Criteria versus guidelines: Are we doing the best for our paediatric patients?

Abstract

The National Institute for Health and Care Excellence guidance for the provision of cochlear implants (NICE Technology Appraisal Guidance 166. Cochlear implants for children and adults with severe to profound deafness. 2009. National Health Service National Institute for Health and Clinical Excellence.) are used to develop candidacy criteria by public health funding bodies within the UK. Often the guidance is interpreted as strict ‘criteria’ whereby clinicians adhere to specific audiometric thresholds without accounting for the acceptable range of performance on individual tests or a child's functional development. In this paper four clinical paediatric case studies are described from two cochlear implant centres which serve to illustrate difficulties in applying NICE guidance as strict criteria. These are presented in the context of recommending more flexible interpretation based on the content of the current guidance along with considerations of circumstances where NICE guidance might be adapted to optimise use of cochlear implant technology within a national framework.

keywords:

• Cochlear implant
• Borderline
• Reverse slope
• NICE guidance
• Non-traditional
• Candidacy
• Criteria

Introduction

The National Institute for Health and Care Excellence guidelines (NICE, 2009) recommend simultaneous bilateral cochlear implants (CI) for children who have a severe to profound deafness, defined as only hearing sounds that are louder than 90 dB HL at frequencies of 2 and 4 kHz without acoustic hearing aids, as part of a multi-disciplinary assessment including speech, language, and functional listening skills.

Interpretation of the guidance to create criteria based on discrete audiometric thresholds alone does not account for the acceptable range of performance for specific test measures. There is a recognised critical difference (i.e. test–retest variability) between 5 and 10 dB for a given threshold when performing pure tone audiometry (Schmuziger et al., 2004; Stuart et al., 1991). This means that a strict > 90 dB HL criteria does not allow for the expected variation of pure tone audiometry when tested on two different occasions, making an absolute criterion potentially less sensitive at identifying those who can reliably hear sounds only when they are louder than 90 dB HL.

In addition, criteria based on audiometric thresholds alone do not account for a child's functional performance; the audibility of the speech signal must be considered. Poor discrimination of speech compromises outcomes even if detection levels appear adequate (Stiles et al., 2012). Use of the Auditory Speech Sound Evaluation (ASSE) discrimination test (Govaerts et al., 2006) enables clinicians to consider how audible the speech signal is for a child. Govaerts recommends cochlear implantation if a child cannot discriminate several of the sound pairs. Following implantation, children should be able to discriminate 7/7 sound pairs. Stiles suggested the use of an aided audibility measure based on the speech intelligibility index (SII). The SII is a measure of the proportion of speech signal information which is audible to the hearing-impaired listener with their hearing aids. Aided SII results of children with hearing aids have been shown to correlate with outcomes in language development and speech understanding whereas language outcomes cannot be predicted efficiently from a child's hearing threshold levels. For children with mild to moderate hearing loss, Stiles et al. (2012) found that those with an SII of less than 0.65 were at greater risk of delays in vocabulary development. Therefore, relative to the pure tone average audiogram, the aided SII provides a more valid estimate of the child's access to speech and potential benefit from their hearing aid.

Ascertaining the benefit from paediatric hearing aid fittings in a timely manner can be challenging, as babies and young children do not give reliable feedback and may not be able to carry out speech testing which would clinically inform on the adequacy of the hearing aid fitting (Bagatto et al., 2010). Tests of infant speech discrimination and speech and language assessments have an important role in inferring future performance with acoustic amplification. NICE guidance recognises the relevance of functional assessment when assessing the suitability of CI for children (sections 4.3.2 and 4.3.3). Therefore, some flexibility in interpretation of the guidance would take into account test–retest variability, and enable other factors to be considered; e.g. aided SII, functional aided testing, speech and language assessments, and essential observations from the multidisciplinary team.

In the UK, clinicians often refer to children who have a severe loss in the high frequencies, for example, 80–85 dB HL at 2 and/or 4 kHz, as audiologically borderline for CI. These children may be described as ‘out of criteria’ for CI based on their hearing thresholds alone. However, more evidence is emerging to support extending the option of CI to children who are currently classed in the UK as ‘borderline’ (Carlson et al., 2015; Gratacap et al., 2015; Lovett et al., 2015).

Reverse-slope hearing loss is not widely reported in research; there is some anecdotal evidence of benefit from CI in a very small number of patients (Bauman, 2015). However there are a number of studies that have shown a high likelihood of cochlear dead regions in hearing losses like this. Vinay (2007) looked at almost 600 ears and found cochlear dead regions in nearly 60% of subjects when thresholds were in excess of 70 dB HL. That means these patients could be optimally fitted with hearing aids, but there is limited benefit as there are no primary auditory neurons to transmit that information to the central auditory system. CIs overcome this issue by bypassing the primary auditory neurons in order to stimulate the hearing nerve directly.

Four clinical paediatric case studies from two CI departments were selected to illustrate difficulties in applying NICE guidance as strict criteria. These include an audiologically borderline case for CI, one patient with a reverse-slope hearing loss, and two cases who fit the standard audiometric criteria but for whom CI was not recommended despite being ‘in criteria’. Routine practice in these departments is to assess and make clinical recommendations for treatment, regardless of the potential outcome of commissioning discussions.

Method

Cases were selected through a database search of patients who have pre-implant data reflective of most recent clinical practice. However, some cases are rare (e.g. reverse-slope hearing loss) and therefore the database search spanned patients assessed and implanted between 2010 and 2015. Information provided by the whole multi-disciplinary team, including the child (where appropriate), family, and local support team, was collected and collated for this review. For the purposes of this paper, children who have audiometric thresholds at 90 dB HL or greater at 2 and 4 kHz were categorised as being ‘in criteria’. Children whose audiometric thresholds are better than 90 dB HL at 2 and 4 kHz were categorised as being ‘out of criteria’.

Results

A summary of the four cases is shown in Table 1.

Table 1 Summary of cases presented. CI was recommended for patients A and B. CI was not recommended for patients C and D

	Type of hearing loss	Aetiology	Hearing thresholds at 2 and 4KHz (dBHL)	Hearing aid use established?	Speech perception results	Speech intelligibility index	Functional presentation	Developmental concerns/co-morbidities?	In ‘Criteria’ for CI?	Fits NICE guidelines for CI?	Recommend implantation?	Why?
Patient A	Borderline progressive severe to profound	Unknown but possibly genetic (Strong family history. Consanguineous marriage)	R 95/85 L 95/90	Yes	3/7 ASSE	R 0.31 L 0.25	Severe language delay with the gap between chronological age and language age equivalent widening	No	No	Yes	Yes – simultaneous bilateral	Poor speech perception and severe receptive and expressive language delay without improvement over time
Patient B	Reverse slope severe to mild, progressive	Strong family history of hearing loss. Clinical presentation of Waardenberg Syndrome although unconfirmed by genetic tests	R 70/60 L 70/60	Yes	R 5/7 ASSE L 4/7 ASSE	Not tested	Significant language delay with the gap between chronological age and language age equivalent widening	No	No	Yes	Yes – unilateral (L ear)	Poor speech perception and significant receptive and expressive language delay without improvement over time
Patient C	Traditional – moderate to profound	Bacterial meningitis contracted as a neonate	R 95/90 L 100/100	Yes	22/22 ASSE AST = +7.3 dB SRT	Not tested	Above average language and no academic concerns	No	Yes	No	No	Excellent functional performance despite poor understanding of speech in noise
Patient D	Traditional – profound	Meningitis contracted at 6 months	R >120/>120 L >120/>120	No	Unable to test	Not tested	No spoken language. Significant delay in sign language	No	Yes	No	No	No hearing aid use, duration of deafness, no spoken language, family's unrealistic expectations

Audiologically borderline patients

Patient A

Patient A was referred at the age of 2.7 years. She passed the Newborn Hearing Screen when administered just after birth, but did not develop spoken language. She was diagnosed with a severe-to-profound hearing loss at the age of 2.6 years, which deteriorated over the course of the assessment. The cause of deafness was thought to be genetic, due to a strong family history, although this remains unconfirmed. Patient A's thresholds at 2 and 4 kHz were 95 and 85 dB HL for the right ear and 95 and 90 dB HL for the left ear. She struggled with basic sound discrimination on the ASSE, being only able to discriminate 3/7 sound pairs. The SII was very poor for both ears; the right ear had a SII 0.31, and the left had a SII of 0.25. Patient A was unable to participate in the Automated McCormick Toy Discrimination Test because her language was not advanced enough. Results of the Pre-School Language Scale (PLS-4) suggested that Patient A had a severe expressive and receptive delay at initial speech and language assessment and made only one month's progress during the 10-month period before the follow-up assessment took place. Whilst Patient A was out of criteria audiologically, she was still within NICE guidance due to her poor functional performance.

Patient A was fitted with bilateral CIs. At the time of writing, she has had just over 3 months use and optimal listening programmes in her sound processors are still being established. Data logging and parental report suggests she wears both processors most waking hours. She has a developing repertoire of single nouns and rehearsed phrases and has developed a short-term auditory memory for two items with familiar vocabulary. Formal language assessments and speech perception measures are due at the 6-month interval.

Reverse slope hearing loss

Patient B

Patient B was referred at 5.6 years of age, due to concerns about his language development. He passed the Newborn Hearing Screen but a hearing loss was detected at 3 years old. There was a strong family history of hearing loss which had the clinical presentation of Waardenburg but genetic testing had not confirmed this. Patient B had a progressive bilateral severe to mild (250 Hz–8 kHz) reverse-slope sensori-neural hearing loss with thresholds at 2 and 4 kHz of 70 and 60 dB HL in both the right and left ears. He was unable to discriminate all sound pairs on the ASSE (4/7 left ear and 5/7 right ear). SII was not used in the clinic at the time Patient B was assessed.

Language development was monitored over time by the local speech and language therapist and the CI team. PLS-4 results suggested a 1 year 11 month overall language delay. Standard scores were 78 for receptive language and 55 for both expressive language and combined language results (the normal range for standard scores on the PLS-4 is 85–115). Additionally, results from the British Picture Vocabulary Scale (BPVS) demonstrated a receptive delay of at least 2 years. Repeat testing reported by the local speech and language therapist indicated language progress was not being maintained month on month and so his language gap was increasing.

The CI team along with Patient B's parents and the local support team concluded that the complexity and speed of language presented in both educational and social situations was already proving extremely challenging and he was at significant risk of falling further behind his peers because of compromised access to auditory information. Whilst Patient B was audiologically out of criteria, a unilateral CI to be worn with the contralateral hearing aid was recommended based on his poor functional performance and poor predicted progress.

Following a successful funding application, Patient B was fitted with a left sided CI. His progress was assessed at 1 and 2 years post-implant:

• The Brief Assessment of Parental Perception (BAPP) is a parent questionnaire developed by the UK Implant Centre Psychology Group and aims to gather information related to device use and quality of life. It is administered at one-year post-implant. Patient B's parents reported full time use of the device and improvements in four out of five quality of life measures, rating behaviour, contentment, communication, and learning as ‘much better’, with no change reported for getting along with friends.

• The Clinical Evaluation of Language Fundamentals (CELF UK 4) was used post-implant instead of the PLS-4, because Patient B was by then too old for this test. A scaled score of 7–13 on each CELF subtest is the normal range. Testing suggested the language gap was closing for receptive language, with a 12-month delay and scaled score of 6. For expressive language there remained a significant delay of 1.5 to over 3 years, with scaled scores ranging from 2 to 6, depending on the subtest used. However at two years post-implant, when expressive subtests were repeated, the gap had closed. The delay now ranged from 2 years 10 months to above average for his age. Scaled scores ranged from 6 to 13, with two of the three subtests now within the normal range.

Patient B prefers his CI to his hearing aid and by two years post-implant had started to reject the aid in his non-implanted ear. Further assessment from the team supported the view that a second implant may be beneficial for listening to speech in background noise and for sound localisation, with test results demonstrating marginal improvements without the contralateral hearing aid. However, the process of obtaining funding between the first CI and applying for the second CI had changed and funding was declined.

‘Traditional’ candidates

Patient C

Patient C was 13 years old when referred with a moderate to profound sensori-neural hearing loss. He was born at 28 weeks and was in special care for 9 weeks. His hearing loss was thought to be acquired following bacterial meningitis which he contracted during this period. His thresholds at 2 and 4 kHz were 95 and 90 dB HL for the right ear and 100 dB HL at both 2 and 4 kHz for the left ear. SII was not used in the clinic at the time Patient C was assessed. He was able to discriminate 22/22 sound pairs on the ASSE but struggled in background noise on the Automated Sentence Test (AST) achieving a speech reception threshold (SRT) of + 7.3 dB (The SRT represents the signal to noise ratio at which the child can understand speech in noise. For Patient C the signal needs to be 7.3 dB above the noise for him to discriminate speech at the level prescribed by the AST).

However, Patient C had made remarkable progress with spoken language development. His scores on the BPVS were reported to be above average for his age and his speech was clear. There were no concerns about academic achievement. Although Patient C's hearing thresholds were in criteria, the assessment suggested that functional performance with hearing aids was excellent. The team concluded that CI might improve his listening in noise but following discussion with the child and his family, the decision was to continue with his hearing aids.

Patient D

Patient D was referred at 8.9 years old. She presented with bilateral profound hearing loss with no response at the maximum output of the audiometer at 2 and 4 kHz. Her hearing loss was thought to be the result of meningitis contracted at 6 months. A scan showed no ossification. Hearing aids had hardly ever been worn. She had no spoken language and significantly delayed sign development. There were no aided auditory responses to sound, and the family's motivation for CI was unrealistic. Despite all this, Patient D was in criteria, however, following multi-disciplinary assessment, the conclusion was that CI was not in Patient D's best interest; it was more important to support her interaction, communication, social and emotional development, as well as her academic achievement in a School for the Deaf.

Discussion

An approach of establishing candidacy criteria solely on the basis of audiometric thresholds does not account for the limited test–retest reliability of audiological measurements or a child's functional performance. This would appear to contradict NICE guidance. Section 4.3.2 states ‘The Committee concluded that decisions about the appropriateness of CI should take into consideration a person's functional hearing and the benefit they gain from acoustic hearing aids.’ Additionally, section 4.3.3 states ‘[…] The Committee heard that tests for children should assess whether speech, language and listening skills are appropriate to the age, development stage and cognitive ability of the child.’ One could therefore argue that NICE guidelines allow for a certain level of flexibility which permit other relevant clinical findings to be considered, such as those described in the four case studies in this paper.

Public health funding bodies in the UK have an extremely difficult task when making decisions about how to interpret NICE guidance to form criteria. Further clinical evidence is required to give funding bodies confidence to recognise more flexible criteria which remain within the guidance. For children, this evidence could be provided through the use of: (a) clinical tools that provide information about a child's ability to discriminate between sounds rather than just detect the presence of sound (e.g. ASSE); (b) clinical tools that provide an indication of the audibility of sound (e.g. aided SII); and (c) age-appropriate language assessments. Increased confidence in clinical decision making would allow children who are currently out of audiological criteria to be fitted with CIs where a multi-disciplinary assessment has concluded that a child's progress is being limited through inadequate amplification.

Conclusion

Clinical recommendations should be based on thorough, appropriately experienced, multi-disciplinary assessment, where the best interests of the child are at the centre of decision making. Further research is recommended to explore the advantages of CI for children who are unable to reach their potential with hearing aids.

Disclaimer statements

Contributors None.

Funding None.

Conflicts of interest The authors declare no conflicts of interest.

Ethics approval None.