Geographical inequity in out-of-pocket costs for families of children with congenital heart disease requiring neurodevelopment support: an examination of policy scenarios using economic modelling

Abstract

Background

Out-of-pocket costs represent care-limiting financial burden among families of children with congenital heart disease at risk of adverse neurodevelopmental outcomes. Neurodevelopmental follow-up care is often centred around major paediatric cardiac centres which can contribute to geographic inequity for people who reside in rural, regional or remote areas. This study examined out-of-pocket costs in three potential policy scenarios with neurodevelopment-related care location foci being either: specialist paediatric cardiac centre (Tertiary); nearest paediatric centre (Secondary); nearest community-based services (Primary).

Methods

This modelling study used person-level stochastic microsimulations informed by a representative population cohort to estimate out-of-pocket costs under the three policy scenarios. Residence and service provider geographical locations were mapped using geographical information system software. Probabilistic estimates for out-of-pocket costs (direct healthcare, direct non-healthcare, indirect productivity costs) for the first five years of life for four illustrative categories of neurodevelopmental support requirements (surveillance only, mild, moderate and severe neurodevelopmental delays) were generated from 10,000 microsimulations of 1000 families in each scenario.

Results

Out-of-pocket costs for families of children who require neurodevelopmental support were substantial across all scenarios and locations of residence, with indirect productivity costs representing the largest burden. Direct travel costs for attending healthcare appointments were also substantial, and differentially affected those who resided in rural, regional or remote areas. With existing healthcare funding support structures, direct healthcare costs were a comparatively smaller contributor to financial burden. Overall, care provision at the nearest paediatric centre (Secondary) scenario was most likely to mitigate financial burden on families. Out-of-pocket costs for families of children with severe delays were approximately 8.7 to 15.4 times higher than for families of children requiring surveillance only. Findings were robust to sensitivity analyses.

Conclusion

Out-of-pocket costs for neurodevelopmental support for families of children with congenital heart disease are substantial. These costs differentially affect families of children with more severe developmental delays, as well as those who reside in rural, regional and remote areas. Policy strategies to shift loci of services closer to home have potential to alleviate some out-of-pocket costs for families, with services based at the nearest secondary paediatric centre likely to reduce geographic inequity.

Background

Geographical inequities in social determinants of health, healthcare access, the cost of services, and health outcomes are widespread across many regions globally [1–5]. These inequities include differential direct and indirect out-of-pocket costs for accessing healthcare among people who reside in rural, regional, or remote areas [1]. This issue is particularly salient in jurisdictions where large distances between residential locations and specialised healthcare services create barriers to timely access. Out-of-pocket cost inequity can be exacerbated when specialised services are concentrated in major metropolitan centres, requiring families to travel long distances for care. This is particularly relevant for children with congenital heart disease who are at elevated risk of neurodevelopmental delays and may require ongoing neurodevelopmental screening, assessment, referral and intervention in early childhood [3].

Despite advances in virtual care, face-to-face consultations remain necessary for neurodevelopmental assessment and support for many children with congenital heart disease, particularly infants and toddlers [6]. In some regions, including Australia [7], Canada [8], and France [9], efforts have been made to support decentralised models of neurodevelopmental follow-up when care does not need to be solely delivered at metropolitan paediatric cardiac centres [10]. The Congenital Heart Disease Long-term Improvement in Functional hEalth (CHD LIFE) program in Queensland, Australia, is an established example of these efforts that has been previously described and is relevant to the present study [7]. Among Australian states, which each have their own unique geographies and health service profiles, Queensland was selected for this study as an informative use case to examine how alternative policy settings may influence family out-of-pocket costs, specifically in relation to the geographical location of neurodevelopment-related service provision. The state spans sparsely populated rural and remote areas, regional centres, and major metropolitan areas, and has healthcare infrastructure with potential to support service delivery in differing locations. In a recent study, we quantified geographical inequity in access to tertiary-hospital-based specialised services for families in rural, regional, and remote areas across Queensland [3]. Mapping of patient residences and service locations demonstrated that a substantial proportion of high-risk patients live large distances from tertiary-level care [3]. However, that work did not quantify the likely magnitude of out-of-pocket cost differences for families under alternative service location scenarios for face-to-face neurodevelopment care.

While there are many factors that must be considered when evaluating policy response options, including clinical effectiveness, safety, workforce capacity, consumer preferences, and implementation feasibility, the potential impact on family out-of-pocket costs is an important consideration for informing equitable service design. Policy responses that support decentralisation of services, where possible and clinically appropriate, to enable neurodevelopment-related care to be delivered closer to home, are a plausible strategy to reduce geographical inequity [3, 4, 7, 10, 11]. In the context of highly developed healthcare systems with overlapping state and federal initiatives intended to reduce geographic inequities, it is important to anticipate potential consequences for patients and families to avoid negative unintended effects. However, nuanced, policy-relevant economic modelling to estimate the potential financial impacts on patients and families has typically not been present in prior research in the field [12].

The aim of this study was to report analyses that focus on family out-of-pocket costs as a relevant consideration for public healthcare system decision making in the context of the Australian universal healthcare system. We examined the potential impact on family out-of-pocket costs across three policy scenarios using economic modelling. The scenarios differed by the geographical locus for accessing face-to-face neurodevelopment support services: a tertiary cardiac centre (Tertiary scenario), the nearest secondary hospital with specialist paediatric services (Secondary scenario), or local community-based services (Primary scenario). The modelling also explored differential out-of-pocket cost impacts for families residing within and outside the major metropolitan area in the region modelled, and across differing levels of neurodevelopmental support requirements. While information on out-of-pocket costs alone only form one of the many considerations when design policy responses to reduce geographical inequity, it is nonetheless important, and often overlooked.

Methods

Design and modelling approach

Stochastic person-level microsimulations were used to examine the impact of the three different policy scenarios from the perspective of out-of-pocket costs for families of children with congenital heart disease with a time-horizon of the first 5 years of life. We adopted this perspective and horizon as high financial burden is a known barrier to service access that may disproportionately affect economically vulnerable families during important developmental stages prior to commencement of formal schooling.

Policy context

The policy context was the Australian healthcare system where geography-related face-to-face service access inequity for children and families is an important consideration [1, 2]. The state of Queensland (1.853 million km²) was the policy modelling use case. The Australian healthcare system is considered a universal healthcare system. However, there is policy complexity in supporting the provision of equitable service quality and access across vast, sparsely populated geographical areas for children with neurodevelopmental support needs who may benefit from different types of financial reimbursement and subsidies dependent on their location and eligibility.

The Australian healthcare system comprises a mix of publicly and privately funded healthcare providers servicing metropolitan, rural, regional and remote areas. Publicly funded healthcare services, typically funded through federal or state governments, are available to all Australian citizens and permanent residents, with no out-of-pocket direct healthcare costs for children accessing public hospitals. Hospitals and specialist services are typically concentrated in large urban centres, which in this study was the region of South East Queensland, where the state’s only large specialist paediatric hospital, the Queensland Children’s Hospital, was located. Other hospital locations across the state where specialist paediatric medical and allied health services are available, were considered secondary level hospitals. Although there is typically no direct public hospital costs, other out-of-pocket costs can contribute to substantial financial burden for families, including costs associated with travel and lost productivity.

There are several established government-associated financial support schemes intended to reduce financial burden for families with defined eligibility criteria. A state-based travel subsidy scheme is available to pay for or reimburse a portion of the child and carer costs for families who are required to travel > 50 km to access public hospital services [3]. This includes costs associated with flights, accommodation and meals for the child and an accompanying family member or carer to access public hospital services. Two federally funded schemes were also of particular relevance for services accessed in community settings: (i) Medicare [4], which provides some reimbursement for items on the Medicare Benefits Schedule including medical services and up to five allied health consultations annually; and, (ii) the National Disability Insurance Scheme which typically reimburses the entire direct cost of health professional consultations but has strict eligibility criteria related to severe and permanent disability [5].

Community based services in Queensland include some publicly funded services with no out-of-pocket costs; for example, child health clinics [6] that provide developmental assessments. However, most primary care and community based allied health services that deliver ongoing neurodevelopmental supportive care are provided by private healthcare providers. The extent to which families of children accessing those services are required to pay out-of-pocket costs is therefore dependent on the child’s eligibility for the federally funded National Disability Insurance Scheme, Medicare rebates, or private health insurance arrangements. The present study was intended to estimate costs for families currently accessing publicly funded hospital cardiac services that are freely available to all citizens and residents and represent the vast majority of paediatric cardiac service provision for children with congenital heart disease in the region. Consequently, the impact of private health insurance arrangements, which are typically not utilised by families that access public healthcare services, were considered beyond the scope of the present study. This contrasts with the National Disability Insurance Scheme and Medicare rebates, both of which were highly relevant and within scope.

Policy scenarios

The Tertiary policy scenario could be conceptualised as a centralised tertiary care model. In this scenario, families travel back to the tertiary hospital in South East Queensland, where specialist paediatric cardiac services are based, to access allied health and other specialist paediatric services. While this scenario had no direct healthcare costs for accessing public hospital services for consumers with any level of neurodevelopmental delay, it necessitated substantial travel commitments for many families. In this and all scenarios, families were assumed to access general medical practitioners in their local community rather than at a hospital. In this scenario, when allied health encounters from multiple disciplines were required at the tertiary facility, it was assumed that an appropriate level of appointment coordination would occur to allow the family to see multiple health professionals in the same episode of travel.

The Secondary policy scenario could be conceptualised as secondary centre-focused care model. In this scenario, families travel only to the nearest public hospital with specialist paediatric services. Consistent with the first policy scenario, there are no direct healthcare costs for accessing the required services in a secondary or tertiary public hospital for consumers with any level of neurodevelopmental delay. This scenario typically represented less travel distance requirements per encounter for consumers, with travel only required to access the closest (secondary or tertiary) public hospital.

The Primary policy scenario could be conceptualised as a community-based care model. In this scenario, families primarily access neurodevelopment support services in their local region. Because this scenario relies on accessing private community-based health services as an alternative to publicly funded hospitals, direct out-of-pocket healthcare costs were a relevant consideration. This scenario was expected to represent the least requirements for travel to access services per encounter. However, because service provision was dependent on local healthcare provider businesses, it was not assumed that consultation scheduling was coordinated in the same way as trips to a tertiary or secondary hospital.

Neurodevelopmental support requirements

The amount of face-to-face neurodevelopment support required for children with congenital heart disease varies widely in the population. To illustrate how the policy scenarios may impact out-of-pocket costs for children of families with differing severity of neurodevelopmental delay, four illustrative exemplar model parameterisations were used to represent: (i) surveillance only, (ii) mild delays, (iii) moderate delays, or (iv) severe neurodevelopmental delays (Table 1). Parameters in Table 1 were initially prepared by a panel of 4 healthcare professionals (SP, KE, JE, SMM) informed by published literature, local health service data and clinical experience during two workshops and asynchronously (via email) until consensus. The initial parameterisation for illustrative levels of support requirements was also presented to a broader stakeholder and investigator group as a validation step (an additional n = 7 medical, nursing, allied health, or health services researchers) in workshop format or one-to-one discussions. There was consensus among the broader stakeholder group that parameterisations for the exemplar levels were appropriate for illustrating the potential impact across the intended range of neurodevelopmental support requirements.

Table 1.

Exemplar frequency of face-to-face service access profiles for children in the first five year of life across different levels of neurodevelopmental support requirements

Service	Age Range	Frequency of service per year estimate for informing Poisson distribution parameterisation and expected range
		General Surveillance	Mild Delays	Moderate Delays	Severe Delays*
Child Health Clinic (provides developmental assessments)	Probability of use	65%	65%	30%	20%
	0–1 year	3 (0–5)	2 (0–5)	2 (0–5)	1 (0–5)
	1–3 years	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)
	3–5 years	1 (0–1)	1 (0–1)	0.3 (0–1)	0.3 (0–1)
General paediatrician	Probability of use	10%	30%	70%	90%
	0–1 year	1 (0–2)	1 (0–2)	2 (0–4)	2 (0–4)
	1–3 years	0.5 (0–2)	0.5 (0–2)	2 (0–4)	2 (0–4)
	3–5 years	0.5 (0–2)	0.5 (0–2)	2 (0–4)	2 (0–4)
General practitioner	Probability of use	40%	60%	60%	60%
	0–1 year	1.5 (0–2)	1.5 (0–2)	2 (0–3)	3 (0–4)
	1–3 years	1.5 (0–2)	2 (0–3)	3 (0–4)	3 (0–4)
	3–5 years	1.5 (0–2)	2 (0–3)	3 (0–4)	3 (0–4)
Occupational therapy	Probability of use	0%	20%	50%	50%
	0–1 year	0	6 (0–12)	6 (0–12)	12 (0–25)
	1–3 years	0	6 (0–12)	6 (0–12)	12 (0–25)
	3–5 years	0	6 (0–12)	9 (0–12)	12 (0–25)
Physiotherapy	Probability of use	0%	60%	65%	45%
	0–1 year	0	6 (0–12)	9 (0–12)	12 (0–25)
	1–3 years	0	6 (0–12)	9 (0–12)	12 (0–25)
	3–5 years	0	6 (0–12)	6 (0–12)	12 (0–25)
Speech Pathology	Probability of use	0%	60%	65%	50%
	0–1 year	0	As for feeding clinic
	1–3 years	0	6 (0–12)	9 (0–12)	12 (0–25)
	3–5 years	0	6 (0–12)	6 (0–12)	12 (0–25)
Dietetics	Probability of use	0%	20%	30%	30%
	0–1 year	0	As for feeding clinic
	1–3 years	0	3 (0–6)	6 (0–12)	9 (0–12)
	3–5 years	0	3 (0–6)	6 (0–12)	9 (0–12)
Feeding Clinic	Probability of use	0%	30%	50%	60%
	0–1 year	0	3 (0–6)	12 (0–25)	12 (0–25)
	1–3 years	0	0	0	0
	3–5 years	0	0	0	0
Music therapy	Probability of use	0%	0%	10%	20%
	0–1 year	0	0	3 (0–6)	6 (0–12)
	1–3 years	0	0	3 (0–6)	6 (0–12)
	3–5 years	0	0	3 (0–6)	6 (0–12)
Psychology	Probability of use	0%	10%	40%	30%
	0–1 year	0	0	0	0
	1–3 years	0	3 (0–6)	3 (0–6)	3 (0–6)
	3–5 years	0	3 (0–6)	3 (0–6)	6 (0–12)

*National Disability Insurance Scheme eligible

Direct healthcare costs

Direct healthcare costs included only the non-reimbursed out-of-pocket costs for families, with costs and reimbursement amounts dependent on the service delivery facility type as well as eligibility for reimbursement which differed across level of neurodevelopmental delay. For services provided at tertiary or secondary public hospitals in Australia, there was no out-of-pocket direct healthcare costs for families, including general paediatrician consultations in the Tertiary and Secondary scenario. For non-public hospital services, market rates for initial and follow-up consultations were sourced from 29 private providers distributed across the state of Queensland (summarised in Table 2). The Medicare benefits schedule typically reimburses the entire general medical practitioner consultation fee for paediatric services, meaning we parameterised models for all policy scenarios with no direct healthcare out-of-pocket cost for general practitioners. In the Primary Care scenario, general paediatrician consultations were considered to be reimbursed at the scheduled rate for private non-hospital-based consultations (initial consultation = AU$259.40, follow up consultation = AU$74.25). Children with severe neurodevelopmental support requirements were considered eligible for the National Disability Insurance Scheme when accessing services through private providers. All children with congenital heart disease, in all scenarios, were considered eligible for the Medicare Benefits Schedule allied health reimbursement for a maximum of 5 consultations (reimbursement rate = AU$60.35) per year, which was notably lower than the typical cost of allied health consultations fees (Table 2).

Table 2.

Gross mean and standard deviation (SD) market rate costs (2024 AU$) of paediatric initial and follow-up consultation in community-based settings prior to reimbursement or subsidy unless otherwise indicated

	Initial consultation		Standard consultation
	Mean	SD	Mean	SD
Child Health Clinics*	n/a	n/a	n/a	n/a
Paediatrician	$380.00	$16.33	$205.00	$14.72
General Practitioner†	n/a	n/a	n/a	n/a
Occupational Therapy	$289.10	$70.90	$205.46	$15.50
Physiotherapy	$291.33	$71.49	$183.00	$41.53
Feeding clinic‡	n/a	n/a	n/a	n/a
Speech and language pathology‡	$384.67	$85.73	$200.50	$59.82
Dietician‡	$198.00	$16.57	$128.50	$12.02
Music Therapy	$193.99	$25.08	$177.44	$22.94
Psychology	$241.67	$14.34	$200.00	$33.35

* Publicly funded services that typically do not have any out-of-pocket costs for citizens and residents; † Services bulk-billed under Medicare Benefits Schedule with no direct out-of-pocket consultation fee for children, amount reimbursed to practitioner varies according to duration of encounter; ‡ Feeding clinics are publicly funded clinics inclusive of Speech Pathology and Dietetics as part of multi-disciplinary feeding-related care predominantly in the first year of life with no out-of-pocket costs when accessed through the public healthcare system

Direct travel costs

Direct out-of-pocket travel costs were calculated per trip and the approach was dependent on the estimated one-way travel time and distance by road and consistent with the state-based patient travel subsidy scheme eligibility criteria [7]. Travel was considered to occur by road using a private vehicle and costed per kilometre according to Australian government recommended rates (Table 3) for travel distances < 500 km and < 300 min each way. If travel by road would require > 500 km or > 300 min, the trip was considered to occur via accessing fully reimbursed commercial flights eligible under the patient travel subsidy scheme, which also includes an accompanying adult. Trips requiring air travel or road travel > 300 km or > 240 min each way were also considered to require overnight accommodation. Market accommodation rate parameters were derived from 13 family-suitable accommodation providers from the area surrounding the Queensland Children’s Hospital. Other relevant direct out-of-pocket costs are described in Table 3, including hospital parking (market rates), local transit (market rates), meals and incidentals for overnight trips (Australian Taxation Office rates). For families eligible for the patient travel subsidy scheme, it was assumed that the maximum permitted reimbursement amount for eligible expenses were claimed on each occasion of travel.

Table 3.

Unit costs (2024 AU$) for direct travel, including mean and standard deviation (SD)

Direct travel	Units	Out-of-pocket costs		Distribution notes
		Mean	SD
Transit private vehicle	Kilometres	0.85 per km	n/a	Sampling with replacement from cohort
Overnight accommodation	Count one per trip - >4 h or distance > 300 km	216	62	Gamma (up to 140 subsidy available per night)
Transit flights	Count if one way > 6 h	238	85	Gamma (with $150 per flight minimum)
Local ground transport if flight (e.g., taxi)	2 per flight	72	11	Gamma
Hospital car parking tertiary	Per visit (up to 7 h)	24	-	Fixed rate up to 7 h in one day
Hospital car parking secondary	Per visit (up to 7 h)	8		Fixed rate up to 7 h in one day
Car park primary	Nil out-of-pocket costs	-	-	-
Additional cost of meals and incidentals family if overnight	Count one per trip - >4 h or distance > 300 km	64	28	Gamma

Indirect productivity costs

The human capital approach was adopted to estimate indirect costs associated with lost productivity consistent with prior recommendations [8] for out-of-pocket cost estimates for families of children accessing healthcare. This approach places monetary value on time taken to travel to and attend healthcare-related activities. Productive time lost was estimated on the basis of time required to travel in each scenario, and the number of appointments required. Productivity time loss was estimated at two days per trip requiring an overnight stay (> 300 km or > 240 min each way), one day for other tertiary and secondary hospital encounters (up to 300 km or 240 min), and half a day for encounters in the primary care non-hospital encounters. To value productivity costs, day rates were derived from the Australian Bureau of Statistics Queensland earnings table [9] with a mean of AU$368 per day and assuming a gamma distribution with standard deviation of AU$72.

Modelling approach

Each microsimulation represented n = 1000 families and locations for accessing primary, secondary or tertiary care in a congenital heart disease population from diverse geographical regions of Queensland, Australia. The structure of the microsimulation involved five broad components to appropriately represent the complexity of the system being modelled for families. First, residential location characteristics were sampled with replacement from a patient cohort representative of this clinical population [2]. Second, service use requirements across relevant healthcare professions and services for the first five years of life across the four illustrative levels of neurodevelopmental support requirement were simulated based on illustrative support requirements (Table 1). Third, travel distances and times were estimated from residence locations to each of these service types based on access at the states’ tertiary paediatric centre, secondary hospitals or community based services in alignment with the three policy scenarios. Fourth, the most appropriate mode of transport (road or air) and time required to access each consultation under each scenario was determined. Lastly, resource use and costs were estimated (Tables 2 and 3) to determine out-of-pocket direct healthcare costs, direct travel costs and indirect carer productivity losses associated with accessing these neurodevelopmental services after adjustment for government funding sources and subsidies dependent on eligibility criteria.

The sampling with replacement for location of residence and service characteristics was derived from our prior study that mapped families and services across this state (maps presented in Fig. 1 adapted from the prior study [2]). The cohort sampled for location of residence included n = 957 patients, with residential locations greatly dispersed across the state and 289 (30.2%) lived outside of the urbanised South East Queensland (Fig. 1b) region. We used Geographic Information Systems software (ArcGIS online [10]) to map patients’ neighbourhood of residence, primary care, secondary care, and tertiary care services [2]. Distance and travel times were derived from address fields to estimate road travel distance (kilometres) and time (minutes) using a Google Maps Application Programming Interface (API) [2].

Fig. 1.

Geographical representation of patient residence locations and road travel time (in minutes) to: a) tertiary specialist paediatric hospital; b) secondary specialist paediatric services; c) primary care providers

This stochastic microsimulation approach was chosen due to its flexibility in representing the appropriate level of complexity across the policy scenarios, including overlapping insurance and subsidy schemes, as well as the ability to include a range of data sources. This approach has benefits in the context of comparative analyses, as it allows key impacts to be explored and policy solutions to be refined prior to large scale investment in new policy implementation. For probabilistic sensitivity analyses, uncertainty intervals were estimated by generating 10,000 iterations of the 1000 families for each of the four illustrative levels of support requirements in each of the three policy scenarios (i.e., 12 simulation models inclusive of all regions in the state). These 12 simulation models were then repeated, sampling only from families in the South East Queensland region defined by hospital and health service catchment boundaries (Fig. 1b), and then repeated again sampling from only families who lived outside of South East Queensland.

Validation of model parameterisation and sensitivity analyses

As an additional validation step for parameter estimates, 37 carers of children with congenital heart disease from across Queensland, known to the Queensland Paediatric Cardiology Service and CHD Life program [11], provided consent to be interviewed via the telephone about their experiences of out-of-pocket costs associated with travel to attend healthcare appointments. Reports from this sample fell within the expected distributions used for policy scenario model parametrisation based on their location of residence and services accessed. In addition to the probabilistic sensitivity analysis, one-way sensitivity analysis was also conducted to examine the potential influence of 25% higher or lower estimates in productive time lost per appointment, but patterns of findings remained the same (not presented).

Results

Estimates of the mean and 95% uncertainty intervals for of out-of-pocket costs in the first 5 years of life for direct healthcare, direct travel, and indirect productivity costs for families from all residential locations are presented in Table 4. Out-of-pocket costs estimates for families who reside inside and outside of the urban South East Queensland region are presented in Table 5. In summary, out-of-pocket costs for families of children who require neurodevelopmental support were substantial across all scenarios and locations of residence. It was noteworthy that productivity costs for carers required to attend healthcare consultations was the main driver of financial burden for families regardless of location of residence or policy scenario. The pattern of this finding was robust to sensitivity analyses. Direct travel costs for attending healthcare appointments were also substantial, and differentially affected those who resided in rural, regional or remote areas outside of the South East Queensland metropolitan area. With existing healthcare funding support structures, direct healthcare costs were a comparatively smaller contributor to total family financial burden relative to the travel costs and productivity losses across all scenarios. Overall, estimates of out-of-pocket costs for families of children with severe delays were approximately 8.7 to 15.4 times higher than for families of children who required only surveillance.

Table 4.

Direct and indirect out-of-pocket cost estimates (per family, all residence locations) of children with different levels of neurodevelopmental support requirements in the first 5 years of life with 95% uncertainty intervals (UI)

	Policy Scenario	Cost type	Out-of-pocket costs (2024 AUD) accrued in first 5 years of life
			General Surveillance	Mild Delays	Moderate Delays	Severe Delays
All residence locations	Tertiary Cardiac-centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	255.14 (223.28, 291.10)	3433.38 (3207.07, 3668.49)	5041.96 (4737.37, 5351.04)	6503.46 (6102.63, 6916.69)
		Indirect (Productivity)	2899.15 (2757.61, 3042.68)	16385.61 (15874.49, 16905.07)	23532.66 (22864.73, 24206.36)	29562.75 (28640.37, 30478.55)
		Total costs	3154.29 (3000.73, 3308.22)	19818.99 (19164.57, 20480.67)	28574.61 (27711.76, 29457.26)	36066.21 (34874.76, 37262.50)
	Secondary centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	261.74 (222.51, 307.17)	1301.60 (1150.68, 1471.00)	1971.91 (1756.54, 2226.26)	2518.60 (2242.35, 2846.04)
		Indirect (Productivity)	2827.23 (2690.14, 2967.51)	12802.61 (12482.02, 13125.01)	19350.66 (18938.79, 19760.84)	24481.25 (23901.30, 25059.03)
		Total costs	3088.97 (2937.52, 3241.09)	14104.20 (13738.32, 14472.38)	21322.57 (20844.64, 21804.16)	26999.85 (26338.90, 27674.19)
	Primary / community-based Mean (95%UI)	Direct (Healthcare)	33.99 (26.78, 41.50)	1327.40 (1297.40, 1356.98)	2032.02 (1997.78, 2065.20)	1168.58 (1132.15, 1204.09)
		Direct (Travel)	262.46 (222.62, 308.67)	1363.17 (1125.38, 1634.63)	2436.75 (2010.58, 2918.22)	3046.60 (2500.74, 3685.14)
		Indirect (Productivity)	2875.56 (2733.77, 3016.74)	19224.32 (18593.34, 19856.36)	35536.26 (34564.46, 36494.33)	44644.90 (43240.20, 46071.89)
		Total costs	3172.22 (3020.70, 3328.65)	21914.88 (21211.23, 22637.96)	40005.02 (38908.56, 41103.18)	48860.09 (47308.28, 50423.35)

Table 5.

Direct and indirect out-of-pocket cost estimates (per family) of children with different levels of neurodevelopmental support requirements in the first 5 years of life, for families residing inside and outside the South East Queensland urban centre with 95% uncertainty intervals (UI)

	Policy Scenario	Cost type	Out-of-pocket costs (2024 AU$) accrued in first 5 years of life
			General Surveillance	Mild Delays	Moderate Delays	Severe Delays
Families in south-east (most populated) region	Tertiary Cardiac-centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	182.97 (170.87, 194.94)	1887.62 (1814.96, 1962.85)	2785.37 (2693.01, 2880.52)	3577.91 (3447.72, 3709.76)
		Indirect (Productivity)	2889.61 (2725.91, 3056.91)	14098.61 (13670.71, 14517.12)	19973.05 (19469.84, 20472.35)	24918.94 (24217.25, 25603.09)
		Total costs	3072.58 (2899.86, 3248.67)	15986.23 (15514.07, 16453.13)	22758.42 (22209.96, 23312.05)	28496.85 (27717.19, 29269.83)
	Secondary centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	170.01 (159.22, 180.77)	909.99 (875.92, 944.79)	1378.96 (1334.44, 1425.43)	1773.58 (1709.47, 1838.03)
		Indirect (Productivity)	2842.31 (2680.66, 3006.80)	12868.61 (12487.07, 13249.01)	19445.24 (18965.24, 19928.59)	24601.55 (23913.30, 25277.75)
		Total costs	3012.32 (2843.29, 3184.87)	13778.59 (13375.62, 14179.11)	20824.20 (20321.64, 21331.57)	26375.14 (25655.98, 27081.89)
	Primary / community-based Mean (95%UI)	Direct (Healthcare)	33.98 (25.77, 43.05)	1327.56 (1291.21, 1362.59)	2032.34 (1991.54, 2072.13)	1168.95 (1125.83, 1211.33)
		Direct (Travel)	168.37 (157.48, 179.23)	654.64 (609.56, 702.96)	1105.53 (1025.79, 1191.47)	1370.95 (1264.30, 1485.24)
		Indirect (Productivity)	2889.27 (2725.63, 3056.24)	19324.28 (18590.22, 20070.59)	35708.64 (34576.96, 36860.59)	44868.72 (43238.24, 46564.08)
		Total costs	3091.63 (2918.06, 3268.14)	21306.49 (20522.23, 22095.03)	38846.51 (37677.03, 40032.66)	47408.62 (45732.91, 49145.38)
Families outside south-east region	Tertiary Cardiac-centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	430.10 (330.27, 547.72)	7174.63 (6616.70, 7746.55)	10501.21 (9823.55, 11205.10)	13582.01 (12673.21, 14546.43)
		Indirect (Productivity)	2925.15 (2659.66, 3199.42)	21923.05 (20699.86, 23177.79)	32133.63 (30587.09, 33675.77)	40787.78 (38621.43, 42957.88)
		Total costs	3355.25 (3049.74, 3665.83)	29097.67 (27566.26, 30625.33)	42634.84 (40777.95, 44463.69)	54369.79 (51695.35, 56991.84)
	Secondary centre based Mean (95%UI)	Direct (Healthcare)	Nil	Nil	Nil	Nil
		Direct (Travel)	484.98 (359.35, 635.96)	2252.35 (1772.10, 2797.32)	3408.45 (2699.29, 4224.86)	4325.03 (3410.30, 5356.67)
		Indirect (Productivity)	2794.63 (2542.26, 3051.30)	12649.39 (12024.80, 13277.78)	19123.06 (18337.04, 19919.14)	24193.89 (23085.56, 25267.00)
		Total costs	3279.61 (2976.64, 3593.92)	14901.74 (14118.90, 15697.72)	22531.50 (21509.36, 23598.59)	28518.92 (27099.14, 29975.10)
	Primary / community-based Mean (95%UI)	Direct (Healthcare)	33.90 (21.34, 48.02)	1327.22 (1270.56, 1381.91)	2032.16 (1970.44, 2094.28)	2067.68 (2329.16, 2492.29)
		Direct (Travel)	491.04 (363.62, 643.67)	3080.50 (2319.70, 3973.48)	5647.31 (4253.28, 7245.25)	7084.81 (5303.20, 9136.91)
		Indirect (Productivity)	2840.67 (2581.95, 3103.34)	18994.68 (17798.02, 20182.83)	35104.05 (33300.20, 36972.59)	44106.41 (41469.25, 46798.67)
		Total costs	3365.62 (3054.86, 3687.10)	23402.99 (21896.97, 24946.74)	42783.53 (40490.01, 45220.32)	52359.97 (49165.03, 55753.48)

The Tertiary scenario had the greatest direct travel out-of-pocket cost estimates for families at all levels of neurodevelopment support requirements due to many families having to travel substantial distances to access the specialist paediatric hospital (Tables 4 and 5), and there were also substantial indirect productivity losses associated with this. This potential burden was offset in-part by travel subsidies and potential for fewer trips for consultations when multiple healthcare professionals were able to be seen in a single visit if more than one healthcare discipline was required. Direct healthcare costs were not borne by families when accessing public hospital services.

The Secondary centre-based scenario was the most favourable overall with lowest total out-of-pocket costs across each of the illustrative levels of potential neurodevelopmental support requirements (Table 4). Lost productivity was a particularly important consideration for families residing outside of South East Queensland, contributing to greater risk of geographical inequity in out-of-pocket costs (Table 5). The Secondary scenario included no direct healthcare out-of-pocket costs for healthcare professional consultations, which were provided at publicly funded secondary hospitals in this scenario. Consistent with the Tertiary scenario, the Secondary scenario also benefited from potential access to multiple healthcare professionals during a single visit to a secondary centre if more than one healthcare discipline was required. For families of children with mild, moderate or severe delays, out-of-pocket costs were higher in the Primary scenario than the Secondary scenario. This was partly attributable to direct healthcare costs associated with accessing private healthcare providers in local communities. Large productivity loss estimates were also important in the Primary care scenario, with some families having to make separate trips for each healthcare professional appointment across private healthcare businesses rather than accessing all services at a single facility with co-ordinated scheduling.

Discussion

The financial burden of direct and indirect out-of-pocket costs for accessing neurodevelopmental services for families of children with congenital heart disease in the first five years of life was substantial in all policy scenarios. The presence of a universal healthcare system with funding structures intended to reduce inequity in financial burden for families meant direct healthcare costs were not the major contributor to out-of-pocket costs. The financial burden was instead primarily driven by travel costs and productivity losses associated with carer time-lost to travelling and attendance at healthcare provider appointments which differentially affected families required to travel greater distances to access services, particularly for families of the most neurodevelopmentally vulnerable children with severe delays. This study has made a useful contribution to the literature by robustly estimating this out-of-pocket cost burden for the first time across a range of scenarios, which has particular relevance for families of children with congenital heart disease who may be experiencing financial distress.

Despite not eliminating geographical inequity in financial burden, the Secondary scenario that supported decentralised loci of care towards locations closer to home is likely to have benefit in minimising family out-of-pocket costs, particularly for families residing outside of the South East Queensland urban centre. Families who reside in rural, regional or remote areas outside of South East Queensland would continue to bear greater burden in all scenarios, as did families with children with more severe neurodevelopmental delays. An important finding that may not be considered in many public healthcare policy scenario analyses, was the large potential impact of indirect productivity costs when carers are required to travel long distances or when many appointments are required. These findings are particularly salient for families who may fear expenses they cannot afford due to travel for care, as well as potential employment repercussions associated with missed work. It is also important to acknowledge this modelling focused on one large Australian state, and the estimates generated may not directly generalise to dissimilar regions or healthcare systems.

The family out-of-pocket cost perspective used in this study was adopted by design to explore the potential inequity in financial burden on families. However, it is noteworthy that the Secondary and Primary scenarios are likely to also offer financial benefits for the healthcare system more broadly through minimising the cost of travel subsidies paid, and the typically higher cost of care delivery in tertiary hospitals in comparison to secondary hospitals or community-based care. Other key factors beyond family financial burden and health system costs remain important for policy decisions intended to support decentralisation of neurodevelopmental support services, including inter-related aspects of workforce capacity and access, clinical effectiveness and individual consumer preferences [1, 2, 12–15]. An important policy consideration in this regard is the dependence of decentralised care on local availability and timely access to suitably skilled and experienced health professionals. Where this is not available in rural, regional or remote communities, travel may remain the only viable option for timely face-to-face service access that is consistent with current standards of care recommendations for childhood-onset heart disease [16].

An additional important policy consideration is the need for flexibility to appropriately accommodate individual preferences and needs of consumers and families which may not be based solely on financial burden and were beyond the scope of the present analyses. Individual family circumstances related to employment arrangements, paid leave, vehicle ownership, other caring responsibilities, cultural identity, and service wait times are also likely to influence consumer preferences related to healthcare access. Similarly, the level of trust and familiarity with individual healthcare providers is another important consideration that may influence consumer preferences, and this impact may be heterogenous depending on the experiences of individual families. For one family, a trusted relationship with a local healthcare provider may enhance consumer experience, adherence and effectiveness of neurodevelopmental supportive care despite that provider being less experienced with assessment and interventions for this clinical population [2, 15, 17]. For another family, they may have greater trust in the provision of care at a high-volume tertiary paediatric centre where they have familiarity arising from the acute and sub-acute cardiac care received. While individual consumer preferences of this nature, or unknown potential differences in clinical effectiveness, could not be accommodated into these high level policy scenario analyses, it is noteworthy that discrete choice experiments to examine preferences for the design and delivery of neurodevelopmental support for children highlighted that family out-of-pocket costs is an important factor influencing consumer preferences [18, 19].

Strengths and limitations

A strength of the present policy scenario analyses was being able to simulate patient populations drawn from the geographical distribution of families and services in a complex healthcare system, and incorporate a range of data sources and relevant reimbursement structures to produce illustrative estimates across a range of neurodevelopmental support requirements. The focus on out-of-pocket costs was appropriate for addressing the aims of this study, but this limited scope did not explore quality, safety, workforce capacity, consumer experience or comparative clinical effectiveness of care delivered in different locations, which are also important considerations for policy implementation in this context. Quantitative evidence in this regard is sparse, and this remains a priority for prospective clinical research with this patient population. Similarly, there remains opportunity to prospectively validate aspects of the modelling, including assumptions regarding the magnitude of realised productivity losses. This study aligned with prior recommendations in using the human capital approach for valuing productivity losses, it is possible that this approach may result in higher estimates of productivity losses than some alternatives including the friction cost approach. Private health insurance was not explicitly modelled within the scope of this study. This is likely to have had negligible effect on the findings as there were no direct healthcare costs in the Tertiary and Secondary scenarios, and private health insurance rebates would not have been required for the Primary surveillance scenario or Primary severe delays scenario where patients were National Disability Insurance Scheme eligible, and direct healthcare costs were already relatively low in the two remaining scenarios (Primary mild and moderate delays). The application of private health insurance rebates and premiums rather than the Medicare Benefits Schedule rebates in these two scenarios, where direct healthcare costs already represented only a small fraction of out-of-pocket costs, would have been unlikely to have substantially influenced study findings, interpretation or conclusions. Furthermore, additional studies examining consumer and family preferences, and the extent to which these findings are generalisable to other regions, would help clarify the applicability of policy recommendations across geographical regions and heterogenous healthcare systems.

Conclusions

Despite the presence of a supportive universal healthcare system in Australia, including generous travel subsidies and considerable funding for direct healthcare costs, out-of-pocket costs for neurodevelopmental support for families of children with congenital heart disease are substantial. These costs differentially affect families of children with more severe developmental delays, as well as those who reside in rural, regional and remote areas. Policy strategies to shift loci of services closer to home have potential to alleviate direct out-of-pocket costs associated with travel for healthcare. It was also noteworthy that indirect productivity losses for carers were substantial in all policy scenarios, particularly for families of children with more severe developmental delays and families who are experiencing financial stress.

Author contributions

Policy scenario analysis conception and primary drafting of manuscript: SMM Research program initiation and funding acquisition: SMM, BJA, KE, RJ, SK, BA Data and information acquisition: SMM, BJA, KE, SP, SS, SK, BA Policy scenario design: SMM, BJA, KE, RJ, SP, BA Critical review and manuscript editing: All authors.

Funding

This research program is funded by the Australian Medical Research Futures Fund grant #ARGCHDG0035. The funder had no role in the conduct of the work or decision to publish.

Data availability

This policy scenario analysis does not report individual patient data, but rather reports findings from microsimulations and associated estimates for the described policy scenarios, the parameters of which are reported in the contents of the manuscript tables and figures. Relevant components of raw code or instructions to replicate the content of these microsimulations will be available upon reasonable request by contacting the corresponding author.

Declarations

Ethics approval and consent to participate

Relevant human research ethics committee approvals were obtained from the Queensland Children’s Hospital human research ethics committee (HREC/21/QCHQ/79154, HREC/22/QCHQ/88866, and HREC/2024/QCHQ/106509) for this work. This included a waiver of individual consent for use of de-identifiable administrative data consistent with the ethical standards of the relevant national guidelines on human experimentation (National Health and Medical Research Council’s (NHMRC) National Statement on Ethical Conduct in Human Research (2007). Informed consent was provided by family members prior to being interviewed.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.