be found. But the state cannot pretend that the testing is needed so that a proven treatment can be provided to those afflicted with disease. Since the program is research, it needs to be treated like research: It needs oversight from a group of individuals who are free of ties to advocacy groups that are promoting further testing or to surgery centers and to drug companies that provide treatment. In the case of Krabbe testing in New York, the state failed to set up any oversight board, and the informal group that has emerged has close ties to Kelly’s advocacy group, Hunter’s Hope. In sponsoring the oversight group, Hunter’s Hope has generously provided what the state should have established and did not, but the help has come at the cost of objectivity.
The basic argument for screening is complicated if we take into account the interest of the public at large. As the use of screening primarily for research purposes shows, all such programs must be evaluated in the context of public health in general. Insofar as any screening program uses limited resources, it must be evaluated not just on its own merits but also relative to other uses to which those resources might be put. Screening mandates rarely come with separate funding; therefore, most of the cost of the screening and follow-ups is drawn from a larger pool of money. Thus, to fund screening, cuts will be required in programs that do not have the same political cachet. In Mississippi, for example, the funding for increased newborn screening came at the same time as cuts in Medicaid, and infant deaths rose by 65 the following year.5 While a direct link between these two occurrences cannot be established, their correlation highlights the reality that, whenever resources are limited, we will have to make hard decisions about which programs to fund. As a society, we often pretend that we do not have to make such choices, but until we create a different health care system, increases in screening will affect other programs, and we will have to decide which of many worthwhile programs to eliminate.
Despite these caveats, screening programs can still be a public health boon. The key is to avoid getting caught up in the technological advances that make more tests possible or in the advocacy politics that create demand for more diseases to be tested. Instead, we must weigh carefully what we think a screening program can accomplish and what its costs will be, both for the people tested, whether or not they are affected, and for the public at large.