Comparative costs of hospitalisation among infants at high risk for respiratory syncytial virus lower respiratory tract infection during the first year of life

Abstract

Objective: This retrospective cohort study compared the total cost of hospitalisation due to respiratory syncytial virus (RSV) lower respiratory tract infection (LRI) during the first year of life between late-preterm (33–36 weeks gestational age [wGA]) and term (≥37 wGA) infants.

Research design and methods: A large national claims database of commercially insured members was examined to identify hospital admissions associated with RSV between January 2003 and June 2007 among infants at high risk for RSV LRI, including late-preterm infants. Hospital use and costs were compared with those of a reference cohort of term infants with RSV.

Results: The cost of hospitalisation for RSV among late-preterm infants with at least one hospital admission associated with RSV (n=173) was twice that of term infants (n=1,983; $20,269 vs. 9,635; p< 0.001). The mean length of stay was also higher (5.3 vs. 3.4 days; p< 0.001). Approximately 21.9% of hospitalisations for late-preterm infants included an intensive care unit admission compared with 9.6% among term infants (p< 0.001).

Limitations: Reliance on ICD-9 codes to identify potential cohort members may result in misclassification and underreporting the cohort size for conditions of interest.

Conclusions: Hospitalisation costs and length of stay due to RSV LRI were significantly greater among late-preterm infants compared with term infants and higher than general estimates previously reported in the broader paediatric population.

Key words::

• cost analysis
• hospitalisation
• length of stay
• premature infant
• respiratory
• syncytial virus infections
• viral bronchiolitis

Introduction

Respiratory syncytial virus (RSV) is the leading cause of hospitalisation in infants younger than 1 year of age¹. The mean cost per case of hospitalisation per child or infant with RSV has been reported to be $3,400–29,000, with increased costs for infants with lower gestational age, admission to an intensive care unit (ICU), use of mechanical ventilation, or the presence of chronic conditions^2–5. Costs among several populations at high risk for developing complications of RSV infection, including infants with congenital heart disease (CHD), infants with chronic lung disease of prematurity (CLD), and particularly late-preterm infants (34 to < 37 wGA, who represent nearly three-fourths of all preterm infants), remain poorly defined⁶. Late-preterm infants demonstrate varied physiological and anatomic immaturity^7,8, which can lead to greater morbidity, rehospitalisation rates, and total healthcare costs during the first year of life compared with term infants^9–11. In a study of hospital resource use among RSV-infected infants, late-preterm infants had increased hospital stays, ICU admissions, and intubation rates than term infants¹². However, costs associated with healthcare use among RSV-infected infants across gestational ages were not examined.

The average cost of hospitalisation for RSV lower respiratory tract illness (LRI) during the first year of life among late-preterm compared with term infants is reported in this descriptive study. Understanding the healthcare needs and costs among infant populations at high risk for developing RSV LRI (including those with CHD, CLD, and premature infants) is the first step in optimising their management.

Methods

This was a retrospective, non-interventional cohort study involving analysis of claims data from The MarketScan Research Database (Thomson Reuters, New York, NY, USA). This national database contains person-level enrolment and medical claims from a geographically diverse, commercially insured population of approximately 11 million insured individuals per year. The database is de-identified and compliant with the Health Insurance Portability and Accountability Act. The data were used to provide information in aggregate form and claims cannot be linked back to any specific subject to provide individual health information, thus the study did not require institutional review board review and is exempt per guidance provided by the US Department of Health and Human Services^13,14.

Records from all infants younger than 1 year of age were examined and admissions associated with RSV during their first RSV season based on diagnosis codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM; Table 1) were identified. Hospitalisations for RSV were identified using diagnosis codes for RSV (079.6), acute bronchiolitis due to RSV (466.11), or pneumonia due to RSV (480.1) across four RSV seasons (from January 2003 to June 2007).

Table 1. List of conditions and corresponding ICD-9-CM or DRG codes.

Description of condition	ICD-9-CM or DRG code
Respiratory syncytial virus	079.6
Acute bronchiolitis due to respiratory syncytial virus	466.11
Pneumonia due to respiratory syncytial virus	480.1
Congenital heart disease	425.4, 745.0, 745.4–6, 746.0, 746.3, 747.0, 747.1, 745.1–3, 746.1, 746.2, 746.7, 747.4
Chronic respiratory disease arising in the perinatal period	770.7
Early-preterm	Defined by a DRG code indicating other than term status (389) and an ICD-9-CM diagnosis code specifying < 33 wGA on the birth claim (765.21–765.26)
Late-preterm	Defined by a DRG indicating other than term status (389) and an ICD-9-CM diagnosis code specifying 33–36 wGA on the birth claim (765.27–765.28)
Term	Defined by a DRG for other term (389) or normal newborn (391) on the observed birth admission claim and no diagnosis codes indicating gestational ages < 37 wGA, or DRG for neonate with other significant problems (390) and a diagnosis code indicating gestational age .37 wGA (765.29)

DRG, diagnostic related group; ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification; wGA, weeks gestational age.

Four high-risk cohorts were examined: infants with any diagnosis codes related to CHD, infants with a diagnosis code for CLD, and early-preterm (≤32 wGA) and late-preterm (33–36 wGA) infants. Whereas the various gestational age cohorts are mutually exclusive, the CHD and CLD cohorts were not exclusive with respect to gestational age. For example, an infant with CLD or CHD may also have been included in the early- or late-preterm cohorts. Term infants (≥37 wGA) were considered low-risk and were the reference group.

Descriptive statistics for the average cost of hospitalisation for RSV LRI, length of stay and other hospital use outcomes, age at hospitalisation, sex, and comorbid conditions, including congenital/perinatal diseases of the cardiovascular, respiratory, digestive, musculoskeletal, and nervous systems were prepared using SAS statistical software version 9.1 (SAS Institute, Cary, NC, USA). For continuous variables, the mean (SD) and median are presented; percentages are presented for categorical variables. The average cost of hospitalisation per infant was determined. The cost data reflect the allowed amount for the hospital and professional components of inpatient treatment adjusted to July 2008 using the hospital inpatient services items of the Consumer Price Index for Urban Wage Earners and Clerical Workers (seasonally adjusted). Allowed costs represent the total compensation the providers expect to receive for the services performed and are not charges.

Outcomes of interest for each of the four high-risk cohorts were compared with the term cohort. Analyses were performed using a distribution-free Wilcoxon test for the continuous variables (α=0.05) and a two-sided t-test for the categorical variables (α=0.05).

Results

All first-season hospital admissions for RSV LRI among infants with CHD or CLD, and early-preterm, late-preterm, and term infants were identified (Table 2). Early-preterm infants and those with CHD or CLD were generally 1 month older (approximately 4 months) at the time of the first hospital admission compared with term infants (p< 0.01).

Table 2. Patient characteristics, hospitalisation costs, and resource use among high-risk compared with term infants diagnosed with respiratory syncytial virus during 2003–2007.

Characteristic	Infants at high risk for hospitalisation due to RSV				Reference group
	CHD	CLD	Early-preterm, ≤32 wGA	Late-preterm, 33–36 wGA	Term, ≥37 wGA
Number, n	151	41	40	173	1,983
Mean (SD) age at start of RSV hospitalisation, days	118 (81)*	127 (73)^†	121 (66)^†	99 (77)	95 (73)
Male infants, %	60.3	63.4	70.0	70.5^†	58.6
CLD or other respiratory comorbid conditions^‡, %	9.3*	100.0*	42.5*	2.9*	0.2
CHD or other cardiovascular comorbid conditions, %	100.0*	58.5*	47.5*	17.3*	3.4
All other comorbid conditions, %	23.8*	19.5*	25.0*	9.2*	1.0
Mean (SD) seasonal cost of hospitalisation per infant^§, USD	27,661 (59,642)*	25,119 (49,248)*	15,693 (16,633)*	20,269 (37,449)*	9,635 (28,419)
Median, USD	8,386*	9,623^†	8,018^¶	6,842^†	5,364
Interquartile range, USD	13,254	21,601	16,163	14,066	5,073
Mean (SD) seasonal total length of hospitalisation per infant, days	6.7 (8.6)*	7.0 (9.5)*	4.9 (3.4)*	5.3 (5.3)*	3.4 (3.4)
Average RSV hospitalisations per infant, n	1.13*	1.10	1.03	1.06	1.05
ICU use per hospitalisation
Admissions with any ICU   stay, %	21.8*	13.3	19.5^¶	21.9*	9.6
Mean (SD) time in ICU, days	4.5 (4.8)	7.2 (4.3)^¶	5.4 (3.2)^¶	5.7 (4.8)*	3.4 (4.4)
Mean (SD) total time   hospitalised, days	9.8 (7.7)^†	8.7 (2.9)^¶	7.0 (2.6)	10.4 (5.8)*	6.4 (5.9)
Any cause rehospitalisation ≤30 days after initial RSV hospitalisation, %	15.2*	19.5*	15.0*	5.2	3.2

One admission for $695,227 associated with an early-preterm infant with CLD was identified as an outlier and excluded from all calculations. Pairwise comparisons include estimates from each of the high-risk groups with the term group.

CHD, congenital heart disease; CLD, chronic lung disease of prematurity; ICU, intensive care unit; RSV, respiratory syncytial virus; USD, US dollars; wGA, weeks gestational age.

*p< 0.001; ^†p< 0.01; ^‡Comorbid conditions include diseases of digestive, musculoskeletal, and nervous systems; ^§Cost of hospitalisations reflects allowed costs adjusted to July 2008 by CPI-Medical (CWSR0000SS5702); ^¶p< 0.05.

Respiratory (42.5%), cardiovascular (47.5%), and other comorbid conditions (25.0%) were significantly more prevalent among early-preterm compared with term infants (each < 4%; p< 0.001) and late-preterm infants were more likely to have cardiovascular (17.3%), respiratory (2.9%), and other types of comorbid conditions (9.2%) than term infants (p< 0.001); however, the majority of late-preterm infants were not found to have any underlying comorbidities. More than half of the infants with CLD (58.5%) had cardiovascular conditions. Coexisting digestive, muscular, and neurological conditions were also observed among infants with CLD (19.5%) or CHD (23.8%).

The cost of hospitalisation for RSV LRI among late-preterm infants was twice that of term infants ($20,269 vs. 9,635; p< 0.001). Late-preterm infants also had significantly longer mean durations of hospitalisation than term infants (5.3 vs. 3.4 days; p< 0.001) and 21.9% of hospitalisations of late-preterm infants required ICU admission vs. 9.6% for term infants (p< 0.001).

Infants with CHD had the highest cost associated with admission for RSV LRI among all high-risk groups ($27,661) followed by infants with CLD ($25,119). Among infants hospitalised for RSV, those with CHD had a significantly higher number of admissions associated with RSV during their first season compared with term infants (1.13 vs. 1.05; p< 0.001). The length of hospital stay for each of the high-risk cohorts (4.9–7.0 days) was significantly greater than for term infants (3.4 days; p< 0.001). Early-preterm infants and those with CHD or CLD were at significantly greater risk for readmission within ≤30 days of discharge than term infants. Infants with CHD and early- and late-preterm infants were also twice as likely to have hospitalisations that included admission to the ICU than term infants (p< 0.05).

Discussion

The Institute of Medicine of the National Academies, Washington, DC, has called for additional research on the economic, acute, and long-term consequences of prematurity¹⁵. This descriptive study supports the Institute of Medicine's emphasis on the vulnerability of late-preterm infants by demonstrating the increased costs associated with acquiring more severe RSV disease and the disproportionate use of healthcare resources in late-preterm infants with RSV LRI compared with term infants. RSV LRI results in costly hospitalisations in the first year of life of late-preterm infants, who generally were not thought to have an appreciably elevated risk of more complicated and severe respiratory tract infections. A study by Stewart et al¹⁶ has shown that total first-year healthcare costs for infants 33–36 wGA with RSV hospitalisation can be approximately 4 times higher compared with infants 33–36 wGA without RSV hospitalisation ($31,408 vs. 8,263, respectively). Because the majority of late-preterm infants did not have underlying comorbid conditions, this study suggests gestational age may play an important role in increased costs and resource use by late-preterm infants compared with term infants, a finding supported by a study by Gilbert et al¹⁰ in which hospital costs remained substantial even in infants 34–36 wGA with low neonatal morbidity. Willson et al¹⁷ showed that mean hospitalisation costs and length of stay were highest in infants diagnosed with bronchiolitis 33–35 wGA compared with infants ≤32 wGA and those ≥36 wGA.

The hospitalisation costs determined in this study are generally higher than costs reported in the literature^{3–5,17–23}. However, this study advances previous work in the field by providing allowed amounts (not charges^2,24–28), recent cost data, and mean costs stratified by high-risk groups. Providing cost data among the generally low-risk population of children infected with RSV younger than 5 years^5,19 may underestimate costs for high-risk groups^21–23,29. One study reported an average hospitalisation cost per stay of $3,355 for infants 32–35 wGA, but this population deliberately excluded infants with CHD or CLD³. Willson et al¹⁷ showed a mean cost of $14,137 in infants 33–35 wGA, but this population included all infants diagnosed with bronchiolitis. The length of stay shown in this study is comparable to those reported elsewhere, depending on population examined (2–8.5 days)^{1–5,12,16–19,26,28,30}. Reports of cost and length of stay vary widely from country to country^31–46 and may rely on too many confounding factors (e.g. different health delivery systems) to allow relevant comparison.

This study was intended to provide a description of the hospital costs due to RSV among infants of various gestational age and the results presented are not controlled for confounding factors. The use of the claims database limits the definition of late-preterm in this study to the ICD-9 coding range of 33–36 wGA (codes 765.27–765.28). In addition, reliance on ICD-9 codes to identify potential cohort members may result in misclassification and consequently underreporting the cohort size for conditions of interest. Although this study did not find a difference in rehospitalisation rates within ≤30 days of discharge after RSV admission between late-preterm and term infants, other relevant factors, such as severity of disease, socioeconomic factors, or outpatient use were not examined. The findings of this study also do not reflect local standards of care and admission practices independent of disease severity for preterm infants, which may vary geographically. The study sample consisted of a commercially insured population, and additional information on healthcare costs is needed among RSV-infected infants insured by Medicaid. Finally, the long-term economic burden beyond that associated with the initial hospital admission for RSV LRI awaits future studies.

Conclusions

The average cost of hospitalisation with RSV LRI among late-preterm infants was twice that of term infants and within the range of estimates of other high-risk groups. Furthermore, the costs of hospitalisation for RSV LRI incurred by each high-risk cohort were higher than those reported in broader populations of young infants and children with RSV^5,24. Late-preterm infants comprise approximately three-fourths of all prematurely born infants, and their number is rapidly increasing⁶. They will continue to represent a significant portion of patients in hospitals and critical care settings and will consume a disproportionate share of healthcare resources among young infants. Additional information characterising the susceptibility of this group of preterm infants is needed to relieve their major healthcare burden.

Transparency

Declaration of funding: This study was funded by MedImmune, Gaithersburg, MD, USA.

Declaration of financial/other relationships: M.L.F., C.B.H. and A.J. have disclosed that they have served as consultants for MedImmune; C.B.H. has also received grant support from MedImmune. A.S.M. and P.J.M. have disclosed that they are employees of MedImmune.

The JME peer reviewer 1 and 2 have not received an honorarium for their review work on this manuscript. Peer reviewer 1 has disclosed that he is employed by and owns stock in Abbott Labs, Abbott Park, IL USA; reviewer 2 has disclosed that she has no relevant financial relationships.