Article Text

First estimates of the potential cost and cost saving of protecting childhood hearing from damage caused by congenital CMV infection
  1. Eleri J Williams1,2,3,
  2. Joanne Gray4,
  3. Suzanne Luck5,6,
  4. Claire Atkinson5,
  5. Nicholas D Embleton2,3,
  6. Seilesh Kadambari7,
  7. Adrian Davis8,
  8. Paul Griffiths5,
  9. Mike Sharland7,
  10. Janet E Berrington2,3,
  11. Julia E Clark1,9,10
  1. 1Department of Paediatric Immunology and Infectious Diseases, Great North Children's Hospital, Royal Victoria Hospital, Newcastle-upon-Tyne, UK
  2. 2Newcastle Neonatal Service, Royal Victoria Infirmary, Newcastle-upon-Tyne, UK
  3. 3Faculty of Medical Sciences, Newcastle University, Newcastle-upon-Tyne, UK
  4. 4Faculty of Health and Life Sciences, Northumbria University, Newcastle-upon-Tyne, UK
  5. 5Centre for Virology (Royal Free), University College London Medical School, London, UK
  6. 6Department of Paediatrics, Kingston Hospitals NHS Foundation Trust, London, UK
  7. 7Paediatric Infectious Diseases Research Group, St George's, University Hospital of London, London, UK
  8. 8Public Health England, London, UK
  9. 9Department of Paediatric Infectious Diseases, Lady Cilento Children's Hospital, Brisbane, Queensland Australia
  10. 10University of Queensland, Brisbane, Australia
  1. Correspondence to Dr Eleri Williams, Department of Paediatric Immunology and Infectious Diseases, Great North Children's Hospital, Royal Victoria Hospital, Newcastle-upon-Tyne NE1 4LP, UK; Eleri.Williams{at}


Background Congenital cytomegalovirus (cCMV) is an important cause of childhood deafness, which is modifiable if diagnosed within the first month of life. Targeted screening of infants who do not pass their newborn hearing screening tests in England is a feasible approach to identify and treat cases to improve hearing outcome.

Aims To conduct a cost analysis of targeted screening and subsequent treatment for cCMV-related sensorineural hearing loss (SNHL) in an, otherwise, asymptomatic infant, from the perspective of the UK National Health Service (NHS).

Methods Using data from the newborn hearing screening programme (NHSP) in England and a recent study of targeted screening for cCMV using salivary swabs within the NHSP, we estimate the cost (in UK pounds (£)) to the NHS. The cost of screening (time, swabs and PCR), assessing, treating and following up cases is calculated. The cost per case of preventing hearing deterioration secondary to cCMV with targeted screening is calculated.

Results The cost of identifying, assessing and treating a case of cCMV-related SNHL through targeted cCMV screening is estimated to be £6683. The cost of improving hearing outcome for an infant with cCMV-related SNHL through targeted screening and treatment is estimated at £14 202.

Conclusions The costs of targeted screening for cCMV using salivary swabs integrated within NHSP resulted in an estimate of cost per case that compares favourably with other screening programmes. This could be used in future studies to estimate the full economic value in terms of incremental costs and incremental health benefits.

  • Costing
  • Infectious Diseases
  • Audiology
  • Virology
  • General Paediatrics

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic

  • Congenital cytomegalovirus (cCMV) is the most common non-genetic cause of sensorineural hearing loss (SNHL).

  • In infants with cCMV, antiviral treatment started in the first month of life improves audiological outcome.

  • Targeted salivary screening for cCMV within the established newborn hearing screening programme (NHSP) in England is feasible and acceptable to mothers.

What this study adds

  • Cost estimates of targeted salivary screening for cCMV within the established NHSP in England.


Cytomegalovirus (CMV) is the most common congenital infection, estimated to affect 0.3% UK births.1 ,2 Findings at birth vary, but only around 13% will be detected on routine neonatal examination,3 the majority remaining undiagnosed at this age (ie, are ‘asymptomatic’). However, long-term consequences are not uncommon in both symptomatic and asymptomatic cases: sensorineural hearing loss (SNHL) occurs in 10%–15% overall.3–5 Although the individual risk of SNHL is higher in infants with symptomatic congenital CMV (cCMV) infection, numerically more cases of cCMV-related SNHL are found in infants who are ‘asymptomatic’ at birth, given their relative frequencies.6 This hearing loss is variable, but, in approximately half, is detectable at birth,4 ,7 ,8 offering a potential route to detecting some ‘asymptomatic’ cCMV, using targeted screening of infants who do not achieve a clear pass on newborn hearing screening.

If correctly timed, this could allow identification of cCMV within the treatment window where an antiviral drug has been shown to improve the final hearing outcome.9 In infants with evidence of central nervous system (CNS) infection (microcephaly, intracranial calcifications, abnormal cerebrospinal fluid, chorioretinitis or hearing deficit), intravenous ganciclovir given before 1 month of age reduced hearing deterioration from 68% from baseline in the untreated group to 21% in treated group.9

We recently proposed a system for identifying neonates with cCMV-related SNHL who are, otherwise, asymptomatic (the Benefits of Extended Screening Testing (BEST) study,10 proposed screening pathway shown in figure 1). Testing 404 infants who did not pass their universal newborn hearing screening programme (NHSP) tests using salivary CMV PCR identified six (1.5%) with cCMV. This targeted approach identifies infants who will benefit from early treatment and minimises identifying infants with cCMV who are truly asymptomatic (without SNHL) where no current evidence of treatment benefit exists. The cost of screening for cCMV in any format (targeted or universal) has not been published, but is important for policy-makers. This analysis aims to explore the healthcare costs and potential benefits of this approach.

Figure 1

Proposed pathway of targeted screening for cCMV within established NHSP. The first box is coloured red to highlight the screening intervention that is proposed. The second box is red to identify the condition that the screening process is trying to identify. AABR, automated auditory brainstem responses; AOE, automated otoacoustic emissions; cCMV, congenital cytomegalovirus; CMV, cytomegalovirus; NHSP, newborn hearing screening programme; SNHL, sensorineural hearing loss.


To conduct a cost analysis of targeted screening and subsequent treatment for cCMV-related SNHL in an, otherwise, asymptomatic infant, from the perspective of the UK NHS.


Data sources

We combined data from two sources to populate our analysis (figure 1) of targeted and integrated screening. Data regarding national numbers of referrals and cases detected of SNHL were obtained for 1 year (2010–2011) from the NHSP ( We assumed that 100% of liveborn infants are screened through NHSP within the first 21 days of life, and 100% of those ‘referred’ for more testing after NHSP are screened for cCMV. The BEST study10 provided data regarding number of referrals and neonates diagnosed with cCMV and cCMV-related SNHL. This English study screened infants ‘referred’ after NHSP for cCMV within 21 days of life, excluding infants with known cCMV. Of the 411 parent–infant pairs enrolled into the study, 404 (98%) returned a sample for CMV PCR testing. Six infants (6/404 (1.5%)) ‘referred’ from NHSP were correctly identified to have cCMV through screening, none falsely. Three of those six infants had confirmed SNHL, and the others were truly ‘asymptomatic’ at birth with normal hearing on further assessment. Combining these data sources, we have estimated numbers of referrals and numbers of children who would be diagnosed with cCMV-related SNHL over 1 year (figure 2). For the analysis, we assumed that infants identified with cCMV are followed up using standard NHS routes, and investigated and managed in accordance with published guidelines.11 Audiological investigation and management is assumed to be the same as current practice.

Figure 2

Populations predicted from a baseline population of 100 000 infants by the costing model for targeted cCMV screening. cCMV, congenital cytomegalovirus; NHSP, newborn hearing screening programme; SNHL, sensorineural hearing loss; VGV, valganciclovir.

Cost analysis

The cost analysis includes the costs incurred by the NHS of identifying and treating cCMV-related SNHL. Specifically, the healthcare costs of identifying cCMV-related SNHL in an, otherwise, asymptomatic newborn infant by targeted screening integrated into the NHSP system and the costs associated with ‘protecting’ the hearing of this, otherwise, asymptomatic infant.

cCMV detection options

CMV can be identified in dried blood spot (DBS), urine or saliva. Salivary screening for cCMV by PCR is very sensitive and specific (97.4% and 99.9%, respectively).12 DBS samples have poor sensitivity,13 and urinary samples are difficult to obtain.10 ,14 This analysis, therefore, assumes that salivary swabs are optimal for screening, and as the sensitivity and specificity are so high,12 we have not accounted for false-positive and false-negative cases. False-positive screenings would not be subjected to treatment, and thus, not incur treatment costs, since the next step on the pathway is confirmatory testing at a cost of only £345.44 per case.

Healthcare costs

The NHS cost of the screening strategy includes costs of performing the swab (NHSP staff time, the initial swab, saliva PCR test), contacting parents and arranging appointments. They also include the cost of initial assessment of infants identified with cCMV, including an appointment with a paediatric consultant, investigations as per local guidelines11 and a second follow-up appointment for discussion of results. The cost of cranial MRI scan was also included. The actual time needed to take and process the samples and later collate and communicate the results are estimated from BEST experiences (table 1; for expanded version, see online supplementary table S1A).

Table 1

Healthcare costs of targeted screening for congenital CMV-related SNHL (further details of cost elements provided in online supplementary tables S1A, B)

The total cost of treating and following up all infants identified with cCMV (regardless of their hearing status) was calculated using current guidelines11 (table 1, for expanded version; see online supplementary table S1B). For infants with cCMV-related SNHL, these comprised of costs relating to antiviral treatment (based on oral valganciclovir at 16 mg/kg twice daily for 6 weeks, the preferred route for ‘otherwise well’ infants), and enhanced blood monitoring and clinical reviews. The amount of valganciclovir required was estimated using the 50th centile weight of a male infant aged 4–10 weeks,15 assuming weekly weight and dose adjustment. There are no additional audiological costs in this group as all infants with SNHL have routine follow-up regardless of CMV status. For infants with cCMV, but no SNHL, costs include a paediatric review and additional audiological monitoring.16

All costs were estimated at 2011 prices and discounted (at 3.5%)17 where required using a range of sources and as specified.

Two NHS costs were calculated. The first was the cost of identifying cases of cCMV-related SNHL by implementing the targeted screening strategy per 100 000 infants undergoing NHSP. The second was the complete cost of identifying cCMV (as per first cost) plus the cost of treating and following up cases of cCMV (both with cCMV-related SNHL and truly asymptomatic infants) by implementing the targeted screening strategy per 100 000 infants undergoing NHSP. From these total costs, it was possible to calculate the cost of screening per case of cCMV-related SNHL identified and treated.

The desired outcome associated with identifying and treating infants with cCMV-related SNHL is that of protecting these infants’ hearing from further deterioration. In order to calculate the cost to ‘protect’ these cases of hearing loss, the number of infants whose hearing would be assumed to be ‘protected’ by treatment is estimated using the only published randomised controlled trial on cCMV treatment efficacy where 47% of babies had their hearing ‘protected’.9 The cost to ‘protect’ one case of cCMV-related SNHL was thus: total cost of implementing screening, treatment and follow-up in 100 000 infant population/(number of infants treated with cCMV-related SNHL in this population ×0.47).

As a sensitivity analysis, this cost is also calculated assuming that the rates of cCMV are half and double that identified in the BEST study. In addition, in order to capture variability and uncertainty in relation to costs associated with the interventions involved, further sensitivity analysis was conducted by examining the effects of a change in cost (by using extreme values with factors of 0.5 and 2) of the investigations (haematology, biochemistry, virology, therapeutic drug monitoring and radiological procedures) and staff time (NHSP screeners, paediatrician and audiologists).


Figure 2 shows populations modelled on a population of 100 000 newborns in the proposed integrated and targeted screening model. Accepting our assumptions of referral rates, participation and cCMV detection rates from a population of 100 000 infants undergoing NHSP, 2294 infants will be ‘referred’, 34 have cCMV and 17 of these have cCMV-related SNHL.

The NHS costs of implementing proposed screening for identification and treatment are summarised in table 2. The variation in cost assuming different rates of cCMV, investigatory and staff costs are presented in table 3.

Table 2

Costs to the NHS per case

Table 3

Costs to the NHS per case identified when key assumptions are varied

In the costing model, 17 infants were offered treatment for cCMV-related SNHL; of which, eight (47%) have their hearing ‘protected’ through early detection and treatment. Therefore, the NHS cost to ‘protect’ a case of cCMV-related SNHL is calculated as £14 202. If the efficacy of valganciclovir treatment for improving the hearing outcome was only half of that of the published study, the cost to improve hearing outcome would be £28 404 (table 3).


This is the first published analysis to examine the potential healthcare costs of targeted screening for cCMV. This approach captures those infants who might benefit from treatment, but who are currently missed or diagnosed too late for this to be offered. The costs of this screening and assessment process are in keeping with other current national screening programmes (£31 000 per case of congenital hearing loss,18 £42 000–356 000 per case of phenylketonuria19 and £5000 per case of cystic fibrosis detected20) emphasising the need for further consideration of this approach.

We estimate that the cost of ‘protecting’ a case of childhood SNHL from cCMV identified and treated through a national targeted screening programme would be ∼£14 000. In comparison, detailed health economic analysis suggests that the societal cost of bilateral hearing impairment in children aged 7–9 years rises from £9120 to £21 179 per year from moderate to severely affected children,21 and the lifelong cost of a paediatric cochlear implant22 is £82 000–108 000 (recommended in cases of severe bilateral hearing impairment from cCMV23 ,24).

A rational economic argument can, therefore, be presented for reducing the severity of hearing impairment in children at a cost of only £14 000 per case. Identifying and treating infants with cCMV-related SNHL at birth, for each case of prevented progression, could over the lifetime of the child produce cost savings to the healthcare system. There are no published cost estimates of the cost of unilateral deafness, which makes up a considerable proportion of deafness in cCMV,7 but there is increasing evidence that these children have lower intelligence coefficients,25 and are at increased risk of academic failure,26 and future health economic analyses should take this into consideration by incorporating a wider societal perspective than only that of healthcare.

There are several limitations of this study. As a cost analysis, despite it contributing to the paucity of evidence, it is not a full economic evaluation where a comparative analysis of alternative courses of action in terms of costs and consequences are analysed and presented incrementally. As such, inferences regarding the relative efficiency of this targeted screening programme cannot be made. Furthermore, this analysis is based on healthcare costs only and takes no account of family and wider societal costs of this targeted screening approach. There are potential costs to the family in terms of further appointments and anxiety, but also potential resource savings in improved hearing outcomes (and possibly developmental outcomes27). A full economic evaluation involving a suitable comparator such as usual care that accounts for the longer term costs and benefits of the screening programme is needed. It could be argued that the most efficient method would be a decision analytic model built on the evidence base regarding prevalence data and long-term costs and benefits of early treatment. However, this paper contributes to the economic evidence base regarding NHS cost estimates of targeted screening for cCMV and the costs of treating this group of infants in order to ‘protect’ a case of cCMV-related SNHL.

The population estimates used in this study are from a relatively small cohort where the prevalence of cCMV in the ‘referred’ population was 1.5%: half that identified from similar studies in the USA.28 ,29 Even if we consider that our prevalence may be wrong by a factor of 50%, and the rates of cCMV are half (0.8%) or double (3.4%) of that used in our model, the cost of identifying a case of cCMV-related SNHL varied between £9224 and £5413, and the cost of ‘protecting’ a case of cCMV-related SNHL varied between £19 601 and £11 502—still within a range that is comparable with other national targeted screening programmes, and is worthy of consideration for further research. The ‘protective’ benefit of treatment with oral valganciclovir is estimated from evidence from intravenous ganciclovir in infants with symptomatic cCMV at birth. Although this included neonates with SNHL, babies with other features of CNS cCMV and without SNHL at birth were included, a different population to that being considered here, and so effect size may vary. There are no current published studies examining the efficacy of treatment with valganciclovir for infants with cCMV SNHL alone. Although the duration of beneficial effect on hearing is as yet unclear, it would seem to be for at least 2 years. However, valganciclovir as a treatment for cCMV is now the gold standard in cCMV treatment trials and common UK practice. A recent study suggests that there might be enhanced audiological benefit from prolonged courses of valganciclovir compared with standard courses (6 months compared with 6 weeks),30 which would increase the drug cost by approximately a fivefold factor (allowing for infant growth), and involve additional blood testing costs, but the exact monitoring regime that would be recommended for such a prolonged treatment course is unknown, as is the proportion of children whose hearing would be improved with longer treatment, making further economic comment impractical at this time. Even if the efficacy of valganciclovir in our population was half of that of ganciclovir, the cost per hearing loss ‘protected’ would still be ∼£28 000, and remain worthy of further exploration.

We also assumed that all infants are screened by NHSP within 21 days of life, none is ‘missed’ or screened later. Most infants do complete their initial NHSP hearing screening within the first 21 days of life; however, ∼15% of infants are screened later (personal correspondence). Most of these have spent >48 h in neonatal intensive care units (NICU), and are assumed to be a ‘sick’ or preterm population. Although this group have increased risk of SNHL, any role of cCMV is unclear currently.31–34 The BEST study also excluded this group, and we do not feel this will significantly affect our findings. This ‘NICU’ group clearly require further separate study.

In conclusion, this is the first published analysis showing the basic healthcare costs of a realistic mechanism of targeted screening for cCMV-related SNHL within England. Considering the low cost and significant potential benefits associated with identifying these infants, we suggest that the feasibility and subsequent evaluation of national roll-out of such a programme warrants further urgent exploration.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors This study was conceived by EJW, JEB and JEC. The study was designed and undertaken by EJW, JG and JEB. EJW, JG, JEB and JEC wrote the first draft. All authors have reviewed and revised subsequent drafts.

  • Funding This study was funded by Sparks, the children’s medical research charity ( Sparks grant reference 09NGH01.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.