Improving paediatric asthma outcomes in rural settings: clinical effectiveness of a multicomponent asthma care model

Abstract

Introduction

Asthma is a leading cause of childhood hospitalisation and emergency department (ED) presentations in Australia. Barriers persist in accessing paediatric asthma care, particularly in rural areas. We implemented and evaluated a multicomponent model to reduce preventable hospital presentations across rural New South Wales.

Methods

We conducted pre–post evaluation of children aged 5–12 years with asthma exacerbations across 13 hospitals in four rural NSW health districts. Children received: 1) a standardised discharge pack, 2) general practitioners (GP) and school linkage, 3) text reminders, and 4) nurse-led virtual home visits for uncontrolled asthma. Primary outcome was healthcare utilisation (ED presentations, admissions and GP visits) from electronic records, comparing matched pre–post periods at 1, 3, 6 and 12 months. Secondary outcomes included asthma control and quality of life at 6 and 12 months. Analyses used generalised estimating equations and mixed-effects models to calculate adjusted rate ratios (aRRs) and mean differences.

Results

188 children participated (62% male, mean age 7.6 years). ED presentations decreased 93% at 1 month (aRR 0.07, 95% confidence interval 0.04–0.12), 85% at 3 months (aRR 0.15, 0.10–0.22), 78% at 6 months (aRR 0.22, 0.16–0.30) and 69% at 12 months (aRR 0.31, 0.25–0.40). Hospital admissions decreased 97%, 95%, 90% and 81%. Uncontrolled asthma decreased from 73% to 37% (6 months) and 30% (12 months). Asthma control and quality of life improved significantly at both timepoints.

Conclusion

This multisite study achieved sustained reductions in preventable hospitalisations and meaningful improvements in asthma control and quality of life. This model offers a replicable approach for underserved populations.

Shareable abstract

A multicomponent asthma care model delivered sustained reductions in preventable hospitalisations and meaningful improvements in asthma control and quality of life for rural children. The model offers a replicable approach for underserved populations. https://bit.ly/3OZaB0S

Introduction

Asthma affects approximately 1 in 10 children worldwide and remains a leading cause of childhood hospitalisation and emergency department (ED) presentations [1, 2]. In Australia, where hospitalisation rates are highest among children under 15 years, most admissions are potentially preventable [3]. Readmissions remain common, occurring in over 20% of Australian children within 12 months of index hospitalisation, with greatest risk in the 3 months after discharge [4].

Evidence-based asthma management requires sustained use of preventative medication, comprehensive self-management education, trigger mitigation and regular follow-up to prevent hospitalisation [5]. Achieving this standard of care, however, remains challenging for many families, particularly in rural settings where children experience disproportionate asthma burden and face substantial barriers [6]. Despite being a national health priority [7], rural children have reduced primary care access, limited paediatric specialists and greater distances to healthcare services meaning that families often rely on the ED for acute asthma exacerbations [6, 8, 9]. Care in these acute-care settings focuses on immediate symptom resolution with limited capacity for ongoing preventative management [9]. These access challenges are compounded by systemic gaps in hospital discharge processes, including inconsistent asthma education, limited provision of asthma action plans (AAPs), variable follow-up arrangements and near-absent communication between hospital and general practitioners (GPs) [10, 11, 12]. Consequently, families are frequently discharged from hospital without adequate knowledge or support for effective self-management, leading to preventable repeat hospitalisations and increased disease burden [13].

Global evidence demonstrates that multicomponent asthma models of care incorporating self-management education, care coordination, and environmental assessment can reduce hospital presentations by up to 80% [14]. We previously demonstrated that a co-developed model incorporating these components reduced frequent asthma ED re-presentations by 56% in a tertiary paediatric hospital [15]. However, like our study, existing evidence is predominantly from metropolitan settings [16–18]. Given significant access and resource barriers facing rural populations, the effectiveness of such models in geographically dispersed, resource-constrained rural areas remains uncertain.

Adapting evidence-based programmes for rural contexts requires modification of delivery methods to overcome access barriers and resource constraints, while maintaining fidelity to core components [19]. Digital health technologies offer potential solutions, though their role in paediatric asthma care requires further evaluation [20]. We adapted our urban model of care [15] and implemented for rural delivery through virtual services [21], evaluating whether this approach achieved comparable reductions in healthcare utilisation and improved asthma control, health-related quality of life and medication adherence among rural children across New South Wales (NSW), Australia.

Methods

Study design and setting

We conducted a quasi-experimental pre–post clinical effectiveness study across 13 hospitals in four rural/regional NSW Local Health Districts (LHDs): Illawarra Shoalhaven, Northern NSW, Southern NSW and Western NSW (figure 1). The LHD catchments range from regional centres to very remote areas [22].

FIGURE 1.

Map of study sites showing geographic distribution across four rural and regional local health districts (LHDs) in New South Wales (NSW), Australia. Sydney is shown for reference.

Study population

Children aged 5–12 years with physician-diagnosed asthma presenting to ED or admitted to study hospitals within participating LHDs were eligible. Exclusions included complex comorbidities (e.g. significant development delay) which may affect completion of child-reported outcome measures, those under specialist respiratory care, non-English-speaking families requiring interpreters or those without permanent housing (e.g. residing in temporary accommodation). Participants were followed for 12 months post-enrolment, serving as their own control in pre–post comparisons.

All study hospitals used electronic medical records (EMRs). Participants were identified through daily research team audits of EMRs and approached in person prior to discharge whenever feasible. Families not reached in person received an initial text message sent to introduce the study and gauge participation interest, followed by a maximum of three telephone attempts within a 7-day window. Parents of eligible children provided informed consent electronically via REDcap [23]. Families not reached within this 7-day period were categorised as “unable to contact”.

Study procedure

Children aged 5–12 years old discharged from study hospitals following asthma exacerbations were provided standard care as per NSW Health guidelines [24], including a short-acting beta-agonist (SABA) weaning plan and hospital discharge letter, plus a standardised asthma discharge resource package adapted from previous work [16]. This package contained a personalised written AAP in colour [25], an educational booklet on triggers and management, a QR code linking to online educational resources, a symptom diary and standardised GP and school letters detailing best practice management tips.

Only families who received the asthma discharge resource package prior to discharge were then invited to participate in the study. Consenting participants received three additional post-discharge components (figure 2) [15, 21]: 1) Parents received a text message within 7 days of discharge containing online resource links and reminders for GP follow-up; 2) GP and school letters were proactively emailed by local staff to nominated providers; 3) children with uncontrolled asthma at enrolment, defined as a Childhood Asthma Control Test (C-ACT) or Asthma Control Test (ACT) score ≤19, were offered a nurse-led virtual home visit (VHV) within 3 months of discharge. Senior nurses across study LHDs completed standardised training provided by specialist asthma nurses in asthma education delivery and virtual environmental assessment. Using the NSW Health myVirtualcare videoconferencing platform [26], these trained nurses conducted a 30-min individual session following a standardised checklist adapted from Royal Brompton and US Environmental Protection Agency resources [21]. VHVs systematically addressed inhaler and spacer technique demonstration and verification, AAP review, virtual home environment assessment (dust mite, mould, pet dander, smoke and indoor air quality issues) [27], tailored trigger reduction strategies and reinforcement of regular GP follow-up and preventer adherence. Non-attending families were offered one additional appointment.

FIGURE 2.

Components of the multicomponent asthma care model and post-discharge delivery timeline. C-ACT: Childhood Asthma Control Test; ACT: Asthma Control Test.

Outcome measures

Demographics

Demographic characteristics including age, sex, comorbid medical conditions, language spoken at home, country of birth, parental level of education, Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD) [28] and Aboriginal and Torres Strait Islander status were collected from EMRs and baseline questionnaire. IRSAD scores were derived from residential postcodes.

Primary outcome

Healthcare utilisation

Asthma-related hospital admissions and ED visits were primarily analysed using EMR data, which allowed comparison of matched pre- and post-enrolment periods at 1, 3, 6 and 12 months. Parent-completed questionnaires, collected at baseline, 6 and 12 months after enrolment, captured ED visits, admissions and GP visits in the preceding 6 months. Parent-reported admissions and ED visits were used to corroborate EMR data.

Secondary outcomes

Asthma control

Asthma control was assessed using the ACT (≥12 years old) or C-ACT (5–11 years old). The total score range is 0–25 (ACT) and 0–27 (C-ACT), respectively, with scores ≤19 indicating uncontrolled asthma [29]. A minimal clinically important difference (MCID) of ±2 points was applied for both scales [30].

Quality of life

Quality of life was assessed using the mini Pediatric Asthma Quality of Life Questionnaire (miniPAQLQ) [31], a 13-item instrument (score 1–7, higher scores indicate less impairment). Parents reported school and work absences due to asthma with a 6-month recall, and activity limitation with a 2-week recall.

Medication adherence

The Brief Medication Questionnaire (BMQ) [32] assessed adherence barriers across three domains (score 0–3, higher scores indicating more barriers). Parents reported days of preventer and SABA in the past week, and oral corticosteroid (OCS) use over the preceding 6 months. Hospital prescription data from EMR were used for corroboration where available.

Sample size

We targeted recruitment of 270 children [21]. Our final sample of 188 children provided 80% power to detect a rate ratio of 0.85 for ED presentations rates in pre–post comparison with 5% significance (two-sided), based on Poisson calculations for repeated measures.

Statistical analysis

Descriptive statistics were presented as mean±sd for continuous variables, with median and interquartile range (IQR) reported for count data due to skewed distributions. Categorical variables were presented as n (%). All tests were two-tailed with α=0.05.

Hospital admissions and ED visits from the EMR, and GP visits from parent-completed questionnaires, were analysed as count data using Poisson regression with generalised estimating equations (GEEs) to account for within-subject correlation and clustering by LHD. Results are reported as adjusted rate ratios (aRRs) with 95% confidence intervals (CIs).

Continuous outcomes including asthma control (C-ACT/ACT) and burden of illness measures (miniPAQLQ, school/work absence, OCS use) were analysed using linear mixed-effects models with random intercepts for individual participants and LHD to account for clustering. Pairwise comparisons between timepoints (baseline versus 6 months, baseline versus 12 months, and 6 months versus 12 months) were performed using estimated marginal means with Tukey adjustment to control for multiple testing.

Binary secondary outcomes including uncontrolled asthma (C-ACT/ACT ≤19), salbutamol overuse (>2 times per week), and preventer medication use were analysed using logistic regression with GEEs. Adjusted odds ratios with 95% CIs compared 6- and 12-month timepoints to baseline.

All models were adjusted for age, sex, chronic comorbidities, recruitment year, season of recruitment and LHD.

Sensitivity analysis examined concordance between EMR and parent-reported healthcare utilisation data using Pearson correlation coefficients. EMR data were prioritised for primary outcomes due to complete ascertainment and absence of recall bias. Assessment of missing data showed no systematic bias (all p>0.20), supporting complete case analysis for EMR data and available case analysis for questionnaire-based outcomes. Analyses used R v.4.2.2 (R Foundation for Statistical Computing, Vienna, Austria).

Ethics

This study was approved by the Sydney Children's Hospital Network (2022/ETH01946) and Greater Western NSW Human Research Ethics Committees (2022/ETH00841).

Results

Demographics

We recruited 188 out of 627 eligible participants (figure 3). Children were predominantly male (n=116, 61.7%) with a mean age of 7.6±2.3 years. Baseline characteristics are presented in table 1. Recruited participants were comparable to non-recruited participants (n=439) in age, sex and geography but more socioeconomically disadvantaged (IRSAD 3.0 versus 3.5, p<0.001) and had higher baseline hospitalisation rates (61% versus 44%, p<0.001) (supplementary table S1). The most common reason for non-recruitment was “unable to contact” (n=268, 61.0%).

FIGURE 3.

Study participant flow diagram. EMR: electronic medical record.

TABLE 1.

Baseline characteristics

Characteristics (n=188)	Value
Age, years	7.6±2.3
Female	72 (38.3)
Comorbidities
Yes	44 (23.4)
No	144 (76.6)
IRSAD decile	3±1
Aboriginal and Torres Strait Islander	31 (16.5)
Recruitment year
2023	107 (56.9)
2024	81 (43.1)
Season of recruitment
Summer	36 (19.1)
Autumn	78 (41.5)
Winter	33 (17.6)
Spring	41 (21.8)
Parent sex, female	169 (93.4)
Parental age
20–29 y	22 (11.7)
30–39 y	106 (56.4)
40–49 y	47 (25.0)
50–59 y	3 (1.6)
N/A	10 (5.3)
Parental education level (n=170)
School	40 (23.5)
Higher education (diplomas and graduate certificates)	70 (41.2)
University (undergraduate and postgraduate degrees)	60 (35.3)
Language spoken at home
English	161 (86.1)
Other	27 (13.9)
C-ACT score	17.1±4.1
Uncontrolled asthma (C-ACT/ACT score ≤19)	134 (72.8)
Prior admission(s) in last 12 months	0.8±0.8
Virtual home visit
Eligible	134 (72.8)
Completed (among eligible)	62 (46.3)

Data are presented as n (%) or mean±sd. ACT: Asthma Control Test; C-ACT: Childhood Asthma Control Test; IRSAD: Index of Relative Socioeconomic Advantage and Disadvantage.

Healthcare utilisation

Pre-enrolment, participants averaged 1.6±0.9 ED visits and 0.8±0.8 hospital admissions in the previous 12 months, and 1.5±1.5 self-reported GP visits in previous 6 months. Descriptive statistics for all timepoints are presented in table 2.

TABLE 2.

Healthcare utilisation at pre- and post-enrolment

Outcome	Pre-enrolment (12 months)	1 month post	3 months post	6 months post	12 months post
ED visits
n	188	188	188	188	188
Median (IQR)	1 (1–2)	0 (0)	0 (0)	0 (0)	0 (0)
Mean±sd	1.6±0.9	0.1±0.3	0.2±0.5	0.3±0.7	0.2±0.5
Hospital admissions
n	188	188	188	188	188
Median (IQR)	1 (0–1)	0 (0)	0 (0)	0 (0)	0 (0)
Mean±sd	0.8±0.8	0.0±0.1	0.0±0.2	0.1±0.3	0.1±0.3
GP visits #
n	185			159	117
Median (IQR)	1 (0–3)			0 (0–2)	0 (0–2)
Mean±sd	1.5±1.5			0.9±1.2	0.9±1.2

ED: emergency department; GP: general practitioner; IQR: interquartile range. All sample sizes n=188 for ED presentations and hospital admissions (complete electronic medical record ascertainment). ^#: GP visits collected via parent survey at baseline, 6 and 12 months with 6 months recall; sample sizes reflect questionnaire response rates.

Healthcare utilisation decreased significantly post-enrolment (table 3). For ED visits, aRRs were 0.07 (95% CI 0.04–0.12) at 1 month, 0.15 (95% CI 0.10–0.22) at 3 months, 0.22 (95% CI 0.16–0.30) at 6 months and 0.31 (95% CI 0.25–0.40) at 12 months (all p<0.001). Hospital admissions showed similar patterns, with aRR 0.03 (95% CI 0.01–0.08), 0.05 (95% CI 0.02–0.12), 0.10 (95% CI 0.06–0.17) and 0.19 (95% CI 0.13–0.29) at 1, 3, 6 and 12 months (all p<0.001). Parent-reported GP visits at 6 months showed aRR 0.60 (95% CI 0.48–0.75, p<0.001).

TABLE 3.

Adjusted rate ratios for healthcare utilisation

Comparison	Outcome	Adjusted rate ratio (95% CI)	p-value
Pre–post comparisons
1 month pre versus post	ED visits	0.07 (0.04–0.12)	<0.001
	Admissions	0.03 (0.01–0.08)	<0.001
3 months pre versus post	ED visits	0.15 (0.10–0.22)	<0.001
	Admissions	0.05 (0.02–0.12)	<0.001
6 months pre versus post	ED visits	0.22 (0.16–0.30)	<0.001
	Admissions	0.10 (0.06–0.17)	<0.001
	GP visits	0.60 (0.48–0.75)	<0.001
12 months pre versus post	ED visits	0.31 (0.25–0.40)	<0.001
	Admissions	0.19 (0.13–0.29)	<0.001
Post comparison
6 months post versus 12 months post	ED visits	0.78 (0.51–1.19)	0.250
	Admissions	1.15 (0.62–2.16)	0.655
	GP visits	0.99 (0.78–1.27)	0.948

CI: confidence interval; ED: emergency department; GP: general practitioner. Models adjusted for age, sex, chronic comorbidities, recruitment year, season and local health district. Rate ratios <1.0 indicate reduced utilisation at follow-up.

There were no significant differences between 0–6 and 6–12 months post enrolment for ED visits (aRR 0.78, 95% CI 0.51–1.19, p=0.250), hospital admissions (aRR 1.15, 95% CI 0.62–2.16, p=0.655) or GP visits (aRR 0.99, 95% CI 0.78–1.27, p=0.948), suggesting sustained effects through the post-enrolment period.

Sensitivity analysis

EMR and parent-reported (PR) healthcare utilisation data showed moderate concordance at 6 months (ED visits: Pearson r=0.63, admissions: r=0.58, both p<0.001) and 12 months (ED: Pearson r=0.50, admissions: r=0.46, both p<0.001). PR estimates were more conservative than EMR at 6 months (ED visits: aRR 0.22 versus 0.44, admissions: aRR 0.10 versus 0.36, all p<0.001). Both data sources showed no significant differences at 12 months (EMR: ED visits p=0.25, admissions p=0.66; PR: ED p=0.02, admissions p=0.18) (table 4).

TABLE 4.

Secondary outcome measures

Outcome	Baseline mean±sd/n (%)	6 months mean±sd/n (%)	12 months mean±sd/n (%)	Adjusted effect 6 months to baseline (95% CI)	Adjusted effect 12 months to baseline (95% CI)	p-value
Asthma control
ACT score (0–25)#	15.2±5.6 n=13	21.4±5.9 n=8	20.7±4.5 n=13	6.7 (3.41–10.08)¶	7.2 (3.80–10.51)¶	0.001
C-ACT score (0–27)	17.1±4.1 n=171	20.9±4.7 n=151	20.7±4.9 n=101	3.7 (2.96–4.61)¶	3.5 (2.55–4.45)¶	<0.001
Uncontrolled, n (%)	134 of 184 (72.8)	54 of 159 (34.0)	34 of 113 (30.1)	aOR 0.20 (0.16, 0.24)+	aOR 0.16 (0.12, 0.21)+	<0.001+
Health-related QoL
miniPAQLQ score	3.9±1.5 n=183	5.4±1.4 n=156	5.6±1.3 n=112	1.42 (1.15, 1.69)¶	1.8 (1.50, 2.10)¶	<0.001
Medication adherence
BMQ total (0–3)§	0.63±0.72 n=180	0.50±0.68 n=147	0.55±0.62 n=108	−0.11 (−0.24, 0.02)¶	−0.08 (−0.22, 0.06)¶	0.267
SABA use >2/week, n (%)	122 of 145 (84.1)	37 of 145 (25.5)	29 of 145 (27.1)	aOR 0.08 (0.04, 0.13)+	aOR 0.07 (0.04, 0.13)+	<0.001+
Preventer use, n (%)	138 of 182 (75.8)	159 of 180 (88.3)	161 of 180 (89.4)	aOR 2.44 (1.67, 3.56)+	aOR 2.71 (1.84, 3.98)+	<0.001+
Other outcomes
School absence (days)	6.8±6.1 n=182	4.3±6.3 n=160	3.5±4.7 n=116	−2.58 (−3.70, −1.47)¶	−3.57 (−4.80, −2.33)¶	<0.001
Work absence (days)	4.3±5.2 n=182	2.3±4.9 n=160	2.3±3.8 n=116	−2.15 (−3.04, −1.25)¶	−2.10 (−3.11, −1.10)¶	<0.001
OCS use (days)	4.9±5.5 n=178	3.0±5.0 n=155	2.8±4.2 n=117	−2.05 (−3.13, −0.98)¶	−2.58 (−3.76, −1.41)¶	<0.001
Activity limitation (days)	4.1±3.7 n=182	1.2±2.7 n=156	1.3±3.1 n=117	−2.87 (−3.58 to −2.16)¶	−2.83 (−3.61 to −2.05)¶	<0.001

CI: confidence interval; ACT: Asthma Control Test; C-ACT: Childhood Asthma Control Test; QoL: quality of life; miniPAQLQ: Mini Pediatric Asthma Quality of Life Questionnaire; aOR: adjusted odds ratio; BMQ: Brief Medication Questionnaire; OCS: oral corticosteroid; SABA: short-acting beta-agonist. ^#: ACT sample size (n=13) is insufficient for robust inference; results should be interpreted with caution. ¶: Mean difference from linear mixed-effects model with Tukey adjustment. ⁺: Adjusted odds ratio (6 months versus baseline and 12 months versus baseline). §: BMQ range 0–3; higher scores indicate greater non-adherence risk. All models adjusted for age, sex, chronic comorbidities, recruitment year, season and local health district.

Asthma control

ACT was completed by 13 participants aged ≥12 years and improved significantly at each timepoint (table 5). The remaining 175 participants aged <12 years completed a C-ACT. C-ACT scores increased from 17.1±4.1 at baseline to 20.7±4.9 at 12 months (adjusted mean difference (aMD) 3.5, 95% CI 2.55–4.45, p<0.001), with no significant difference between 6 and 12 months (aMD −0.29, p=0.56). Uncontrolled asthma (C-ACT/ACT ≤19) decreased from 73% (134 of 184) at baseline to 30% (34 of 113) at 12 months, representing a 43-percentage point reduction (adjusted odds ratio (aOR) 0.16, 95% CI 0.12–0.21, p<0.001).

TABLE 5.

Sensitivity analysis comparing EMR and self-report data

Outcome	Time comparison	EMR (n=188)		Parent-reported#		Correlation (95% CI)¶
		aRR (95% CI)	p-value	aRR (95% CI)	p-value
ED visits
	Baseline versus 6 months	0.22 (0.16–0.30)	<0.001	0.44 (0.33–0.59)	<0.001	0.63 (0.53–0.72)
	6 months versus 12 months+	0.78 (0.61–1.19)	0.25	0.63 (0.46–0.89)	0.02	0.50 (0.35–0.63)
Hospital admissions
	Baseline versus 6 months	0.10 (0.06–0.17)	<0.001	0.36 (0.23–0.56)	<0.001	0.58 (0.46–0.67)
	6 months versus 12 months+	1.15 (0.62–2.16)	0.66	0.71 (0.45–1.16)	0.17	0.46 (0.31–0.59)

aRR: adjusted rate ratio from Poisson regression with generalised estimating equations; EMR: electronic records. All models adjusted for age, sex, chronic comorbidities, recruitment year and season of recruitment. ^#: Self-report: n=161 at 6 months, n=121 at 12 months. ^¶: Pearson correlation coefficient between EMR and self-reported data (all p<0.001). ⁺: 6 months versus 12 months comparison assesses effects in post-enrolment period (aRR≈1.0 indicates no rebound).

Quality of life

MiniPAQLQ scores improved from 3.9±1.5 at baseline to 5.6±1.3 at 12 months (aMD 1.8, 95% CI 95% CI 1.50–2.10, p≤0.001) (table 5). There was sustained improvement between 0–6 and 6–12 months (aMD 0.38, p=0.01). School absences decreased from 6.8±6.1 to 3.5±4.7 (aMD −3.57, 95% CI −4.80 to −2.33) and parental work absences from 4.3±5.2 to 2.3±3.8 (aMD −2.10, 95% CI −3.11 to −1.10) at 12 months. Activity limitation improved from 4.1±3.7 to 1.3±3.1 (aMD −2.83, 95% CI −3.61 to −2.05) at 12 months. There were no significant differences in school/work absences or activity limitation between 0–6 and 6–12 months (all p>0.05).

Medication adherence

BMQ scores showed decreasing mean scores (baseline: 0.63±0.72, 6 months: 0.50±0.68, 12 months: 0.55±0.62), though changes were not statistically significant (p>0.05). Overall, BMQ scores remained low, indicating minimal adherence barriers (table 5).

PR preventer use increased from 76% (138 of 182) at baseline to 89% (161 of 180) at 12 months (p<0.001), with significantly higher odds of preventer use reported at 12 months (aOR 2.71, 95% CI 1.84–3.98, p<0.001) compared with baseline. All baseline preventer users maintained therapy through to 12 months. Among 44 participants not using preventers at baseline, 21 (48%) initiated preventer therapy by 6 months and 23 (52%) by 12 months. Analysis of PR medication changes revealed therapy escalation in 27% (49 of 180) and de-escalation occurred in 6% (11 of 180) based on GINA treatment steps [5].

At enrolment, 84% (122 of 145) of participants reported frequent SABA use (>2 times/week), decreasing significantly to 27% (29 of 107) at 12 months (aOR 0.07, 95% CI 0.04–0.13, p<0.001). OCS use decreased from 4.9±5.5 at baseline to 2.8±4.2 at 12 months (aMD −2.58, 95% CI −3.76 to −1.41, p<0.001), with no significant difference between 0–6 and 6–12 months (aMD −0.51, p=0.39). These improvements paralleled reductions in uncontrolled asthma and improvements in C-ACT scores.

Discussion

This multisite study demonstrated that comprehensive evidence-based asthma care incorporating standardised discharge resources and post-discharge follow-up achieved substantial improvement in asthma health outcomes for rural children. These findings are particularly important given that improving health outcomes for children is a national priority [7].

Our model resulted in significant reductions in healthcare utilisation, with hospital admissions and ED presentations decreasing at all measured time points. This is particularly notable as Australia has one of the highest rates of asthma admissions globally [33], most of which are potentially preventable [3]. Readmission risk is highest within 28 days of index hospitalisation [4], yet our model achieved a 97% decrease in admissions during this critical window, with sustained reductions at 12 months (69% fewer ED visits, 81% fewer hospital admissions). Our results are comparable with other multicomponent urban-based studies that have demonstrated 29–85% reductions in healthcare utilisation [16–18]. This may reflect similar core elements of education, action plans and environmental mitigation. It is difficult to know which components made the biggest difference; however, consistency of benefits across studies suggests this combination of education, AAPs and environmental assessment is effective. Importantly, our outcomes were achieved across rural populations, settings where multicomponent asthma models have rarely been implemented or evaluated. The parallel improvements seen across healthcare utilisation, asthma control, quality of life and functional outcomes further strengthens confidence that our model addressed fundamental management gaps.

Beyond healthcare utilisation, our model achieved meaningful clinical impact. Asthma control improvements surpassed the MCID [30], with the proportion of children experiencing uncontrolled asthma decreasing substantially from baseline through to 12 months post-enrolment. Quality of life improvements similarly exceeded thresholds of meaningful change [31], translating to reduced school and parental work absences which has important economic implications. Our predominantly socioeconomically disadvantaged rural population experiences high asthma burden and reduced healthcare access [6, 34]. The success of this adapted model across populations experiencing established barriers to asthma care, demonstrates that appropriate adaptation and technology-enabled delivery can address rural–urban health disparities in paediatric asthma care.

Medication use patterns revealed important insights. At baseline, most children were prescribed preventers (76%) and self-reported adherence barriers remained minimal throughout the study, yet 73% had uncontrolled asthma. This differs from much of the literature, which identifies low preventer use as a primary driver of poor control [35]. Despite preventer prescriptions, SABA overuse, a recognised indicator of poor asthma control and predictor of acute exacerbations [36], was highly prevalent at enrolment (84%) and decreased significantly to 27% at 12 months post-enrolment. The persistently low rates at both follow-up timepoints suggest the model was associated with sustained improvements rather than simply reflecting natural recovery from exacerbation. Combined with stable preventer use and C-ACT improvements, these findings suggest children achieved and maintained better symptom control. The high baseline SABA use despite preventer prescriptions may reflect both the acute exacerbation and underlying factors including suboptimal preventer dosing (27% escalated), poor inhaler technique, inadequate medication understanding, adherence gaps despite self-report, uncontrolled triggers or a combination of these factors [37]. We did not systematically measure these factors, limiting our ability to identify specific mechanisms. However, our model's education emphasis through booklets, action plans and multimedia resources may have addressed knowledge and skill gaps, contributing to the observed reductions in SABA use.

Structured education and personalised AAPs are cornerstones of paediatric asthma care [38], yet AAP provision rates remain low in routine practice [3], particularly in rural settings where additional barriers like time constraints and limited specialist input contribute to suboptimal care [6]. Our standardised, technology-enabled approach demonstrates this evidence-based care model can achieve comparable outcomes in resource-constrained rural settings, offering a replicable model for reducing preventable asthma hospitalisations in underserved populations.

This clinical effectiveness evaluation forms part of a broader mixed-methods programme including implementation [39, 40] and an economic evaluation (under review), with findings collectively informing the pathway toward systematic, routine delivery across diverse settings.

Limitations

Several limitations warrant consideration. The quasi-experimental pre–post study design without concurrent control group limits causal inference and cannot exclude secular trends or regression to the mean, though recruitment over 15 months with seasonal adjustment and sustained improvements across outcomes strengthen confidence in our model's effects. Of 627 eligible participants, 188 (30%) were recruited. Recruitment challenges inherent to busy rural hospitals including high locum staff turnover in some sites and rapid patient flow contributed to non-recruitment [39, 40]. While recruited participants were demographically representative, they had higher baseline hospitalisation (61% versus 44% with prior admission, p<0.001), potentially limiting generalisability to children with milder asthma (supplementary table S1). Exclusion of children with medical, language and social complexity further limits generalisability. Questionnaire attrition (38%) at 12 months reduced power for some secondary outcomes, though missing data analysis showed no systematic bias. Small sample size for ACT (n=13, participants ≥12 years) limits inference for this age group. The 12-month follow-up may be insufficient to assess long-term sustainability, and geographic scope was limited to four NSW health districts. Whether these findings translate to non-rural settings or healthcare systems with different infrastructure requires evaluation.

Conclusion

This multisite study demonstrates that technology-enabled multicomponent asthma care can achieve sustained reductions in preventable hospitalisations and meaningful improvements in asthma control and quality of life in rural settings, with outcomes comparable to urban programmes. This model offers a replicable approach for addressing rural–urban health disparities in paediatric asthma care.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

00389-2026.SUPPLEMENT