ABSTRACT
Objective: To evaluate the prevalence of chronic respiratory morbidity (CRM) in preterm infants (born ≤28 weeks gestational age (GA)) and compare healthcare resource utilization and costs among infants with/without CRM, and with/without bronchopulmonary dysplasia (BPD).
Methods: Commercial claims data from the Truven MarketScan database were retrospectively analyzed. Included infants were born ≤28 weeks GA and admitted to a neonatal intensive care unit (January 2009–June 2016). Continuous insurance eligibility was required from birth through 1 year (CRM/no CRM cohorts) or ≥3 months (BPD/no BPD cohorts) CA or death.
Results: CRM analysis included 1782 infants; 29.0% had CRM. BPD analysis included 2805 infants; 61.1% had BPD. The mean birth hospital length of stay was longer in infants with CRM versus those with no CRM (p < 0.0001). In infants with CRM or BPD, hospital readmission rates were significantly increased versus those without (both p < 0.0001). Total health care costs were significantly higher in infants with CRM (p = 0.0488) and BPD (p < 0.0001) versus those without. After birth hospitalization, outpatient visits and hospital readmissions accounted for most of the costs for the CRM and BPD cohorts.
Conclusion: CRM and BPD following extremely preterm birth impose a significant health care burden.
1. Introduction
An estimated 15 million infants are born preterm (at <37 weeks gestational age (GA)) each year (based on 2010 data), of whom 0.78 million were born extremely preterm (<28 weeks GA) [Citation1]. Premature birth has a negative effect on postnatal lung development, and infants born extremely preterm are at high risk of developing chronic respiratory morbidity (CRM) [Citation2,Citation3].
Bronchopulmonary dysplasia (BPD) is a chronic lung disease of preterm infants characterized by early lung injury that results in a need for supplemental oxygen [Citation4]. BPD is the most common complication of extremely preterm birth [Citation5,Citation6] and can be an early indicator of CRM [Citation2,Citation4]. Infants with BPD remain at risk for impaired lung or respiratory function after discharge from the neonatal intensive care unit (NICU).
CRM typically manifests as recurrent respiratory symptoms requiring treatment with pulmonary medications such as bronchodilators, a need for supplemental oxygen, and frequent hospitalizations and physician visits, especially during the first 2 years of life [Citation2,Citation7]. In the Prematurity and Respiratory Outcomes Program (PROP) study, 46–69% of extremely preterm infants had persistent respiratory morbidity at 1 year [Citation2]. Respiratory morbidity in that study was based on assessment of hospitalizations, home respiratory support, respiratory symptoms, and/or administration of respiratory medications. Persistent lung function abnormalities may also increase the risk of chronic obstructive pulmonary disease as the infants age [Citation8,Citation9]. Although CRM may result from having BPD, CRM can be seen in children not previously diagnosed with BPD. CRM, especially in children with BPD, imposes additional burden on health care systems with increased use of direct and indirect medical care [Citation10–15]. Published studies have shown that healthcare costs among preterm infants with BPD are substantial [Citation10,Citation16,Citation17]. Literature on the resource utilization and NICU costs for extremely preterm infants with CRM are currently limited.
In this study, we used data from a US insurance commercial claims database (MarketScan) to evaluate the prevalence of CRM among extremely preterm infants through 12 months corrected age (CA). In addition, we compared birth hospital length of stay, hospital readmissions, and health care costs among infants with or without CRM, and among infants with or without BPD.
2. Methods
2.1. Study design and data source
This was a retrospective analysis of the Truven Health MarketScan claims database.
The database contains health insurance claims and electronic medical records of ~61 million employees, dependents, and retirees with primary or Medicare supplemental coverage across the United States. The database includes enrollment history and claims for medical (provider and institutional) and pharmacy services, with inpatient services captured at both the claim and summarized stay level. The database contains actual payer costs, and patients’ out-of-pocket expenses, including deductibles, co-pays, and co-insurance. The data are considered representative of all census regions in the United States, predominantly the South and North Central (Midwest) regions.
The study comprised four cohorts: infants with CRM, infants without CRM, infants with BPD, and infants without BPD. Infants eligible for inclusion were born at ≤28 weeks GA and admitted to an intensive care unit (ICU) or NICU within 1 day after birth between January 2009 and June 2016. Admission to the ICU or NICU was identified using hospital standard revenue codes. Infants who survived to 36 weeks postmenstrual age (GA + chronological age) were eligible for inclusion. Infants with a diagnosis of congenital heart disease, congenital diaphragmatic hernia, or other congenital malformations were excluded in both the CRM/no CRM and BPD/no BPD cohorts. However, infants with a specific diagnosis of patent ductus arteriosus or patent foramen ovale were not excluded, as these conditions are highly prevalent among preterm infants and the infants have otherwise structurally normal hearts. Diagnoses of BPD and CRM were not mutually exclusive.
Health care costs were based on payment to the insurance company and were directly provided in the claims data. Costs were annualized and inflated to 2017 USD using the US Medical Service Consumer Price Index. The follow-up time used to calculate costs for infants with complications (CRM or BPD) and without BPD or CRM was from birth to 1 year CA. All analyses used SAS Enterprise Guide 7.1 (SAS Institute Inc, Cary, NC, USA).
2.2. CRM/no CRM cohorts
CRM was defined as one or more of the following events in at least two quarters (3-month time intervals) from discharge of birth hospitalization through 1 year CA: hospital readmission or emergency room (ER) visit due to a respiratory cause, use of at least one respiratory medication for ≥30 days (defined as having been prescribed ≥30 days of ≥1 respiratory medication) or use of home respiratory/oxygen support. CA represents the age of the child from the expected date of delivery and was calculated by the following formula: age ‒ ((40 weeks – GA in weeks) × 7)/365. Health care costs associated with birth hospitalization, outpatient visits, ER visits, and hospital readmissions were evaluated for infants with CRM versus no CRM. All infants included in the CRM/no CRM group had continuous insurance eligibility from birth to 1 year CA or death, whichever happened first.
2.3. BPD/no BPD cohorts
BPD was identified by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) or ICD-10-CM diagnosis codes (which included chronic lung disease) during the period from birth to death or end of continuous eligibility, whichever occurred first. The codes used to identify BPD were ICD-9-CM 770.7 and ICD-10-CM P27.1. Severity categorization for BPD was not available. Health care costs associated with birth hospitalization, outpatient visits, ER visits, and hospital readmissions were evaluated for infants with BPD versus no BPD. The BPD/no BPD group had continuous insurance eligibility from birth through at least 3 months CA or death, whichever occurred first. Outcomes were evaluated utilizing a person/time approach due to variable length of follow-up.
2.4. Statistical analyses
Comparisons of patient characteristics/complications between the CRM/no CRM and BPD/no BPD cohorts were conducted using the Wilcoxon rank-sum tests for continuous variables and Chi-squared tests or Fisher’s exact tests for categorical variables. Deaths were not included in the analysis. Generalized linear models adjusted for patients’ characteristics and complications with a negative binomial distribution and log link were used to compare number of health care resource utilization events and length of stay between cohorts.
Logistic regression models, adjusted for patients’ characteristics and complications, and offset for follow-up time, were used to compare the presence of health care resource utilization events between cohorts. Adjusted generalized linear models with a Tweedie distribution and log link were used to compare health care costs between cohorts.
3. Results
3.1. Patient population
The MarketScan database included 112,187 premature infants (<37 weeks GA) born between January 2009 and June 2016. Of these, 44,751 were admitted to an ICU or NICU within 1 day after birth based on hospital revenue codes. Overall, 2805 eligible infants were born at ≤28 weeks GA and were included in the current analysis ().
Figure 1. Sample selection flowchart of the Truven MarketScan database
3.2. CRM/no CRM analysis
For the CRM/no CRM analysis, 1782 eligible infants, with continuous eligibility from birth to over 1 year CA or death, were born at ≤28 weeks GA. Of these, 1265 (71.0%) did not have CRM, and 517 (29.0%) had CRM. The prevalence of CRM within 1 year CA increased with decreasing GA (27–28 weeks, 18.9%; 25–26 weeks, 34.4%; 24 weeks, 49.0%; ≤23 weeks, 51.3%). BPD was present in 86.7% of infants with CRM and 53.3% of infants without CRM (). The percentage of infants experiencing CRM within 1 year CA was 29.0% (n = 517/1782) if ER visits were included in the definition, and 27.7% (n = 494/1782) if ER visits were not included in the definition.
Table 1. Patient characteristics during birth hospitalization
Use of ≥1 respiratory medication for ≥30 days in the 1-year CA group was the most common CRM event (80%), followed by home respiratory/oxygen support (57%) and rehospitalizations or ER visits (41%). Among infants with CRM receiving respiratory medications, 53.2% used inhaled steroids, 33.5% inhaled bronchodilators, 30% diuretics, 6.0% methylxanthines, 5.0% leukotriene receptor antagonists, 2.9% systemic corticosteroids, and 1.0% pulmonary vasodilators. The frequency of CRM events varied by quarter through 1 year CA; use of home respiratory/oxygen support decreased substantially over time, while rehospitalization and ER visits decreased gradually and use of respiratory medication remained stable (). Significantly higher rates of complications (intraventricular hemorrhage, retinopathy of prematurity, patent ductus arteriosus, and necrotizing enterocolitis) were observed among infants who experienced CRM compared with infants without CRM.
Figure 2. Prevalence of events by quarter among 517 infants with chronic respiratory morbidity (CRM) who experienced ≥2 events through 1 year corrected age
Infants with CRM had a significantly longer mean length of stay during the birth hospitalization, compared with infants without CRM (98 ± 33 vs. 78 ± 31 days; p < 0.0001; ). In addition, hospital readmission rates after NICU discharge, for all causes as well as for respiratory causes specifically, were significantly higher for infants with CRM versus those without (all cause, 40.4% [209/517] vs. 14.2% [180/1265]; respiratory cause, 30.9% [160/517] vs. 6.5% [82/1265]; both p < 0.0001; ).
Table 2. Length of stay during birth hospitalization, and hospital readmissions after discharge through 1 year corrected age, among extremely preterm infants in the CRM/no CRM cohorts
All health care costs were substantially higher for infants with CRM versus those without (). The total health care cost including birth hospitalization was 813,759 USD ± 802,648 for infants with CRM and 586,294 USD ± 597,315 for infants without CRM (with CRM vs. without, p = 0.0488). After the birth hospitalization, which accounted for the highest costs (762,350 USD ± 783,287 vs. 560,782 USD ± 591,848; p-value not significant), the largest drivers of increased costs (all causes) were outpatient visits (27,395 ± 25,343 USD vs. 18,207 USD ± 16,521; p < 0.0001) and hospital readmissions (21,956 ± 82,629 USD vs. 6597 USD ± 43,912; p < 0.0001). Increases in costs from respiratory causes were due to hospital readmissions (13,939 ± 53,733 USD vs. 2443 USD ± 32,504; p < 0.0001), ER visits (1318 ± 2992 USD vs. 281 USD ± 983; p < 0.0001), and outpatient visits (3907 ± 13,650 USD vs. 1276 USD ± 6535; p < 0.0001).
Figure 3. Comparison of total health care costs by presence of chronic respiratory morbidity (CRM) through 1 year corrected age in extremely preterm infants
3.3. BPD/no BPD analysis
Overall, 2805 infants were eligible for inclusion in the BPD/no BPD analysis. Of these, 1715 (61.1%) had BPD, and 1090 (38.9%) did not have BPD. Among infants with BPD, 60.8% were born at ≤26 weeks GA; among infants without BPD, 29.7% were born at ≤26 weeks GA (). The rates of other complications (CRM, intraventricular hemorrhage, retinopathy of prematurity, patent ductus arteriosus, and necrotizing enterocolitis) were significantly higher in infants with BPD compared with those without.
The mean birth hospital length of stay was longer for infants with BPD versus those without; the difference was not statistically significant (95 ± 33 vs. 68 ± 23 days; p = 0.4308; ). In infants with BPD, hospital readmission rates after NICU discharge were significantly higher than those without BPD for all causes (26.1% [447/1715] vs. 12.0% [131/1090]; p < 0.0001) as well as for respiratory causes specifically (16.9% [290/1715] vs. 6.4% [70/1090]; p < 0.0001; ). All health care costs, including birth hospitalization, were significantly higher for infants with BPD versus those without (). After birth hospitalization (732,019 USD ± 767,371 vs. 429,436 USD ± 382,726; p < 0.0001), outpatient visits (21,498 USD ± 21,979 vs. 13,362 USD ± 12,059; p < 0.0001) and hospital readmissions (17,730 ± 89,631 USD vs. 5135 USD ± 31,938; p < 0.0001) accounted for the majority of all-cause costs for both groups. Respiratory causes resulted in increases in costs from hospital readmissions (11,064 ± 74,605 USD vs. 1643 USD ± 12,258; p < 0.0001), ER visits (750 ± 3296 USD vs. 326 USD ± 1099; p < 0.0001), and outpatient visits (2206 ± 9478 USD vs. 828 USD ± 2080; p < 0.0001).
Figure 4. Total health care costs by presence of bronchopulmonary dysplasia (BPD) through 1 year corrected age in extremely preterm infants
Table 3. Length of stay during birth hospitalization, and hospital readmissions after discharge through 1 year corrected age, among extremely preterm infants in the BPD/no BPD cohorts
4. Discussion
In this study, we evaluated the impact of prematurity on the burden of lung disease beyond the NICU stay using real-world data from a large claims database. Using CRM definition criteria similar but not identical to those in the PROP prospective cohort study [Citation2], we retrospectively identified patients and determined their health care resource utilization and costs.
In PROP (N = 724), ~69% of infants had respiratory morbidity (CRM); a lower incidence of 46% was calculated for a secondary outcome of respiratory morbidity that excluded clinical respiratory symptoms [Citation2]. In the CRM/no CRM analysis of the present study, 29.0% of infants had CRM and the incidence of CRM increased with decreasing GA. The difference in findings between PROP and the current study may be attributable to study design. Our study was retrospective and claims-based, while PROP was prospective and collected firsthand data. Our definition of CRM was somewhat more restrictive in that it included the use of respiratory medication for ≥30 days. In PROP, information about respiratory medications was collected in postdischarge questionnaires, and the definition of respiratory morbidity did not specify the length of use. A lower prevalence was, therefore, expected for our study.
The lower hospital readmission rate in the BPD cohort (26.1%) compared with the CRM cohort (40.4%) can likely be explained by the shorter duration of follow-up for the BPD cohort, or that the diagnosis of BPD is given at hospital discharge but then drops off as the patient ages during the first year. In addition, the incidence rate for hospitalizations in infants with CRM (0.44 per person-year) was higher than that in infants with BPD (0.22 per person-year), which could be because CRM is defined partly on the basis of hospitalizations. Based on the quarterly data for frequencies of each of the three CRM components, respiratory medication use remained relatively constant through 1 year CA, which highlights its importance as a measure of long-term respiratory morbidity over time. In contrast, rehospitalizations or ER visits and home respiratory/oxygen support contributed most in earlier months, thus reinforcing the importance of including multiple outcomes when evaluating interventions and treatment benefits related to pulmonary morbidity.
Medical advances in neonatal care have led to increased survival rates of preterm infants born at earlier GAs, which could be associated with increased burden on both short- and long-term health care resource utilization. Results from this study showed that the birth hospital length of stay was longer, and hospital readmission rates after NICU discharge were significantly higher, for infants with CRM or BPD versus those without. In addition, total health care costs (including birth hospitalization) were significantly higher for infants with CRM or BPD versus those without. After birth hospitalization, the highest costs were due to outpatient visits and hospital readmissions (for both all-cause and respiratory-cause analyses), which suggests that these events could be relevant endpoints in clinical trials of CRM. Results from our study are in agreement with those from other studies that show that respiratory complications in preterm infants result in increased burden on health care resources [Citation10–15].
In the United States, direct medical costs for extremely preterm infants have been reported to range from 70,451 to 100,389 USD per patient (2005 US dollars) [Citation18]. Management of BPD in particular has been a difficult and persistent challenge over time, with few therapeutic advances or changes in prevalence over the past 20 years [Citation19,Citation20]. One study calculated that, after adjusting for potential confounders, BPD in infants with a body weight <1500 g was associated with a 31,565 USD increase in direct NICU costs (adjusted to 2009 US dollars), which was higher than costs associated with brain injury, necrotizing enterocolitis, and late-onset sepsis [Citation10]. Given that infants with CRM and BPD may experience long-term morbidities into adulthood, further direct and indirect costs to society could extend beyond costs incurred in the first year.
Among infants with CRM using respiratory medications in this study, the most frequently used class of medication was inhaled steroid (53.2%), followed by inhaled bronchodilator (33.5%), diuretic (30%), and methylxanthine (6.0%). In comparison, a recent analysis of PROP showed that for infants <29 weeks GA, inhaled bronchodilators were the most frequently used respiratory medication (30%) in the first year postdischarge, followed by inhaled steroid (22%), systemic steroid (15%), and diuretic (12%) [Citation21]. The use of pulmonary vasodilators was low (<2%) in both our study and PROP.
We acknowledge a number of limitations in our study. The MarketScan database is comprised of data from multiple sources from the US, the majority being large employers, followed by health plans including Medicaid and Medicare data. As a result, findings may not be generalizable to other populations. As with all claims database analyses, an ICD-9-CM or ICD-10-CM diagnosis code on a medical claim is not an attestation that the patient has the diagnosis, because the code may represent a rule-out diagnosis or may be recorded incorrectly. This could explain the higher than expected rates of BPD diagnosis (>60%) in this study compared with the rate (~40%) reported in other recent studies [Citation22–24]. Another limitation is that, as the data are de-identified, the exact date of birth is not available and needs to be approximated using the admission date of the earliest inpatient stay. In addition, for preterm infants identified by ICD-9-CM diagnosis codes, GA was approximated using the upper bound of the GA range of each diagnosis code. These approximations may contribute to small errors in calculating CA. An additional consideration is the variability of clinical practices for treating BPD across centers, as well as changes in clinical practice over the period of observation (2009 to 2016), which may have affected the findings of the current analysis. However, the current findings are considered to be reflective of real-world practice, owing to the large sample size of the study. Finally, due to varied length of follow-up among selected infants in the BPD/no BPD analysis, the health care costs need to be annualized. This approach may result in some bias if infants with follow-up periods <1 year had much higher or lower costs compared with those with longer follow-up periods.
5. Conclusion
In conclusion, CRM was found to be prevalent among extremely preterm infants, with events through 1 year CA indicating long-term respiratory morbidity. The results also indicate that CRM and BPD place a significant burden on health care systems that extends beyond the initial birth hospitalization.
Declaration of interest
ME Mowitz was a paid consultant to Shire, a Takeda company, in relation to this study. L Han is an employee of and holds stock/stock options in Shire, a Takeda company. R Ayyagari, W Gao, J Wang, and J Zhao are employees of Analysis Group, Inc., who were paid consultants to Shire in relation to this study. A Mangili and SP Sarda were employees of Shire, a Takeda company, at the time of the study. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Reviewers disclosure
Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.
Author contributions
All authors made substantial contributions to study conception and design, acquisition of data, or analysis and interpretation of data; participated in revising this manuscript critically for important intellectual content; and gave final approval of the version to be published.
Acknowledgments
Under direction of the authors, Nasser Malik, an employee of Excel Scientific Solutions, provided writing assistance for this manuscript. Editorial assistance in formatting, proofreading, and copy editing also was provided by Excel Scientific Solutions. Writing and editorial support were funded by Takeda.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Additional information
Funding
References
- Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet. 2012;379(9832):2162–2172.
- Keller RL, Feng R, DeMauro SB, et al. Bronchopulmonary dysplasia and perinatal characteristics predict 1-year respiratory outcomes in newborns born at extremely low gestational age: a prospective cohort study. J Pediatr. 2017;187:89–97.
- Islam JY, Keller RL, Aschner JL, et al. Understanding the short- and long-term respiratory outcomes of prematurity and bronchopulmonary dysplasia. Am J Respir Crit Care Med. 2015;192(2):134–156.
- Jobe AH. Mechanisms of lung injury and bronchopulmonary dysplasia. Am J Perinatol. 2016;33(11):1076–1078.
- Berkelhamer SK, Mestan KK, Steinhorn RH. Pulmonary hypertension in bronchopulmonary dysplasia. Semin Perinatol. 2013;37(2):124–131.
- Jensen EA, Schmidt B. Epidemiology of bronchopulmonary dysplasia. Birth Defects Res A Clin Mol Teratol. 2014;100(3):145–157.
- Ali K, Greenough A. Long-term respiratory outcome of babies born prematurely. Ther Adv Respir Dis. 2012;6(2):115–120.
- Bhandari A, McGrath-Morrow S. Long-term pulmonary outcomes of patients with bronchopulmonary dysplasia. Semin Perinatol. 2013;37(2):132–137.
- Davidson LM, Berkelhamer SK. Bronchopulmonary dysplasia: chronic lung disease of infancy and long-term pulmonary outcomes. J Clin Med. 2017;6(1):4.
- Johnson TJ, Patel AL, Jegier BJ, et al. Cost of morbidities in very low birth weight infants. J Pediatr. 2013;162(2):243–249.
- Landry JS, Croitoru D, Jin Y, et al. Health care utilization by preterm infants with respiratory complications in Quebec. Can Respir J. 2012;19(4):255–260.
- Morrow CB, McGrath-Morrow SA, Collaco JM. Predictors of length of stay for initial hospitalization in infants with bronchopulmonary dysplasia. J Perinatol. 2018;38(9):1258–1265.
- Patel AL, Johnson TJ, Robin B, et al. Influence of own mother’s milk on bronchopulmonary dysplasia and costs. Arch Dis Child Fetal Neonatal Ed. 2017;102(3):F256–F61.
- Klinger G, Sirota L, Lusky A, et al. Israel Neonatal Network. Bronchopulmonary dysplasia in very low birth weight infants is associated with prolonged hospital stay. J Perinatol. 2006;26(10):640–644.
- Smith VC, Zupancic JA, McCormick MC, et al. Rehospitalization in the first year of life among infants with bronchopulmonary dysplasia. J Pediatr. 2004;144(6):799–803.
- Mowitz ME, Ayyagari R, Gao W, et al. Health care burden of bronchopulmonary dysplasia among extremely preterm infants. Front Pediatr. 2019;7:510.
- Lapcharoensap W, Bennett MV, Xu X, et al. Hospitalization costs associated with bronchopulmonary dysplasia in the first year of life. J Perinatol. 2020;40(1):130–137.
- Zupancic JAF. Institute of Medicine (US) Committee on Understanding Premature Birth and Assuring Healthy Outcomes. D, A systematic review of costs associated with preterm birth. In: Behrman RE, Butler AS, editors. Preterm birth: causes, consequences, and prevention. Washington (DC): National Academies Press (US); 2007;688-724.
- Iyengar A, Davis JM. Drug therapy for the prevention and treatment of bronchopulmonary dysplasia. Front Pharmacol. 2015;6:12.
- Latini G, De Felice C, Giannuzzi R, et al. Survival rate and prevalence of bronchopulmonary dysplasia in extremely low birth weight infants. Early Hum Dev. 2013;89(Suppl1):S69–S73.
- Ryan RM, Keller RL, Poindexter BB, et al. PROP Investigators. Respiratory medications in infants <29 weeks during the first year postdischarge: the prematurity and respiratory outcomes program (PROP) consortium. J Pediatr. 2019;208:148–155.
- Poindexter BB, Feng R, Schmidt B, et al. Prematurity and Respiratory Outcomes Program. Comparisons and limitations of current definitions of bronchopulmonary dysplasia for the prematurity and respiratory outcomes program. Ann Am Thorac Soc. 2015;12(12):1822–1830.
- Stoll BJ, Hansen NI, Bell EF, et al. Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network. Neonatal outcomes of extremely preterm infants from the NICHD Neonatal Research Network. Pediatrics. 2010;126(3):443–456.
- Lapcharoensap W, Gage SC, Kan P, et al. Hospital variation and risk factors for bronchopulmonary dysplasia in a population-based cohort. JAMA Pediatr. 2015;169(2):e143676.