June -November 2006 : NHS programs in Belgium


Universal Newborn Hearing Screening in Belgium

by Prof PJ Govaerts, the Eargroup, Antwerp-Deurne, Belgium.



Belgium is a west-European country with approx. 10.300.000 inhabitants.  It is a federal country with 3 semi-autonomous regions, Flanders (6.000.000 inhab), Wallonia (3.300.000 inhab) and the Capital Region of Brussels (1.000.000 inhab).  Medical prevention belongs to the regional responsibilities.  In Wallonia and Brussels no universal neonatal hearing screening programs exist.  Efforts are being made to implement them in the near future though.


History of UNHS in Flanders

In Flanders neonatal hearing screening was started by the Eargroup team in Antwerp under Dr. PJ Govaerts in 1993.  After a pilot phase with several modifications to the techniques being used, the regional Well Baby Clinics (“Kind & Gezin”) adopted the idea and launched a universal program in Flanders in 1998, which became fully operational in 1999.  In Flanders, the annual birth rate is approx 60.000.  Screening is done at the Well baby centres within the first 2 months after birth by means of automated ABR (ALGO®).  In case of unilateral or bilateral fail, a second test is done 1-2 weeks later.  In case of unilateral or bilateral fail on this test, the child is referred to one of a list of  “centres of reference”.  These centres are selected on the basis of geographical and political reasons.  No quality criteria apply.  The centres are bound to follow a diagnostic protocol as defined in a mutual contract between the Well Baby Organisation and the centre.  In case of confirmation of the hearing loss, the child is referred to an auditory rehab centre for hearing aid fitting and family guidance.


Outcome of UNHS in Flanders

In 2003 and 2004 the UNHS tested 90,7 – 93,5% of the newborns (ref: annual reports Kind & Gezin).  NICU children are not being tested!  An additional 3% of the newborns were tested in and by the maternities and approx 2% of children refused to undergo the screening.   The first test was done during the first month of life in approx 70% of the cases and during the second month in another 27-28% of cases.  After the first test, approx 3,5% were enrolled for a second test.  After the second test, approx 3 per thousand newborns remain to be referred for diagnostic work-up.  Approximately 1,0 to 1,2 per thousand are reported to have an ABR-confirmed bilateral hearing loss with thresholds above 40 dB.


What have we learnt?


  1. Do not re-discover the wheel !

When Karl White launched his UNHS program in Rhode Island in 1990, it became clear that the OAE-device that was developed by Kemp et al in the late 80’s would allow for the first time ever to really screen the hearing of all newborns with good test specificity and sensitivity and above all with an acceptable price. Soon after receiving the first reports from Karl White, we decided to invite him to Belgium to teach us how to implement UNHS.  He came over several times and his teaching was of invaluable importance to speed up things in Flanders.  We now see that many countries in Europe and abroad still hesitate to launch UNHS programs because they first want to repeat all the experiments and operational comparisons that were already done 10 years ago and that are well documented.  This loss of time is unjustified.


2.     Work at 2 levels

 We believe that for a UNHS program to succeed, two distinct levels have to be well defined: 1) the central (regional or national) level where the responsibilities for the epidemiology, the monitoring and the “trace and chase” function have to be situated and 2) the local (hospital or well-baby centre) level where the operational competences and responsibilities are situated.  Both levels are complementary and have different tasks.  The central level will never be able to execute the individual screening and to cover sufficiently high numbers.  The central level should keep the database, should define and monitor the quality control and should organize task forces to really trace and chase the babies that are at risk for being lost to follow up.  The local level should execute the screening, inform the parents, and stay motivated.


3.     Be sure to have a solid “after screen program”

In Flanders we are proud of having a UNHS since 1998, but we still experience major difficulties in the quality of the after-screen.  After a double fail on the screen, babies are referred for diagnostic work-up.  Already at the moment of referral, parents need expert counselling.  Only expert counselling will reassure them that everything is under control.  This is essential since parents at that moment have lost their hope and yet, after so many efforts, nobody tells them whether there is a hearing loss or not.  They have plenty of questions that need an answer at that very moment.  Otherwise, the parents will also loose their thrust and compliance.  The diagnostic workup should be done by competent people.  Too much is at stake to allow compromises of whatever kind.  In Flanders parents are referred to a limited list of “centres of excellence” but this list is constructed on political grounds without any qualitative criterion.  This is a disaster.  Objective criteria should be established that check not only for eligibility as centre of excellence, but that also enable continuous monitoring of the existing centres.  After a proper diagnostic workup, children with a hearing loss need family guidance, hearing aid fitting and (re-)habilitation.  It takes time for specialized rehab centres to gain experience in these very young kids, but also in this domain, excellence may exist, is required and should be monitored. Here again, we have learnt how difficult it is to establish criteria of good practise and we still have much work to do.


4.     Do not compromise on the standards

Fortunately many people who are involved in UNHS, are very committed and they really want to do the best.  But not many health care professionals are used to manage large quantities of subjects and data like in screening programs.  It is difficult to keep the same high quality standards that we are used to use in individual clinical practise when we are suddenly dealing with large population data.  Yet it is crucial not to compromise on these standards.  For instance, if epidemiology tells us that the prevalence of bilateral sensorineural hearing loss is 1,2 per thousand newborns, it is not correct to give your UNHS program good marks when it finds 1,2 children per thousand with a hearing loss, not mentioning that also unilateral losses and conductive losses were included in its figures.


5.     Prodigious children have many fathers

The fate of pioneers is to face much opposition as long as their project is immature and the outcome not yet proven.  Once their child appears to be prodigious, many are those who claim paternity J.



In conclusion, Flanders is the first region outside the USA with 60.000 newborns a year and with an operational UNHS program.  The coverage is over 90% and thanks to some local initiatives, it is around 95%.  The screen specificity seems to be high (99,8% with 0,2% false alarms).  The sensitivity is not really known because no robust epidemiological figures are known about the prevalence of congenital hearing loss.  Assuming a prevalence of 0,12%, the sensitivity is between 83 and 100%.  Assuming a prevalence of 2% however, which has been found in well-controlled maternity based studies, the sensitivity is between 50 and 60%.   The use of automated ABR is arguable mainly because of the high cost of the disposables.  The selection of centres of reference on geographical and political grounds rather than on qualitative criteria is even more problematic.  The reported hearing losses are not well documented and the type of deafness (sensorineural or conductive) is not reported.  The lack of transparency of the Well Baby Organisation renders scientific analysis of these data cumbersome.



JJJ Dirckx, K Daemers, K Van Driessche, Th Somers, FE Offeciers, PJ Govaerts.  Numerical assessment of TOAE-screening results: currently used criteria and their effect on TOAE prevalence figures.  Acta Otolaryngol (Stockh) 1996; 116: 672-9.

K Daemers, J Dirckx, K Van Driessche, Th Somers, FE Offeciers, PJ Govaerts.  Neonatal hearing screening with otoacoustic emissions: an evaluation. Acta ORL Belg 1996; 50: 203-9.

G De Ceulaer, K Daemers, K Van Driessche, S Mariën, Th Somers, FE Offeciers, PJ Govaerts.  Neonatal Hearing Screening with transient evoked otoacoustic emissions: a learning curve.  Audiology 1999; 38: 296-302.

PJ Govaerts, G De Ceulaer, M Yperman, K Daemers, K Van Driessche, Th Somers, FE Offeciers.  A two-stage, bipodal screening model for universal neonatal hearing screening.  Otol Neurotol. 2001;22(6):850-4.

G De Ceulaer, K Daemers, K Van Driessche, M Yperman, PJ Govaerts.  Neonatal hearing screening with transient evoked otoacosutic emissions - retrospective analysis on performance parameters.  Scand Audiol 2001; 30 Suppl 52: 109-11.

K Deben, S Janssens de Varebeke, T Cox, P Van de Heyning.  Epidemiology of hearing impairment at three Flemish Institutes for Deaf and Speech Defective Children.  Int J Pediatr Otorhinolaryngol. 2003;67(9):969-75.

I Courtmans, V Mancilla, C Ligny, B Belhadi, E Damis, P Mahillon P.  Hearing screening of newborns. Preliminary results. Rev Med Brux. 2005;26(1):11-6. French