Reasons for investigating the aetiology of PHI in children include; providing parents with information, improving understanding of the progress and natural history of deafness, identification of other pathologies such as Pendred or Usher syndrome to enable timely and appropriate management strategies, genetic counselling and to help in understanding the epidemiology of hearing loss.³

National guidelines for the aetiological investigations of deafness in children have been issued by the NHSP.³ The Guidelines were first published in 2003, but have been revised over the years. Investigations are categorised as levels 1 (box 1) and 2 (box 2) depending on the evidence base. Level 1 investigations should be offered in all cases, level 2 in selected cases depending on the clinical picture.

There is little evidence on adherence to the guidelines across the country. Wilson et al⁴ studied the current practice of aetiological investigation of children with bilateral severe-to-profound sensorineural hearing loss (SNHL) in Wales by sending postal questionnaires to all ear, nose and throat (ENT) consultants and community paediatricians in audiology working in Wales. They found that community paediatricians reported higher usage of visual and developmental assessment, but less of imaging and ECG, whereas ENT consultants were less likely than other specialists to request ECG and urine analysis. They also found that community paediatricians performed more Level 2 investigations, mainly haematology, biochemistry and serology. Their study recommended greater use of ECG, urine analysis and ophthalmology referral.

Yoong and Spencer⁵ audited local performance against national guidelines for aetiological investigation of permanent childhood hearing impairment in Bradford. They found that almost 42% of the children were not offered imaging and about 20% were not referred to ophthalmology. They concluded that factors responsible for non-adherence to guidelines were lack of funding and parental choice.

Parker et al⁶ in 1999 reported a survey of clinical geneticists, in which they noted a great variation both in the services provided and in the recurrence risks quoted in isolated cases of childhood deafness. They sent postal questionnaires to 79 consultant clinical geneticists based at 26 centres across the UK. Just over half the respondents would ask for parental audiograms and only around 10% requested imaging. In their survey, the commonest specific investigations requested were ECG, ophthalmological review, thyroid function tests, virology for congenital infection and urine analysis. They acknowledged the need to improve existing clinical and social understanding of childhood hearing impairment.

The respondents were grouped into five different specialties on the basis of their response to the mandatory question regarding their specialty. There were 15 paediatricians, 11 paediatricians in audiology, 8 community paediatricians, 8 audiovestibular physicians and 10 ENT surgeons.

Of the nine items listed in the level 1 investigations category, all respondents except one ENT surgeon indicated they would routinely take a detailed paediatric and family history. This ENT surgeon commented that he/she would do a routine ENT examination, but the paediatrician in the joint clinic also routinely saw the child. Apart from one each of audiovestibular physician, community paediatrician, ENT surgeon and paediatrician, all others indicated they would do a clinical examination including developmental assessment routinely.

The level 1 investigations showing the most variations between specialties for being requested routinely were family audiograms, Connexin testing and MRI scan of inner ears, with MRI having the maximum variation across specialties. Table 1 summarises the frequency with which the level 1 investigations were requested routinely by different specialists.

Audiovestibular physicians and paediatricians in audiology routinely requested family audiograms and ophthalmology referral more often than others. ENT surgeons were the least likely to request the level 1 investigations routinely. This difference was most notable for requesting MRI scan (20% ENT surgeons and 100% audiovestibular physicians).

Overall, audiovestibular physicians and paediatricians in audiology were more likely to request level 1 investigations routinely as compared to other specialties. This difference was statistically significant (Fisher-Freeman-Halton exact test, p=0.008). The analysis was done by assigning the respondents to two groups, those requesting eight or more level 1 investigations routinely (first group), and those requesting seven or less routinely (second group). Of the total of 19 audiovestibular physicians and paediatricians in audiology, 14 were in the first group and 5 in the second. Of the 33 respondents from other specialties, only 11 were in first group, but 22 were in the second group. A contingency table analysis with χ² and Fisher-Freeman-Halton exact test showed this difference to be statistically significant, with a p value of 0.005 (χ²=7.864535) and 0.0089, respectively. The Fisher-Freeman-Halton exact test is more accurate as the numbers in the groups are relatively small.

Geographically, of the 20 respondents from North West and London, 10 made routine requests for level 1 investigations, whereas only 10 of the 32 respondents from the other regions did likewise. Analysis similar to that done for the specialties, showed this difference to also be statistically significant (χ² (x²)=9.437037, p=0.002). All the respondents (100%) from the North West and London reported requesting family audiograms routinely, compared to little more than one-half to two-third from other regions. The most striking difference was for the routine use of imaging, with none of the respondents from the South West requesting this routinely, compared to more than 80% from North West and London doing so (table 2).

The results from this study show a good awareness towards the importance of taking a detailed paediatric and family history and also carrying out a thorough examination, including developmental examination. There appeared to be major disparity between specialists in requesting family audiograms. One reason for this difference is a lack of resources in some parts of the country as indicated by comments from some of the respondents. One respondent stated that there was an ongoing negotiation and business case with the commissioners for family audiogram. Another respondent indicated that there were funding implications with family audiograms as only the patient would be paid for on payment by result (PBR). The other reason could be a lack of awareness of the importance of certain investigations such as family audiograms. Age-appropriate hearing assessment of first-degree relatives has been recommended even if there are no concerns, as unsuspected abnormalities may be uncovered and the configuration of the audiogram may also show a similar pattern among family members.²

The frequency of routinely testing for congenital cytomegalovirus (cCMV) appears to be less than optimal across all the specialist subgroups. CMV is the most common cause of congenital infections in humans¹³ and is a leading cause of non-hereditary SNHL.^14–15 Children with both symptomatic and asymptomatic congenital CMV can develop SNHL.¹⁶ The hearing loss can be progressive or of delayed onset.¹⁶ Not diagnosing cCMV infection can have important implications as parents may not be counselled about the chance of concurrent disabilities. Another important reason for diagnosing cCMV is that early antiviral therapy has been shown to prevent onset or deterioration of hearing loss in both symptomatic¹⁷ and asymptomatic ¹⁸ cCMV infections. Therefore, routine testing for cCMV is important. The diagnosis of cCMV can also be made retrospectively from the child's dried blood spot (DBS or Guthrie card). This has been shown to be a valid and effective method, with the added advantage that diagnosis can be made after many years as the Guthrie card can be stored for long periods.¹⁹

The most common cause of genetic deafness is mutations in the Gap Junction Beta 2 gene (GJB2), located on chromosome 13q and encoding the protein Connexin 26.^20–21 The 35 delG mutation is the commonest^22–23 and has been reported to account for more than 80% of the GJB2 mutations in the Caucasian population.²⁴ Homozygotes for the c.35delG mutation have been reported to have more significant hearing impairment than other genotypes.^{21 25} In this study, all audiovestibular physicians and community paediatricians and most of the paediatricians and paediatricians in audiology would offer Connexin testing routinely, but only half of ENT surgeons did so. Not diagnosing a case of genetic deafness such as GJB2 would mean that parents may not be counselled about the high risk of recurrence in future pregnancies.

The level 1 investigation which seemed to demonstrate the maximum variation according to specialty was MRI scan of Internal Auditory Meati (IAM). Whereas all audiovestibular physicians and three-quarters of paediatricians in audiology offered this routinely, only two-thirds of community paediatricians, a third of paediatricians and a fifth of ENT surgeons did so. This finding is quite different from that of Wilson et al,⁴ who found that community paediatricians requested imaging less often than ENT surgeons. Local policy and funding seems to be one of the factors influencing the decision of not offering routine imaging, especially from the community as three community paediatricians commented about the difficulty in arranging MRI scans. Decision of the individual clinician was another factor as five ENT surgeons commented that they would arrange for the MRI scan only if cochlear implantation was being considered. However, imaging has been shown to have a high diagnostic yield in the investigation of PHI in children²⁶ and MRI is a level 1 investigation. High-resolution MRI has the added advantage of screening the central nervous system (CNS), enabling other neuro-developmental abnormalities to be detected.²⁷ Characteristic findings on MRI can also lead to a retrospective diagnosis of congenital CMV.²⁸ A dilated vestibular aqueduct (DVA) is the commonest abnormality noted on imaging.² The hearing loss in DVA can be progressive or there may be sudden drops in hearing triggered by minor head trauma.²⁹ Identification of a DVA thus enables the clinician to give parents information on prognosis and to discuss important management strategies such as avoiding contact sports which have a risk of head trauma. DVA may also be associated with thyroid disease in Pendred syndrome and parents need to be counselled about this. Therefore, the use of MRI must not be limited to cases only where cochlear implantation is being considered, but should be offered to all children with permanent severe-to-profound hearing loss.

Most of the respondents routinely referred children with PHI for an ophthalmological assessment. This finding is in contrast to the finding of Wilson et al⁴ that a greater need for routine ophthalmological referral was needed. This may reflect a better understanding of the role and importance of various investigations, possibly contributed to by the NHSP ‘aetiological investigations’ courses over the years. The NHSP Quality Assurance (QA) team visits to local NHSP sites would also have helped to improve the understanding of the role of the various aetiological investigations. Documents such as Quality standards in vision care for deaf children and young people (2009) published jointly by the the National Deaf Children's Society (NDCS) and Sense, add weight to the importance of checking the visual status of children with PHI.³⁰

All audiovestibular physicians and most community paediatricians routinely requested urinalysis or urine dipsticks, but less than two-thirds of ENT surgeons did likewise. This finding was similar to that of Wilson et al,⁴ However, the importance of routine urine analysis in the investigation of PHI in the neonatal period can be debated. The average age of presentation with deafness in Alport syndrome is reported to be 11 years,³¹ a routine urinalysis in the neonate would not necessarily identify this condition and it would need to be repeated in mid-childhood, especially if no cause for the hearing loss has been found or if the hearing loss is progressive.

A bogus question; positron emission tomography (PET) was added to the list of investigations in the questionnaire to increase the validity of the responses. A PET scan is not listed in the investigations recommended in the national guidelines and currently does not form a part of the investigation for PHI in children. Therefore, if any of the respondents had indicated they would offer this routinely, the validity of the rest of their responses could be questioned. However, none of the 52 respondents indicated they would offer a PET scan routinely.

For the question ‘Would your answers have been the same for a child with Unilateral severe-to-profound deafness’; 59% answered ‘no’ and 41% ‘yes’. For the question ‘Would your answers have been the same for a child with mild-to-moderate degree of hearing loss’; 62% answered ‘no’ and 38% ‘yes’. There are now guidelines for investigation of unilateral and mild-to-moderate sensorineural hearing loss³², but these guidelines are not evidence based.

One of the limitations of the study was the response rate of 52%. However, this response rate could be considered quite acceptable as often the response rate to questionnaires without personal (face to face or telephone) contact with the respondent can be as low as 20%.³³ One reason for non-response could be due to the method by which the questionnaires had to be distributed. The questionnaires were sent electronically to all the clinical NHSP leads. However, the clinical lead may not be the person carrying out the investigation in all areas. Even though the covering letter accompanying the questionnaire requested the clinical lead to forward the questionnaire to the appropriate individual, this may not have happened in all cases, thereby reducing the response. Another factor could be that those who responded to the survey are the ones more likely to be carrying out the investigations as per the national guidelines and those who do not comply with the guidelines not responding. Therefore, the real proportion of families offered the full battery of level 1 investigations may be even lower than shown in this survey.

The main limitation of this study was that the MRC Hearing & Communication group, which sent out the questionnaires on behalf of the authors, could not divulge details about the number of recipients in each SHA or provide us with a breakdown of the specialty of all the recipients as they felt this would be a breach of confidentiality. It was not possible for the authors to find out this information as only the MRC Hearing & Communication group had the relevant database. The specialty of the responders was known as this was a mandatory question. Since the specialty of all the recipients was not known, it was difficult to know if any one specialty had a higher percentage of return rates than others, which might create a bias. However, since there was a fairly symmetrical mix of specialty among the responders, the chance of any bias towards any one particular specialty is unlikely. Similarly, the number of responses received from each SHA was known, but the exact number of recipients in each SHA was unknown. Therefore, it is not known if a higher percentage of recipients responded from any particular SHA. Again, as for the specialty of the recipients, since there was a good distribution of responses from the various SHAs, bias towards any particular SHA would not be likely.