B4 School Check hearing screening and middle ear disease: a five-year analysis of prevalence and inequity

The B4 School Check (B4SC) is a national screening programme in Aotearoa New Zealand that identifies health and developmental issues at age four, before children begin school. Eligible children are identified via primary care, school and preschool enrolment data, with families also able to self-refer.¹ The programme is overseen by Health New Zealand – Te Whatu Ora across its regional branches, replacing the former 20 district health boards.

Detection of hearing loss is an integral component of early childhood screening, as timely management of ear disease supports optimal speech, language and learning outcomes.² In Aotearoa New Zealand hearing is formally screened at birth and again during the B4SC.³ While newborn hearing results are routinely reported, data for the four-year screen are not.⁴ Screening occurs in early childhood education settings or community clinics, and all children are offered this opportunity except those already managed by audiologists or otolaryngologists.

Children undergo pure-tone audiometry screening in each ear. Failure to respond to any screening tone triggers tympanometry testing, which assesses middle ear status as peaked (normal) or flat (abnormal, otherwise known as type B).⁵ Type B tympanograms most often indicate otitis media with effusion (OME), but can also occur in the setting of wax impaction, a perforated drum or the presence of a ventilation tube. OME is a collection of middle ear fluid that may follow respiratory infection and can persist, diminishing hearing in home and classroom environments. Left undetected, chronic OME can delay speech, language and literacy development.²

Under B4SC protocols, children with abnormal tympanometry results are either rescreened after 3 months or referred to a primary care provider. This depends on the degree of hearing loss but is also based on the assessment of the technician conducting the screen as to whether the child is at high risk for developmental and learning difficulties. Further interventions beyond referral and their timing are not captured in B4SC data.

International guidelines recommend ventilation tube insertion for OME lasting beyond 3 months, especially if there are developmental concerns.^6,7 In Aotearoa New Zealand, higher rates of persistent middle ear disease are experienced by Māori and Pacific children and those living with higher deprivation.^8–11 Further, there is evidence to suggest that OME can have a more significant impact on long term outcomes for children living in more deprived areas and that these impacts are not spread evenly across our specific populations.^12,13 These groups also have higher rates of hospitalisation for the related illness acute otitis media, yet have lower and later rates of elective grommet insertion.^14,15

Access to both screening and follow-up services is not universal. Barriers include distance to clinics, appointment costs and primary health organisation enrolment status.^16,17 Although proposals for more targeted screening exist,¹⁸ there is a lack of up-to-date, nationwide prevalence data for preschool OME which makes the design of such programmes difficult.

By linking national B4SC tympanometry and hearing results with individual level demographic data we can describe the true burden and distribution of OME and identify inequities in access and treatment. These insights can inform future screening strategies, resource allocation and service delivery models to deliver more equitable ear health outcomes for Aotearoa New Zealand children.

Methods

All B4SC individual hearing screening records from 1 January 2018 to 31 December 2022 were extracted, capturing pure‐tone audiometry and tympanometry results for each ear (coded in the database as “Pass” or “Fail” for each ear and each metric without further detail), overall screening outcome and whether a child had been rescreened or was a direct referral after their first screen. These records were linked to individual level demographic data: ethnicity (Māori, Pacific, Other), gender and primary health organisation (PHO) enrolment status. We also collated each child’s Health New Zealand – Te Whatu Ora region, rurality (as defined by the Health Services Accessibility Index [HAI] score¹⁷) and data required to grade each individual within the Index of Multiple Deprivation (IMD)¹⁹.

To estimate the cohort either not identified or not successfully engaged by the screening programme, we identified children with any health services user activity recorded by Health New Zealand – Te Whatu Ora in their 4th year (1 January 2018 to 31 December 2021) who lacked a B4SC record, applying the same demographic linkages as above.

All data were de-identified and analysed in SPSS within The University of Auckland’s secure data laboratory.

We compared demographic differences among screened versus unscreened cohorts using Chi-squared tests (α=0.05). Among screened children, we calculated the prevalence of flat (type B) tympanograms as a marker for OME. Sub-group Chi-squared tests and multivariable logistic regression models examined associations between flat tympanometry and demographic factors. In the subset with flat tympanograms, we evaluated whether children had been rescreened or were referred directly to primary care after their first encounter with the programme, stratified by ethnicity and deprivation. A final sub-group analysis assessed PHO enrolment disparities among those with flat tympanograms.

Ethics approval was granted by The University of Auckland Health Research Ethics Committee (AHREC ref AH27066).

Results

Screening uptake and tympanometry outcomes

Between 1 January 2018 and 31 December 2022 276,911 children were identified in the B4SC database. Of these, 261,986 (94.6%) underwent hearing screening and 13,507 (5.2% of those screened) had flat (type B) tympanograms. Among those with flat results, 7,774 (57.6%) were referred directly to primary care and 5,394 (39.9%) had undergone rescreening prior to referral; 286 (2.1%) were already under specialist care and 53 (0.3%) lacked outcome data.

Children not screened

We identified 38,035 four-year-olds (2018–2021) with health services user data but who were absent from B4SC records in the year they would have been eligible for screening. Māori (OR 0.58, p<0.0001) and Pacific (OR 0.49, p<0.0001) children were significantly less likely to be listed in the B4SC database. Higher deprivation was also associated with reduced inclusion: odds of screening declined progressively from IMD decile two through decile 10 relative to decile one (see Figure 1). Multivariable regression analysis confirmed that both ethnicity and IMD decile were independent predictors of unsuccessful identification and engagement by the screening programme.

View Figure 1–3.

Prevalence of flat tympanograms

Overall, 5.2% of screened children had flat tympanograms. Prevalence by ethnicity was 6.9% in Māori (OR 1.86, p<0.0001), 10.0% in Pacific (OR 2.79, p<0.0001) and 3.9% in other ethnicities. Analysis of IMD showed no difference across deciles one to five, but a significant increase in flat tympanograms from decile six onwards (see Figure 2). Rurality (HAI decile) had a minimal effect, with no significant difference between different deciles. Multivariable regression identified ethnicity and IMD as independent predictors of flat tympanometry.

Primary health organisation enrolment

Of the children with flat tympanograms, 741 (5.5%) were not enrolled in a PHO. Māori (OR 0.54, p<0.0001) and Pacific (OR 0.51, p<0.0001) children had significantly lower enrolment rates. Children in IMD deciles seven to 10 also showed reduced enrolment compared with decile one (p<0.05).

Rescreening versus referral

Māori (OR 0.87, p=0.0005) and Pacific (OR 0.81, p<0.0001) children with flat tympanograms were more likely to have undergone rescreening than be referred after their first screen (see Figure 3). Higher deprivation was similarly associated with rescreening; HAI showed no effect. Multivariable regression indicated ethnicity and deprivation effects on rescreening rates were not independent variables.

Discussion

There were inequities identified in our dataset in the B4SC screening process, as well as in underlying disease burden and the subsequent opportunity for management. An estimated 38,000 children over the study years 2018–2021 were potentially eligible for screening based on health services user activity but were not in the B4SC database; about half of these were Māori or Pacific children. When screened, these groups had significantly higher rates of type B tympanograms, indicating hearing loss from probable OME, with Māori being one and a half times more likely and Pacific peoples being two and a half times more likely than other ethnicities. These children were more likely to have undergone rescreening prior to referral and less likely to have access to primary care for consideration of intervention.

Increasing deprivation was also independently and strongly correlated with middle ear disease. The cause of this is likely related to higher incidence of respiratory infections, and links have been made with poor housing conditions including dampness, poor heating and crowding.²⁰

The higher rates of middle ear disease for Māori and Pacific children and those living in deprivation are broadly consistent with previous studies.^8–11 It is likely that the ongoing detrimental effects of colonisation as well as other social and economic determinants of health have a significant role in these ethnic disparities.¹⁶ These determinants also contribute to healthcare access, affecting not only access to primary healthcare after the screening process, but also the opportunity to participate in the screen in the first place, compounding the inequity. It was interesting to note that ethnicity and deprivation were independent risk factors for OME in our dataset; more research would be useful to confirm this.

Māori and Pacific children are significantly disadvantaged by our health system in the cohort studied for this research, suggesting the screening programme is almost twice as likely to fail to offer hearing testing to these groups. As the B4SC database utilises the primary care enrolment database,¹ those not enrolled will possibly be less likely to be offered screening. The fact that a number of children were offered screening despite this suggests that the other pathways into the programme can be effective. This could be investigated more closely with further research to identify these successes. The health impact for this group of missed children can be inferred from other research including wide ranging short- and long-term effects of OME on child behaviours, learned social responses, language and cognition.^13,21

A child may either be referred to primary healthcare or rescreened within the programme when found to have a flat tympanogram. Children in areas of higher deprivation and Māori and Pacific children are significantly more likely to be coded as having been rescreened, rather than as referred at first presentation, within the database. The decision to refer or rescreen is based initially on whether the child can cooperate and is developmentally ready for the test. Guidance from the audiometric results then subsequently directs decision making. For children who respond at higher thresholds but not at lower levels, the technicians then consider whether there are any concerns about speech and language development or other developmental difficulties. This is informed by diagnosed comorbidities but also by parental and educator concern and by the technician’s impression of the child.

This partially subjective assessment potentially allows the introduction of bias and/or regional differences into decision making in the screening pathway. In many places around the country, significant efforts are being made by the regional screening units to engage with local populations, and children may be offered rescreening intentionally due to lack of access to primary care. Children who do undergo rescreening would potentially have a longer lead time from their first screening assessment to subsequent management of their underlying middle ear disease, which may exacerbate the negative impacts of the condition. If the findings in this study do reflect a systematic bias rather than simply underlying differences in audiometric findings, this may indicate an inequitable service which is prolonging pathways to management in those populations that need it most. Unfortunately, reasons for selecting rescreen or refer decision options after flat tympanometry are not well captured in data reporting and warrant further qualitative research to inform policy guidelines.

The data collected were robust as they are linked to each individual’s national health number which does not change regardless of where the child is in the country. By analysing whole population data over a number of years, our findings are likely to be an accurate assessment of the state of middle ear disease in four-year-old children in Aotearoa New Zealand.

There are limitations in our dataset. We do not have the ability to determine why children underwent rescreening. It could be argued that different populations had different audiometric results including variable resolution rates in middle ear effusion between a first and second screening test, leading to a perceived variance in the rates of rescreening. Different practices in different regions may have also led to skewed results or affected the evidence of an inequitable provision of service. Our assessment of the population of children who did not undergo screening at all is an estimate which is probably based on incomplete data. We could not identify children who had no interactions with the health system at all, and transient populations may not have been identified.

Only children who demonstrated evidence of hearing loss proceeded for tympanometry in the screening programme. Type B, or flat tympanograms, were used as a marker for the presence of OME; however, the results can be confounded by other conditions including wax impaction, canal blockage or eardrum perforation. Although some of these conditions can be differentiated by using tympanometry volume, this data is not input to the national database. This may mean that the overall prevalence of OME is different from what was estimated, although these children still required further assessment or intervention to manage their hearing loss.

Our dataset included the period that the country was under COVID-19 restrictions. This will have undoubtedly affected both access to the screening programme and the prevalence of middle ear disease. Internationally, prevalence of OME (and other non–COVID-19 infections) was lower during the pandemic period, hence these results may underestimate current OME statistics.²²

OME related hearing loss remains a significantly prevalent health issue in Aotearoa New Zealand children approaching school age, affecting one in 10 Pacific children, one in 14 Māori children and one in 20 children overall. This analysis of Aotearoa New Zealand–specific nationwide whole-population OME prevalence data based on tympanometry screening should inform the ongoing redevelopment of the preschool screening programmes, as the current screening programme design and implementation has gaps in the provision of equity. Redesign with proportionate universalism is critical, enabling the screening to be delivered in a different way to those that most need it.

Screening programme delivery

More accessible, culturally responsive, targeted programmes have been and are currently being trialled and implemented within select communities around the country.^23–26 These community programmes should be investigated further, and examples that are working well should be embedded in the nationwide delivery model. Ways of seeking enrolment for children currently being missed should be explored.

Screening programme design

Screening programmes need to have clear intervention pathways for children identified with problems. For children who have hearing deficits associated with type B tympanograms, pathways to assessment, referral and intervention should be clear and supported. This should include children who do not currently have access to primary care providers.

We recommend a timely and holistic approach which can be implemented directly from screening. This should include home and community-based speech and language support, online resources for parents and families, listening and communication activities and support for healthy housing and financial welfare.

In addition to data collection, database management and reporting of the B4SC should include type B tympanograms as a marker of middle ear disease in order for this to be monitored and inform future policy.

Alternative workforce models of care, including nurse practitioner community-based clinics, school-based outreach and advanced audiology-led programmes have been demonstrated in Aotearoa New Zealand to improve access for timely investigation and management of childhood ear disease.²⁵ These models are an example of an effective intervention pathway providing care which might occur after screening. Such programmes could be expanded to other targeted areas of the country and may especially improve access for children who are not enrolled with primary care providers after screening if referral pathways were streamlined.

Public health policy

Middle ear disease, including hearing loss from OME, has received little public health attention this century, with varying policy advice.²⁷ Calls for improved public health policy and intervention in the 1980s gained some traction with mobile nurse-led ear services such as “ear-vans”. Many of these services have now been eroded due to lack of funding.

Policy design needs to be based on contextually relevant research. The authors disagree with past Aotearoa New Zealand policymakers who have relied on overseas research concluding active intervention for OME has demonstrated little difference in long-term outcomes.^27–29 Much of that research was conducted in contexts that may not be generalisable, often excluding children who are most vulnerable due to deprivation and other socio-cultural factors such as multilingual households. Indeed, more recent local research suggests that standard follow-up models do not adequately manage children with middle ear disease, disproportionately impacting Māori children.¹¹ Current pathways and priorities have not been adapted for these specific contexts, compounding the impact of hearing loss for these groups.

This study supports other research demonstrating that middle ear disease remains an important cause of childhood hearing loss, especially affecting Māori and Pacific children and those living in deprivation.

We recommend an integrated policy programme with proportionate universalism through the journey from identification to intervention from the range of providers involved. This programme needs to be developed with co-design principles and communities, ideally with cross-sector collaboration such as between the education and health sectors and with support from the ministries of health, education and social development.