Intended for healthcare professionals


Measuring the prevalence of permanent childhood hearing impairment

BMJ 2001; 323 doi: (Published 08 September 2001) Cite this as: BMJ 2001;323:525

The introduction of screening makes this important and timely

  1. Shirley Russ, assistant professor of pediatrics (shirlyruss{at}
  1. Department of Primary Care Pediatrics, Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA

    Papers p 536

    In June 2000 Britain's health minister announced the introduction of universal neonatal hearing screening into the United Kingdom with initial pilot programmes at 20 sites. 1 2 Such universal neonatal screening is now also mandated in 35 of the 50 states of the United States, with legislation in other states pending.3 There is only one existing controlled trial of this approach,4 the remaining studies having compared their results with historical data. Although initial results are promising, they are primarily from hospital centres, often with a strong research interest, with relatively short follow up. Whether these results can be sustained when screening is introduced across whole communities, and when programmes are subjected to long term follow up, remains to be determined.

    Given this setting, it is particularly important that new, universal screening programmes are evaluated adequately. One way to achieve this in the short term is to compare the observed prevalence of targeted hearing impairments being detected by the new screening programme with the expected prevalence in the population. Unfortunately neither the United Kingdom nor the United States currently collect national data on all cases of permanent childhood hearing impairment, so expected prevalence rates are not readily available. Lack of prevalence data also hinders adequate service planning and the ability to monitor changes in patterns of hearing impairment over time.

    In an effort to overcome this problem, Fortnum et al performed an ambitious UK wide study to determine the prevalence of permanent childhood hearing impairment in all children born in 1980-95.5 They report their findings in this week's BMJ (p 536).5 In the absence of a national database, the authors had to resort to other methods of case ascertainment—namely, postal questionnaires to health and education providers. Inevitably with such voluntary reporting systems, there was under-ascertainment, though some compensation was possible with the use of statistical adjustment using capture-recapture analysis.

    The study has two important findings with implications for service providers. Firstly, and not unexpectedly, the observed prevalence of confirmed permanent childhood hearing impairment increased with age. What this means is that if we had been able to follow an entire birth cohort prospectively throughout childhood, the number of cases of confirmed hearing impairment would have increased with each passing year. This increase continues at least until the age of 9 and probably beyond. These findings are consistent with national data collected in Australia and with data from Denmark. 6 7 This rise in cumulative prevalence with age in part reflects delayed diagnosis of stable congenital impairments, with more cases coming to light as time progresses. However, some cases of permanent childhood hearing impairment are acquired postnatally—for example, after meningitis or chemotherapy, while others appear to be either late onset or progressive—for example, certain genetic forms of hearing impairment, some cases resulting from cytomegalovirus, and some cases following neonatal extracorporeal membrane oxygenation. Though all these factors probably contribute to the increasing prevalence of permanent childhood hearing impairment with age, the relative magnitude of the effect of each is uncertain.

    The second finding from the Fortnum study is that the prevalence of confirmed permanent childhood hearing impairment in the United Kingdom among children aged 9-16 years is higher than previous estimates, being at least 1.65/1000 for impairments >40 dB in the better ear and possibly as high as 2/1000. Current yields from neonatal screening programmes in the United Kingdom are about 1/1000, 4 6 so unless these yields improve, significant numbers of additional cases will still need to be diagnosed outside the neonatal period. Child health surveillance systems will need to remain vigilant to detect those children who pass their neonatal screen but subsequently acquire a loss; who have progressive or late onset impairments; who have mild impairments <40dB and hence are below the screen threshold; who migrate from countries not offering universal neonatal screening; or who missed their neonatal screen. Already, close audiological follow up has been recommended for certain risk groups,8 but not all the children who will develop later impairments have easily identifiable risk factors.

    It is important that both parents and professionals have realistic expectations of universal screening and what it can achieve. Whenever there is clinical concern about a child's hearing status—even if the child has passed the neonatal screen—prompt audiological referral should ensue. Good communication between the screening programmes, community health workers, audiologists, and physicians caring for children will be vital components of the new programmes. Only prospective comprehensive longitudinal birth cohort studies following the introduction of universal screening will give accurate information on the size of the problem of late onset, acquired, and progressive losses. Even then, in some cases it may be difficult or impossible to determine whether the loss was congenital or acquired, stable or progressive, particularly for the 40-70% cases where the cause is unknown.911 Clearly, much further research is needed in these areas.

    Consideration should be given to establishing national databases with core data on permanent childhood hearing impairment so long as this can be achieved in a way that maintains patient confidentiality and is acceptable to the community. Comparison of prevalence data between countries has hitherto been hampered by the use of differing case definitions, inclusion and exclusion criteria, and ages and subgroups of the population under study, resulting in widely varying estimates of prevalence, from 1 to 5 or more per 1000. 6 7 1012 Similarly, yields from neonatal screening programmes are reported in various ways, as including or excluding unilateral, mild, and conductive losses, and the screen may or may not have covered the whole population. Closer collaboration among members of the international research community should clarify these issues.

    The introduction of universal neonatal screening offers real hope of earlier diagnosis for children with hearing impairment. Creative uses of information technology will enable us to monitor its effectiveness and provide continuing measures of prevalence to better inform service providers.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    View Abstract