Cerebral palsy (CP) is one of the most common physical disabilities in childhood, occurring in approximately 1 in 700 live births in Australia.¹ It is defined as a motor disability due to a brain injury primarily affecting posture and movement and may be accompanied by sensory, communication, and cognitive impairments.² Many children with CP have comorbidities, some requiring continuing medical management.^{3, 4} Comorbidities can be classified as either co-causal (associated with underlying aetiology of the CP), a complication of CP, or co-occurring (along with the CP). Comorbidities (including co-occurring disorders) are more common in children with CP than in the general population.³

Comorbidities, particularly those with progressive consequences, are associated with increased emergency department presentations and hospital admissions.^5-7 Children with CP present to hospital with more diagnoses and with longer lengths of stay (LOS) than children without CP.^{5, 8} Hospital admissions reduce health-related quality of life for the individual and their family and are costly for the health system.^9-11 Health resources are often concentrated disproportionately on a small group of high-cost users.^{12, 13} Children with complex chronic conditions account for a large proportion of readmissions and prolonged LOS, leading to high inpatient costs.^{9-11, 14} This pattern of high-cost users is likely to be replicated in CP, given its chronicity and frequency of comorbidities. Gaining knowledge of the common reasons for hospital admissions, readmissions, and prolonged LOS is important as it enables health administrators and clinicians to ensure access to timely and appropriate health care that will reduce this burden on individuals and the health system.^{7, 8}

A significant comorbidity (complication) from CP is respiratory disease.^{3, 5, 6, 15} Complications from respiratory disease are the leading source of premature death in CP, with respiratory failure identified as the cause of approximately 59% of deaths.¹⁶ It is a major reason for emergency department presentations and hospital admissions.^{5-7, 15, 17} Repeated hospitalizations for respiratory illness are common in children with CP, with approximately half readmitted again within 1 year.¹⁷ There are several risk factors linked to respiratory hospital admissions, including potentially modifiable factors: oropharyngeal dysfunction, difficulty managing saliva, gastro-oesophageal reflux disease, respiratory symptoms with meals, frequent respiratory symptoms, seizures, and nightly snoring. By identifying which individuals with CP are at risk of respiratory illness and targeting modifiable risk factors, there is an opportunity to reduce admissions and medical costs and improve health-related quality of life through preventive strategies.^{8, 14}

Although respiratory illness is recognized as a common reason for hospitalization in CP, the economic burden for health services is unknown. The aim of this study was to establish the burden of respiratory illness in CP on the Western Australian health care system by quantifying the costs of respiratory admissions in children with CP, compared with non-respiratory admissions. Secondary aims were to establish whether there was a subgroup of health care users contributing disproportionately to hospital costs and to investigate the contribution of respiratory disease to admissions for other (non-respiratory) comorbidities.

METHOD

RESULTS

The cohort consisted of 671 individuals with CP with hospitalizations from 2015 to 2016. There was one individual with two entries in the WARDA database, which did not affect hospitalization data, but did when determining CP severity. For this duplicated entry, one had missing CP severity information, while the other had a recorded value which was therefore assigned to this individual for this study. There were missing data for CP severity for another two individuals.

Fifty-seven per cent of the cohort were male, with 31% having severe CP. The median age at hospitalization was 7 years (IQR 5–12), ranging from birth to 18 years.

There were a total of 2357 EOC, of which 1.4% involved the transfer of an individual to another hospital or a change in type of care. There were 298 individuals with 726 emergency EOC and 588 individuals with 1631 elective EOC (Tables 1 and 2). Although there were more elective EOC, emergency hospitalizations were associated with more bed-days (LOS: emergency 2807 vs elective 2648), with individuals experiencing an emergency EOC staying for longer (mean LOS: emergency 3.9 days vs elective 1.6 days). Same-day EOC made up almost half of elective hospitalizations (47%), but only 5% of all emergency days in hospital. ICU days in hospital were higher for emergency EOC (256 days) than for elective EOC (83 days).

Costs for emergency EOC (Table 1) over this 2-year period were estimated at AU$7 748 718 with respiratory the highest cost area. ‘Respiratory’ accounted for 36% of emergency costs, 35% of emergency bed-days, and the highest number of overnight and ICU days. Of the 298 individuals with CP who had an emergency EOC, 82 were admitted for a respiratory reason, of whom 68.6% had severe CP. The ‘nervous system’ was the second highest cost area with 23% of emergency costs, with other clinical areas each contributing 0.1% to 8.2% of the total emergency costs, with 45.1% of those with ‘nervous system’ hospitalizations having severe CP.

Elective EOC costs (Table 2), over this 2-year period, were less than emergency EOC, with an estimated total of AU$6 738 187. The ICD-10 AM principal diagnostic area of ‘nervous system’ was the highest cost area, with 45% of elective costs, 46% of elective days in hospital, and the highest number of same-day and overnight days. There were 819 same-day EOC observed for ‘nervous system’, with 697 (85%) of these consisting of the primary diagnosis of ‘cerebral palsy and other paralytic syndromes’ (G80–G83) combined with a botulinum neurotoxin procedure code. ‘Musculoskeletal’ was the second highest cost area with 17% of elective costs. ‘Respiratory’ had the highest number of ICU days for elective admissions but was only ranked fourth for costs and fifth for LOS.

The results from the sensitivity analysis are shown in Figure 1 and Tables S2 to S5.

For emergency EOC, the 10 most costly clinical areas remained unchanged, with the top four cost rankings consistent with the primary analysis. ‘Respiratory’ continued to be the leading cause of emergency hospitalizations, accounting for 35.5% to 36.7% of all emergency costs. For elective EOC, the top 10 most costly clinical areas also remained unchanged, with the top three rankings consistent with the primary analysis, and ‘nervous system’ remaining the lead cause of elective EOC. Respiratory hospitalizations ranked fourth among elective EOC for three of the cost models and sixth for the model where ICU admissions were bundled in with hospitalizations with overnight stays.

The 5% of highest cost users (Table 3) consisted of 34 individuals, with 397 EOC (17% of all EOC), 2409 days (48% of all bed-days), at a total cost of AU$6 767 968 (47% of all costs). Of these 34 individuals, 70.6% (n = 24) had severe CP. The most common reason for hospitalization in this group was ‘respiratory’, followed by ‘nervous system’. In this group, ‘respiratory’ contributed to 18% of all costs and 15% of all LOS (830 days), whereas ‘nervous system’ contributed to 12% of all costs and 10% of all LOS (545 days). ‘Respiratory’ and ‘nervous system’ combined in these 34 individuals contributed 30% of all costs for this 2-year period.

Table 4 compares the contribution of respiratory disease to admissions for non-respiratory reasons. For overnight stays, clinical areas with only a primary non-respiratory diagnosis had the lowest average LOS, followed by those with a non-respiratory additional diagnosis, with the highest mean LOS observed in EOC with a respiratory additional diagnosis (mean [SD] all diagnoses: 1.8 [1.4] vs 3.7 [5.3] vs 8.4 [9.6] respectively; p < 0.001). The rate of hospitalizations was more than 5 times higher in those with a respiratory additional diagnosis than in those with a primary non-respiratory diagnosis only (incidence rate ratio 5.7, 95% CI 4.6–7.0; p < 0.001). The trend of increasing mean LOS from primary non-respiratory diagnosis only, to additional non-respiratory diagnoses, to those with respiratory diagnosis was consistent across clinical areas. Clinical areas with a respiratory additional diagnosis had statistically significant higher rates of LOS than those with a primary diagnosis only (all p < 0.05), ranging from 5.2 to 7.0 times the LOS.

DISCUSSION

Although CP is not a respiratory condition, the leading cause of emergency hospitalization among this group was due to respiratory illness. ‘Nervous system’ was the second most common reason for emergency hospitalizations, with a similar number of episodes of care compared with emergency respiratory hospitalizations; however, individuals with respiratory hospitalizations tended to stay longer in hospital, leading to higher costs. In emergency hospitalizations, those with respiratory diagnoses had the highest number of days in hospital, accounting for 36% of all emergency admission costs. This relatively small group (n = 86) of individuals, many of whom had severe CP, had the highest number of days in ICU, further driving up costs.

‘Nervous system’ was the most common and costly reason for elective EOC, with a large proportion of same-day presentations for botulinum neurotoxin procedures. The second most common reason for elective admission was ‘musculoskeletal’. Admissions for botulinum neurotoxin and musculoskeletal procedures are not unexpected in children with CP as these are planned proactively to improve function and prevent complications, with evidence of their efficacy.²¹ This differs from respiratory admissions, which are frequently non-elective and reactive, and therefore lacking coordinated management.^{5-7, 17, 22, 23} Srivastava et al.²⁴ reported that children with medical complexity have fragmented, uncoordinated, and crisis-driven care, aligning with the findings of a recent audit reporting that CP respiratory admissions often lack certainty about the specific diagnosis, which can lead to costly and inefficient management.²⁵

Within this cohort, 34 individuals (5% highest cost users), of whom 71% had severe CP, accounted for almost half of both LOS and total (emergency plus elective) costs, with respiratory diagnoses being the most common reason for admission. Our findings are consistent with other studies where a relatively small number of individuals can have a large impact on overall expenditure.^{13, 26} Given the significant costs associated with a small group of individuals, there are opportunities to reduce health care costs through targeted, coordinated care, and consistent management.^{7, 8, 11, 14, 25, 27, 28}

Non-respiratory admissions were complicated by respiratory disease, with admissions associated with an additional respiratory diagnosis associated with consistently longer LOS. It is unknown whether individuals were admitted with respiratory symptoms, developed a hospital acquired infection, had an exacerbation of their respiratory disease, or aspirated while inpatients. Regardless of the primary reason for presentation, proactive respiratory support for children with CP during all inpatient stays, especially those with known CP respiratory risk factors, should become routine care.²⁷

This study focused on costs to health services, but hospitalization is also costly for patients and families, both monetarily and through reduced quality of life, especially unplanned hospital stays. During paediatric hospitalizations, families commonly participate in providing the routine care for their child, while managing work and family commitments. There are increased personal expenses, with parents incurring travel and meal costs in addition to loss of income.^{11, 29} Avoiding hospitalizations and reducing LOS would reduce health care costs, the cost to families, and improve the lives of individuals with CP.^8-11

It has been argued that reducing LOS has limited cost-saving potential for health services, as a high proportion of costs are both fixed and accrued in the first days of admission. Although some costs are fixed, managing respiratory conditions, such as pneumonias, is particularly resource intensive.³⁰ Respiratory admissions can involve costly interventions such as mechanical ventilation and expensive laboratory and consumable items.³¹ Reducing respiratory admissions and LOS would allow the reinvestment of precious health resources elsewhere.³⁰

Recently a consensus statement for preventing and managing respiratory disease in children with CP was developed, encouraging a proactive approach with early identification of respiratory risk factors, continuing surveillance, and multidisciplinary management.³² Given these findings, implementing respiratory management in hospital and community settings for high-risk individuals is supported by the present study. Coordinated care in any setting, especially for those with respiratory risk factors, will have a positive impact on health care costs and families.

A strength of this study is the use of a population-based database of individuals with CP to link CP diagnosis to hospital admission data. Studies relying on ICD codes to identify children with CP fail to capture those children with CP not assigned a CP code.⁸ Among the limitations are using LOS as the resource unit to calculate costs. This does not account for resources such as medications, emergency department costs, surgery, and ventilation. The primary aim of our study depended on the principal diagnosis. However, children are often hospitalized with multiple diagnoses, and determining the principal diagnosis is difficult in complex admissions.⁸ Diagnostic coding is also prone to errors, and codes may be assigned to areas of high reimbursement in preference to codes with lesser reimbursement.³³

CONCLUSION

Respiratory hospitalizations in CP are a significant driver of health care costs. In the paediatric age group, they are a burden for a subgroup of children with CP. In accordance with the consensus statement for preventing and managing respiratory disease, children with CP require respiratory surveillance and management of risk factors. Respiratory review should become routine with coordinated, evidence-based multidisciplinary management to reduce respiratory hospitalization.

FUNDING INFORMATION

The Perth Children's Hospital Foundation and the Cerebral Palsy Alliance.

CONFLICT OF INTEREST STATEMENT

The authors have stated that they had no interests that might be perceived as posing a conflict or bias.