PEDIATRICS Vol. 106 No. 3 September 2000, pp. 616-618

Universal Neonatal Hearing Screening

To the Editor.

Dr Paradise's commentary appearing in the March 1999 issue of Pediatrics raises to my mind 2 questions regarding universal neonatal hearing screening: Is this an appropriate time to institute universal neonatal screening? Would it be more appropriate instead to institute a 2-step screening process consisting of a high-risk questionnaire followed up with neonatal screening when indicated?¹

In general, decisions regarding the selection of screening techniques have largely been made by the test developers or specialty groups, such as the American Speech-Language-Hearing Association, and the rationale for their decisions is often unknown. However, since society is either benefiting from the test or paying the costs of errors, decisions regarding the use of the screening tests should reflect the values of society. This point is of particular significance, as some states have already mandated universal neonatal hearing screening. Therefore, it would behoove health professionals making such decisions to incorporate societal values in the decision-making process. One approach to achieve this is through an approach called decision analysis.^2,3

Screening tests generally produce 4 outcomes: A, correct identification of individuals with the problem; B, overreferrals; C, underreferrals, and D, correct identification of those individuals who are free of the problem. Figure 1 illustrates these outcomes and their relationship to each other.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Screening outcomes.

Decision analysis enables one to determine for any 1 or more screening procedures the costs and benefits of the 4 outcomes and to sum them to obtain an overall picture useful in making comparisons with alternative screening procedures. The analysis entails first the determination of the costs and benefits for individuals falling in each of the 4 cells. For instance, the costs and benefits (factors) associated with each of the cells is illustrated in Fig 2.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Costs and benefits.

Although the costs of screening, rescreening, diagnosis, and treatment may not be so difficult to obtain from the literature, the economic gain is more difficult to assess in view of the fact that about 30% of those with sensory neural hearing loss may have concomitant developmental disabilities such as mental retardation and cerebral palsy.⁴ Factors such as intrinsic gain, parental anxiety, and the value of identifying normal individuals can be obtained by sampling a cross-section of society with questionnaires. The second step is to sum the positive and negative factors in each of the 4 cells. The third step is review the literature for large-scale clinical trials of a test such as one for universal neonatal hearing screening to determine the relative frequency with which subjects fall into each of the 4 cells. The fourth and final step, giving a subtotal for each cell, is to multiply the probability obtained in step 3 times the sum of factors obtained in step 2. In the case of universal neonatal hearing screening, the results of this step may be surprising because the relative low cost of the procedure may be offset by the 90+% of overreferrals. The subtotals for the 4 cells are then summed to obtain the overall benefit of that particular screening procedure. Having completed such an analysis for universal neonatal hearing screening, one can conduct similar analyses for alternative procedures such as that proposed by Paradise and others.

Although the use of decision analysis provides an excellent approach to the selection of screening tests, there are certain limiting factors that may make the selection of certain procedures unfeasible, even though results indicate that the greatest gain can be obtained by using one particular procedure. Examples of such limiting factors are: 1) availability of diagnostic and treatment services (alluded to by Paradise); 2) affordability of services; 3) probability that follow-up diagnosis and treatment will occur; and 4) an overreferral rate that is unacceptable to diagnosticians. One might also question the use of economic criteria in making screening decisions. When financial resources are finite, there are few alternatives in deciding how to allocate health dollars.

In summary, a large clinical trial such as proposed by Paradise should be undertaken at this time. If so, decision analysis could be used evaluating the results of various screening options. In the meantime, it would be wise to curtail any further legislation to make universal neonatal hearing screening mandatory.

William K. Frankenburg
Departments of Pediatrics and Preventive Medicine
University of Colorado School of Medicine
Denver, CO 80237-5075

REFERENCES

1. Paradise JL Universal newborn hearing screening: should we leap before we look? Pediatrics. 1999; 103:670-672 [Free Full Text]
2. Hammond KR, Aldeman L Science and human judgment. Science. 1976; 194:389-396 [Free Full Text]
3. Hershey JC Consequence evaluation in decision analytic models of medical screening, diagnosis and treatment. Methods of Information in Medicine. 1974; 13:197-203 [Medline]
4. Van Naarden K, Decouflé P, Caldwell K Prevalence and characteristics of children with serious hearing impairment in metropolitan Atlanta, 1991-1993. Pediatrics. 1999; 103:570-575 [Abstract/Free Full Text]

To the Editor.

Dr Paradise raises important issues with respect to the identification of infants with congenital hearing impairment that pediatricians should carefully consider. Dr Paradise suggests that most affected infants can be detected through screening of high-risk newborns. The evidence from a number of studies suggests otherwise. Only about half of all affected infants have 1 or more risk factors at birth.^2-5 Moreover, not all infants with risk factors will necessarily be identified through targeted screening. For example, only 60% of infants in the United Kingdom with congenital hearing impairment who have known risk factors (neonatal intensive care unit, family history, craniofacial abnormality), one third overall, are identified through targeted hearing screening.⁶

Dr Paradise proposes an "alternative, comparably large-scale national effort" to evaluate screening of high-risk infants and increased public and professional awareness as alternatives to universal newborn hearing screening (UNHS). Evidence of this kind is already available from the United Kingdom, where the majority of health districts use both targeted newborn hearing screening and a developmental check of all infants' hearing at 6 to 9 months of age. The median age of entry into intervention in these districts is above 18 months of age.⁶ Intervention by age 6 months is important to ensure normal language development.⁷ The UK experience, including a randomized, controlled trial, is that programs using UNHS get the majority of children into intervention by 6 months of age, at a cost per case detected only slightly higher than with targeted screening.^6,8

We agree with Dr Paradise that attention must be paid to providing adequate facilities and professional services for follow-up of children identified with hearing impairment. We acknowledge the importance of the issue of parental concern regarding positive screening results and recommend that programs monitor and address these concerns. We also agree that a public awareness program to help identify infants with hearing loss is important to detect infants lost to follow-up and those who have a late-onset or progressive hearing loss.

We encourage pediatricians to use the framework provided by Paradise in formulating their positions on newborn hearing screening. At the same time, on the basis of the available evidence, we concur with the AAP policy statement⁹ that supports hospitals and communities in their efforts to develop and implement universal hearing detection and intervention programs. In so doing, these programs should address the important questions and concerns that Dr Paradise raises.

Scott Grosse
Mike Adams
June Holstrum
Kim Van Naarden
Coleen Boyle
National Center for Environmental Health
Centers for Disease Control and Prevention
Atlanta, GA 30341

REFERENCES

1. Paradise JL Universal newborn hearing screening: should we leap before we look? Pediatrics. 1999; 103:670-672
2. MMWR Serious hearing impairment among children aged 3-10 yearsAtlanta, Georgia, 1991-1993. MMWR Morb Mortal Wkly Rep. 1997; 46:1073-1075 [Medline]
3. Mauk GW, White KR Giving children a sound beginning: The promise of universal newborn screening. Volta Rev. 1995; 97:5-32
4. Mutton P Early identification of deaf babies. Lancet. 1998; 352:1951-1952 [CrossRef][Medline]
5. Stein LK Factors influencing the efficacy of universal newborn hearing screening. Hearing Loss in Children. 1999; 46:95-105
6. Davis A, Bamford J, Wilson I, et al. A critical review of the role of neonatal hearing screening in the detection of congenital hearing impairment. Health Technol Assessment. 1997;1(10)
7. Yoshinaga-Itano C, Sedey AL, Coulter DK, Mehl AL Language of early- and later-identified children with hearing loss. Pediatrics. 1998; 102:1161-1167 [Abstract/Free Full Text]
8. Wessex Universal Neonatal Hearing Screening Trial Group Controlled trial of universal neonatal screening for early identification of permanent childhood hearing impairment. Lancet. 1998; 352:1957-1964 [CrossRef][Medline]
9. American Academy of Pediatrics, Task Force on Newborn and Infant Hearing Newborn and infant hearing loss: detection and intervention. Pediatrics. 1999; 103:527-530 [Abstract/Free Full Text]

In Reply.

I appreciate Dr Frankenburg's comments, and particularly his endorsement of Fred Bess' and my view^1,2 that deciding for or against universal neonatal hearing screening should await the results of a large-scale randomized trial. Dr Frankenburg's cost-analysis approach is germane and interesting; to his list would need to be added other potential untoward consequences of false-positive identification, as we have discussed previously²: disruption of parent-infant bonding, unfavorable "labeling," unnecessary diagnostic tests, and inappropriate therapeutic procedures. In addition, agreed-upon dollar values for many of the positive and negative factors in such a cost analysis will not be easy to come by.

Dr Grosse and colleagues cite 4 references^3-6 as indicating that only about half of all infants with congenital hearing impairment have 1 or more risk factors (as listed on the Joint Committee on Infant Hearing high-risk register [HRR]⁷) at birth. None of the 4 references contain primary data; each refers, in turn, to other reports. In one of those reports,⁸ indeed only 48% of congenitally hearing-impaired children had a high-risk factor (as listed on an early [1982] version of the HRR⁹), but 33% had been intensive care nursery (ICN) graduates. The authors did not indicate how much overlap existed between the 2 subgroups, but surely the percentage who fell into either the then-current HRR category or the ICN category, or both, would have exceeded 50%. In another of the reports,¹⁰ involving 6- to 9-year-old children with educationally significant sensorineural hearing loss, 50% of the children had had a high-risk factor (again as listed on the 1982 version of the HRR⁹), but the authors indicated that including ICN admission as a risk factor would have raised that value to 63%. If, as seems likely, the study population included certain children with later-onset or acquired hearing loss whose loss could not have been detected as neonates, exclusion of those children from the denominator would have resulted in an even higher percentage classifiable as HRR/ICN, particularly if one referred to the most recent (1994) version of the HRR.⁷ Finally, in the United Kingdom study¹¹ cited later by Dr Grosse and colleagues, which involved a total of 53 781 newborns, 70% (37/53) of the infants eventually identified as having moderate to profound bilateral sensorineural hearing loss had high-risk factors and/or were in special/ICNs. I stand by my statement that most of the infants with sensorineural hearing impairment detectable in the newborn period will be found by screening only newborns who meet HRR criteria and/or are admitted to an ICN.¹²

The statement that "intervention by age 6 months is important to ensure normal language development" refers to the study by Yoshinaga-Itano and colleagues,¹³ limitations of which I have discussed previously.¹⁴ The study consisted of a nonrandomized comparison, some of its findings seem counterintuitive, and its authors considered its results suggestive rather than indicative of a cause-and-effect relationship.

Dr Grosse and colleagues express agreement with concerns that Bess and I have raised about adequacy of facilities and follow-up, about the need for greater public awareness, and particularly about the impact of false-positive identifications. Unfortunately, those concerns appear to be receiving little more than lip service in the drive to implement universal screening.

Once again, I refer the interested reader to a previously published exchange of correspondence (Pediatrics. 1994;94:948-963) that provides a comprehensive summary of arguments for and against universal newborn hearing screening. A more recent exchange (Pediatrics. 1999;104:351-355) adds new details. It seems likely that the arguments will continue to be voiced until such time as the issue is settled one way or the other by an appropriately designed and executed study or, less rationally and far more expensively, by the cumulative weight of favorable or unfavorable experience.

Jack L. Paradise, MD
Department of Pediatrics
University of Pittsburgh
School of Medicine and Children's Hospital of Pittsburgh
Pittsburgh, PA 15213-2583

REFERENCES

1. Bess FH, Paradise JL Universal screening for infant hearing impairment: not simple, not risk-free, not necessarily beneficial and not presently justified. Pediatrics. 1994; 93:330-334 [Abstract/Free Full Text]
2. Bess FH, Paradise JL Reply to letters concerning universal screening for infant hearing impairment. Pediatrics. 1994; 94:959-963 [Abstract/Free Full Text]
3. MMWR Serious hearing impairment among children aged 3-10 yearsAtlanta, Georgia, 1991-1993. MMWR Morb Mortal Wkly Rep. 1997; 46:1073-1076
4. Mauk GW, White KR Giving children a sound beginning: the promise of universal newborn hearing screening. Volta Rev. 1995; 97:5-32
5. Mutton P Early identification of deaf babies. Lancet. 1998; 352:1951-1952
6. Stein LK Factors influencing the efficacy of universal newborn hearing screening. Pediatr Clin North Am. 1999; 46:95-105 [CrossRef][Medline]
7. Joint Committee on Infant Hearing 1994 position statement. Pediatrics. 1995; 95:152-156 [Abstract/Free Full Text]
8. Elssmann SF, Matkin ND, Sabo MP Early identification of congenital sensorineural hearing impairment. Hear J. 1987; 40:13-17
9. Joint Committee on Infant Hearing 1982 position statement. Pediatrics. 1982; 70:496-497 [Abstract/Free Full Text]
10. Mauk GW, White KR, Mortensen LB, Behrens TR The effectiveness of screening programs based on high-risk characteristics in early identification of hearing impairment. Ear Hear. 1991; 12:312-319 [Medline]
11. Wessex Universal Neonatal Hearing Screening Trial Group Controlled trial of universal neonatal screening for early identification of permanent childhood hearing impairment. Lancet. 1998; 352:1957-1964
12. Paradise JL Universal newborn hearing screening: should we leap before we look? Pediatrics. 1999; 103:670-672
13. Yoshinaga-Itano C, Sedey AL, Coulter DK, Mehl AL Language of early- and later-identified children with hearing loss. Pediatrics. 1998; 102:1161-1171
14. Paradise JL Reply to letters concerning universal newborn hearing screening: should we leap before we look? Pediatrics. 1999; 104:354-355