The health and educational costs of preterm birth to 18 years of age in Australia

Preterm birth is the greatest cause of death up to five years of age and an important contributor to lifelong disability. There is increasing evidence that a meaningful proportion of early births may be prevented, but widespread introduction of effective preventive strategies will require financial support.


INTRODUCTION
Preterm birth represents one of the major challenges to contemporary health care. Being born too early is the greatest cause of death in newborns and a significant contributor to ongoing disability. 1 Birth at the very early gestational ages is associated with increased risk of cerebral palsy, blindness and deafness. 2 Being born at later gestational ages, extending to early term, is associated with increased risk of learning and behavioural disorders at school age. 3 Together, the consequences of preterm birth for individuals, families and societies are considerable, both in terms of human suffering and economic consequences.
The importance of quantifying the economic costs of preterm birth has been highlighted in recent years as a result of prevention becoming a realistic possibility. Several clinical strategies have now shown promise. When applied as a multifaceted program across an entire population the rate of preterm birth can be reduced, at least by about 8%. 4 Further advances in potential interventions can be expected to make prevention even more effective.
Decisions to apply resources to prevention, as opposed to just treatment, are best based on assessments that include reliable economic estimates. Several studies have now been published reporting the considerable annual cost of preterm birth across nations, including England and Wales, 5 Canada 6 and the USA. 7 The purpose of this study was to estimate the economic cost of preterm birth to the Australian government and provide financial information enabling appropriate decision making in allocating resources to the prevention and treatment of preterm birth.

Overview
A decision-analytic model was developed to estimate the costs of preterm birth in Australia for a hypothetical cohort of 314 814 Australian children born in 2016. The modelling framework was similar to an approach used for England and Wales, 5 and Canada, 6 but adapted for an Australian context. Costs from birth to 18 years of age were estimated from a government perspective, discounted at 3.5% per annum 8

Model framework
The live cohort enters the model from birth in gestational age groups: extremely preterm (20-27 weeks); moderately preterm (28-31 weeks); late preterm (32-36 weeks) and term (≥37 weeks) ( Fig. 1). After birth, the majority of neonates are discharged and progress to two years of age. A small proportion die in the delivery room, some are transferred for neonatal care and discharged or die before discharge from neonatal care. Hospitalisations and survivorship are modelled annually from two to 18 years of age. Levels of disability at primary and secondary school are imputed based on the relative risk of special education needs from the literature.

Model parameters
Australian data were available for the majority of parameters (Tables A1, A2 in Appendix S1). The number of births in each state by gestational age was obtained from the Australian Institute of Health and Welfare (AIHW) perinatal statistics. 9 The probabilities of death in the delivery room and of neonatal admission were estimated for each gestational age group using the AIHW perinatal statistics. 9 The requirement for neonatal emergency transport was set at one in ten neonatal admissions based on an Australian study of infants born outside tertiary centres. 10 The probability of surviving to discharge from the neonatal hospitalisation was estimated for each gestational age group using the Australian and New Zealand neonatal network data. 11 The annual probability of hospitalisation for each gestational age group was derived from an Australian population-based longitudinal study of hospital admissions until 18 years of age. 12 As that study combined gestational ages below 32 weeks, probabilities of hospitalisation for the 20-27 weeks and 28-31 weeks groups were assumed to be equal, with the effects of this assumption tested in a supplementary sensitivity analysis. Survivorship data were obtained from Australian Life tables, 13 with relative risks applied to preterm birth cohorts derived from the literature. 14

Healthcare costs
Delivery costs by state were obtained from the Independent Hospital Pricing Authority (IHPA) maternity data, which provides cost and number of separations of each maternity diagnostic related group (DRG) by state. 15 To calculate an average cost of delivery by gestational age group and state, DRGs were weighted by type of birth, length of stay and public or private setting based on the latest AIHW perinatal statistics. 9 Average neonatal admission costs by gestational age group were calculated from IHPA data, weighted according to birthweight based on the latest AIHW perinatal statistics. 9 As IHPA does not provide neonatal DRG data by state, national admission costs were varied by state based on the average variation of all state hospitalisation costs. 16 Newborn Hospitalisation costs, excluding delivery and neonatal admissions, were obtained from a data request to IHPA for childhood hospital separations. 16 Average hospitalisation costs by gestational age group were calculated from child DRG separations weighted according to the Australian longitudinal study data using hospitalisations that occurred after discharge from the birth-related admission until 18 years of age. 12 Hospital admissions were reported based on the International Classification of Diseases, and mapped to DRGs. 19 Costs for the 20-27 and 28-31 weeks groups were assumed to be equal.

Wider government costs
The Australian government makes a payment of $2200 to eligible parents for a first stillborn child and $1101 for any subsequent stillborn children. 21 AIHW data show that a stillbirth was preceded by one or more stillbirths in 7% of cases. 22 A survey of the families of stillborn children suggests that 22% of families are likely to be eligible. 23 The average cost to government was therefore estimated at $475 per stillbirth.
Since 2018, Australian schools are provided a disability loading with a base rate of $11 343 per primary student increased by 42% for a supplementary level of disability, 146% for a substantial disability, and 312% for an extensive disability. For secondary students, the base rate is $14 254, and the disability loadings are 33, 116 and 248% respectively. 24 The mapping of schooling disability categories was based on the population-based Scottish study of 407 503 children, which linked school census data with gestational age that highlighted the relative risk of special education needs for preterm births of 1.53, 2.66 and 6.92 for late, moderately and extremely preterm cohorts respectively. 25 All costs are summarised in Appendix S1 (Tables A3, A4).

Analytical methods
Costs were estimated from birth to 18 years of age for each ges-

Incremental costs of preterm births
The incremental cost of an extremely preterm, very preterm and preterm birth relative to a term birth was estimated at $236 036, $89 709 and $25 417, respectively ( Table 1).
The major contributors to the incremental cost of preterm births were neonatal admission costs contributing around twothirds of the cost difference, and schooling costs contributing around a quarter of the cost difference.
The incremental cost differences of preterm birth varied by state, with the Northern Territory, Western Australia and the Australian Capital Territory facing the highest cost differences, while Victoria and New South Wales faced the lowest ( Table 2).

Overall cost burden of preterm births
The overall incremental cost of preterm birth to government was estimated at $1.413 billion for the 2016 birth cohort (Table 3).     26 It was estimated that prolonging each preterm birth by one week could potentially reduce medical expenditures by more than $25 million.

F I G U R E
In the present study, the major contributor to incremental costs was neonatal intensive care, contributing about two-thirds of the cost difference. One-quarter came from additional costs of schooling. The most significant burden of neonatal costs per child comes from those born at the earliest gestational ages who also face greater rates of ongoing disability. However, the effects of preterm birth on subsequent educational needs extends across the gestational age spectrum and includes early term births. By linking school census data to birth outcomes in Scotland, the odds ratios of special educational needs increased progressively across the whole gestational-age-at-birth spectrum from 24 weeks to term. 25    It should no longer be assumed that the high costs of preterm birth are an inevitable consequence of our reproduction. There are many pathways to untimely early birth and some are now amenable to prevention. In 2014, a whole-of-population multifaceted program was introduced across Western Australia to safely lower the rate of preterm birth. In the first full calendar year, the statewide rate of preterm birth was reduced by 7.6% and in the sole tertiary level perinatal centre by 20%. 4 Over the following two years the 20% reduction in the tertiary centre was maintained, but the effect dissipated in the secondary centres where the educational program had not been sustained. These results indicate that the rate of preterm birth can be safely reduced but the benefit depends on ongoing education across the various sites.
The benefits of preterm birth prevention include fewer children with behavioural and learning problems, including need for special education assistance. 27 It would be reasonable to conclude that the costs to government of preterm birth estimated in the current study up to the age of 18 years represent only a fraction of the eventual overall burden to individuals, families and the nation.
In conclusion, our study shows that the annual cost of preterm birth to the Australian government in the first 18 years of life is approximately $1.4 billion. Two-thirds of the costs are from neonatal care, and one-quarter arises from the need for special educational assistance. Our results also show that cost savings of $71 million and $141 million per year may be achieved by reducing the national rate of preterm birth nationally by only 5 and 10% respectively ( Table 4). Prevention of a reasonable proportion of untimely preterm births is now possible, and these data need to be applied when decisions are made to allocate resources to prevention and treatment interventions. Further analyses will be required as the ongoing costs across the public sector and to individuals in adulthood are likely to reveal significant longer-term economic consequences.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.

Supplementary Material
Appendix S1 Table A1. Parameters, probabilities and relative risks of events are shown for the decision-analytic model stratified by gestational age at birth (extremely, moderately and late preterm, and term birth). National probabilities are presented. The state-specific probabilities that were implemented into the model are shown in Table A2.