Navigating the Shoals of Surgical Treatment Decisions in OME

Some highlights including new findings as summarised in a presentation

on the TARGET Trial to BAPO, September 2009

 Mark Haggard, Multi-centre Otitis Media Study Group

This is not another guideline, as there is a good one (NICE GL60, 2008) already in place, itself consistent in main thrust with a previous one provided by Peter Robb and myself for BAPO.  The present set of 8 points concentrates on summaries and principles, to enable the  paediatric otolaryngologist to appreciate what the evidence implies, and to reconcile some of the conflicts between claims of different  articles, and between these and the understanding of pathogenesis or clinical experience. It represents a synoptic view taken in 2010 and will require to be updated, but has been checked by BAPO Council as both reasonable and helpful at this time.  It is likely that in the eventual revision of GL60, NICE will be able to incorporate formally evidence, currently being published, that it could not in 2008; some use is made here of such evidence. Guidelines sometimes seek to reduce clinical discretion, but mostly where the evidence on what works best is very clear, or  g for safety or side-effect reasons. Where there are not perverse incentives towards over-intervention (eg payment b y item-of-service,  not present in UK), clinical discretion (guided by principle and check list, rather than whim) is to be encouraged, because spurious over-precision is also a danger.

1 Reasons for evidence gaps, and limitations to overcome

There is quite a lot of evidence on surgery in OME, some of it of good quality. It does not say exactly what to recommend for the individual presenting child, and there are limits to how closely evidence that is achievable can approach that. The condition is multi-facetted and changes over time, making it hard to research, and it is expensive to document a history in sufficient detail.  OME has been the subject of a narrow view in which mere current diagnosis, rather than assessment of severity and persistence, permitted treatment. In contrast, several findings now confirm the importance of past history as a predictor of impact and hence of the ability to benefit. Such information is limited by (a) lack of general record linkage from community to ENT, (b) fluctuation of the condition in respect of the objective and performance measures that can be taken at ENT, and (c) differing standards for response and imperfect memory in parental response. Procedures that overcome these limits will be the next priority.

2 The present pressures and dilemmas

Pressures to reduce numbers of operations may increase in the present financial climate. In England, many PCTs now have a policy of not funding ventilation tube (grommet) operations unless the NICE criterion of a 25dB HL is met. They need to have and be seen to have some form of restrictive policy in place, but seem insufficiently aware of the changes in practice since over- treatment was first rightly criticised in 1985. We now have an accumulation of evidence justifying using HL as a main basis of decision, which supports the NICE recommendation at a general level of using HL, something which a meta-analysis of trials could not provide in 2005. NICE never intended that this should be black and white; indeed, imagining the 1.25 dB difference between two possible cases with average HL of 24.375 dB and one of 25.625 dB (with 2 ears and 4 frequencies, 5dB steps give 5/8 = 0.625dB as the minimum difference, so this is twice that minimum) to be real and not affected by other considerations is absurd. Rather there should be audit of numbers of operations below 25 dBHL and a check against the NICE list of particular supplementary considerations  (eg manifest developmental educational impact) invoked for cases below 25 dB, and preferably for those up to 30dB also. An HL of 26.625 dB cannot make these irrelevant. Some commissioners� policies have omitted these considerations in their transcription of NICE. This may just be due to limited understanding of these issues. I have for some years advised consultants encountering stances from commissioners not driven by balanced considerations, eg the absurd idea of 6 rather than 3 months of formal watchful waiting (WW), without specification of the type of child, if any, to whom this might be applied.  Such ideas show little awareness of the natural history or the existing gatekeeping in primary care, based on informal watchful waiting, or of changes over the last 25 years. Positive engagement is required with commissioners to implementing the balanced spirit of NICE overall, rather than selectively follow its more restrictive parts only. 

3  Differences between caseloads

There are large differences between a typical NHS caseload and the samples in two trials with null results (Rovers, NL, 2003; Paradise, US, 2005). Those trials are of high quality, but high-quality interpretation and application of them are  difficult to guarantee.  Due to the children being found by active screening from around 1 year, they were younger and their impact milder than in those that come through primary care gatekeeping in the NHS. There are analogous differences in severity between UK regions, that appear to correlate with the local intervention rate, and which certainly underpin variation in the degree of benefit from operations; the more selectivity in, or resistance to, local  referral to ENT, the more benefit.  Local practice variations (�postcode prescription�) can not be eliminated completely. But large variation cannot be justified epidemiologically, and is counter to both effectiveness and equity. Discussion of criteria, operation numbers and audit between commissioners and consultants should be based on an understanding of whether the local intervention rate has been high, medium, or low, relative to national norm for the child population, and attempts should be made to understand the likely reasons for any deviation.

4 De-medicalisation?

Current attempts to contain spread of antibiotic resistant bacteria have had some success in UK, by unhooking parents from the incorrect belief that in RAOM antibiotics are always a necessary or highly effective treatment. Unfortunately, the pressure on services leads to some flight of patients whose expectations are not met to more available but unconventional forms of care, and this supports another bad side-effect: over-reliance on placebo effects in anti-scientific systems like homeopathy, supporting erroneous beliefs. As with adult self-management (cf hearing tactics), paediatric ENT and audiology have a duty to foster understanding and behavioural management of the impact of OME as a supplement or alternative to any operation given, and this can help fill the gap. The need for traditional ENT and audiology services is so large that this is no threat to them.

5 Benefits to hearing

Short-term benefits to hearing from VTs are not in question, if persistence through 3 months has been established by WW and a better-ear hearing of 20dBHL or more (average 24 dB). The magnitude of the difference between treated and untreated groups becomes washed out over time. Despite difficulty in showing proportionality in standard analyses, the amount of such benefit is proportional to the initial hearing loss, justifying the main stance of NICE on an HL criterion. The point is rather that any one such basis of selection cannot be complete. Early parental report of hearing after treatment is exaggerated by contrast and by expectancy bias (a form of placebo effect) and so is over-generalised, but the short-term HL benefit from ventilation tubes is genuine and, in relation to the prognosis if untreated,  is large in magnitude. When accumulated over two years, the lesser but more enduring further benefit of 3-4dB on average from adjuvant adenoidectomy roughly doubles the accumulated hearing benefit for the period. It is difficult to translate this finding into a policy consensus, because various people place differing emphases on the sort and long terms of this finding. Health-economically, the associated reductions by adenoidectomy in further ENT care required make adjuvant adenoidectomy attractive, and  further  squeeze the scope for VTs-alone, between continued WW and VTs+ad.   I have never adopted the purely health-economic stance on this (ie have not recommended adenoidectomy for all), but have sought to move towards co-ordinated criteria for the two procedures that would concentrate and so maximise the benefit from each, within a broader clinical understanding.

6 Nature and location of benefits to general development

Language problems can occur or be exacerbated in persistent OME, but in relation to the mostly transient effects in the usual range of severity and persistence, this problem has been over-generalised and hence over-sold in the past. It is not surprising that the two trials mentioned, whilst finding benefits to hearing, do not find demonstrable benefits to language. TARGET does find benefit to a broader parent-reported measure of development including behaviour and parent quality of life (ie on the well established cases qualifying with and having come through NHS referral). This overall benefit is small, and due mostly to ventilation tubes, but supplemented by adjuvant adenoidectomy. It divides into a moderate benefit to the older (here > 5 yrs) children in the caseload and virtually none in the younger. The same age-dependence is found in a performance test of auditory disability, speech reception threshold in noise, so the principle is trustworthy. However, to interpret �age� here requires recalling that children with prior ENT operations were excluded from TARGET. Thus, what it really shows is that, understandably, the older children who are new patients (and for whom the identification and referral process has been slower than average, as histories rarely start at age 4-5) get more benefit to development. Such children have more cumulative deficit that is remediable by improving the hearing, because of their longer history and circumstances in the family or healthcare system, ie not bringing them through the primary care gatekeeping earlier. To follow the recommendation of NICE and assess for the impact on the child, it is necessary to consider age at presentation in relation to family tendency to seek help and the degree of gate-keeping likely to have been experienced. This, and the considerations in 5 & 7 could all be put into a formal and quantitative guideline, but it is not clear how practical and acceptable using that would be.

7 Physical health in OME

Because a diagnostic contrast with AOM is needed, and because OME is sometimes only semi-symptomatic, in younger cases the physical impact has not been emphasised in recent decades. However not all OME cases are those with such immunologically abnormal mucosa as to be triggered by a single infection into chronic viscous effusion � the classical glue ear. Thus we must be dealing with a continuum of which one end also carries a heavy burden of URTI and/or �super-added� (R)AOM. The ordering here is based on the main pathogenetic sequence, and the issue is not reallocating the diagnosis; some of the cases that qualify on two occasions as OME by B tympanograms and  a measured HL of >/=20 dB better ear have this major recurrent physical health burden also . TARGET underlines this dimension, which is not a distinct dichotomy. Even when the URTI and RAOM scores are adjusted for expectancy bias and so rendered semi-objective, there is a moderate-to-large overall benefit respectively to these scores from adjuvant adenoidectomy (URTI) and from basic ventilation tubes (super-added RAOM). Furthermore, and important for therapeutic rationale, these benefits are strongly related to how severe the specific problem (URTI, RAOM) was in the first place. Adenoid-ectomy has other benefits in OME, but the restriction by NICE of the indications to those with severe/frequent URTI was a judicious deferment of the issue, and remains largely consistent with the emerging evidence.

8 Recommendations of a more specific nature, and clinical tools to support and monitor them

It is unlikely that any satisfactory basis for a cost-per-QALY policy decision on these two treatments in OME will emerge in the near future, because of three uncertainties over: (a) the appropriate balance between the above three domains of outcome; (b) the appropriate balance between short and long terms; and (c) the appropriate balance between benefits and risks/inconvenience with adenoid-ectomy. Nevertheless evidence is emerging for OME impact proportional to disease severity and duration, and for its reduction by treatment, which should prompt the otolaryngologist to select cases well. Forthcoming publications will express (5-7) as recommendations, with cut-offs on simply acquirable clinical measures, plus question responses on characteristics that predict degree of benefit.  Additionally, as the ideal background information in (3)may not be available, ENT departments may be asked by commissioners to conduct audits of caseload severity, case selection  and outcomes. A support package to assist running such audits (OM2-13 questionnaire and spreadsheet, with facilities for data entry, scoring, storage of standard clinical data, transmission and simple analysis) will shortly be available.