BACKGROUND: Hospitalization for lower respiratory tract infections (LRTIs) among children have been well characterized. We characterized hospitalizations for severe LRTI among children.
METHODS: We analyzed claims data from commercial and Medicaid insurance enrollees (MarketScan) ages 0 to 18 years from 2007 to 2011. LRTI hospitalizations were identified by the first 2 listed International Classification of Diseases, Ninth Revision discharge codes; those with ICU admission and/or receiving mechanical ventilation were defined as severe LRTI. Underlying conditions were determined from out- and inpatient discharge codes in the preceding year. We report insurance specific and combined rates that used both commercial and Medicaid rates and adjusted for age and insurance status.
RESULTS: During 2007–2011, we identified 16 797 and 12 053 severe LRTI hospitalizations among commercial and Medicaid enrollees, respectively. The rates of severe LRTI hospitalizations per 100 000 person-years were highest in children aged <1 year (commercial: 244; Medicaid: 372, respectively), and decreased with age. Among commercial enrollees, ≥1 condition increased the risk for severe LRTI (1 condition: adjusted relative risk, 2.68; 95% confidence interval, 2.58–2.78; 3 conditions: adjusted relative risk, 4.85; 95% confidence interval, 4.65–5.07) compared with children with no medical conditions. Using commercial/Medicaid combined rates, an estimated 31 289 hospitalizations for severe LRTI occurred each year in children in the United States.
CONCLUSIONS: Among children, the burden of hospitalization for severe LRTI is greatest among children aged <1 year. Children with underlying medical conditions are at greatest risk for severe LRTI hospitalization.
- CPT —
- Current Procedural Terminology
- ICD-9 —
- International Classification of Diseases, Ninth Revision
- LRTI —
- lower respiratory tract infection
What’s Known on This Subject:
Lower respiratory tract infections (LRTIs), including pneumonia, are in the top 10 causes of death among children in the United States. In high-income countries, 3% to 14% of LRTI hospitalizations have been reported to require admission to an ICU.
What This Study Adds:
During 2007–2011, approximately 31 289 hospitalizations for severe LRTI occurred in children each year in the United States. Children <1 year of age had the highest rates of severe LRTI and accounted for 30% of severe LRTI hospitalizations.
Globally, respiratory infections are a common cause of morbidity and mortality among children. In the United States, lower respiratory tract infections (LRTIs), including pneumonia, were the fifth leading cause of death among 5- to 9-year-old children and the sixth leading cause of death among children aged <5 and 10 to 14 years during 2009.1 Similarly, during 2010, LRTIs were estimated to be the 11th highest cause of life years lost among children in North America.2 In high-income countries, 3% to 18% of pediatric admissions to the hospital are due to LRTI,3,4 and among pediatric LRTI hospitalizations, 3% to 14% have been reported to require admission to an ICU.5–9 Data on rates of pediatric hospitalizations for severe LRTI, including those requiring specialized care, such as ICU or mechanical ventilation, are limited. Such information might be useful for public health emergency planning for pandemics and other emerging respiratory infections, such as Middle East respiratory syndrome coronavirus.10 We describe the rate of LRTI hospitalizations requiring ICU or mechanical ventilation (severe LRTI) among children aged ≤18 years in the United States from 2007 to 2011. In addition, we describe risk factors for severe LRTI among children.
We analyzed data from MarketScan Commercial Claims and Medicaid Databases from 2007 to 2011 (Truven Health Analytics, Ann Arbor, MI). MarketScan collects deidentified individual level data from commercial (∼40 million people per year) and Medicaid (∼6 million people per year) health insurance claims with wide geographical representation of the United States.11,12 We included all children 0 to 18 years old with at least 1 LRTI hospitalization from 2007 to 2011.
LRTI was defined by International Classification of Diseases, Ninth Revision (ICD-9) discharge codes for acute bronchitis (466.0), acute bronchiolitis (466.1), pneumonia (480–486), as well as others (Supplemental Table 7).13,14 LRTI hospitalizations were defined as any hospitalization with an LRTI ICD-9 code as a primary or secondary discharge diagnosis during the study period. Both primary and secondary codes were included to capture young children admitted with dehydration or other syndromes; we assumed LRTI was the underlying cause. Severe LRTI hospitalization was defined as an LRTI hospitalization requiring admission to the ICU or mechanical ventilation at any time during the LRTI hospitalization. Nonsevere LRTI was defined as hospitalizations without ICU admission or requiring mechanical ventilation. Use of mechanical ventilation was determined by either Current Procedural Terminology (CPT) or ICD-9 procedure codes (Supplemental Table 8). Repeat hospitalizations within 14 days of the original LRTI hospitalization were excluded from the analysis. Enrollees could have multiple LRTI hospitalizations throughout the year.
Age is provided in the commercial database. However, the Medicaid database includes only year of birth. For the Medicaid database, age upon entering the calendar year was calculated for all Medicaid enrollees as year of hospital admission, minus year of birth, minus 1, except those who were born in the year of enrollment were zero years during the year of enrollment.
Underlying medical conditions were determined by using ICD-9 discharge diagnoses from hospitalizations and outpatient visits during the 1 year before the LRTI hospitalization. Underlying medical conditions were categorized as cardiac disease, pulmonary disease other than asthma, renal disease, hemoglobinopathies, HIV/AIDS, immunodeficiencies due to conditions other than HIV, malignancy, diabetes, other metabolic disorders, obesity, liver disease, neurologic and neuromuscular disorders, cerebrovascular disease, complications during the mother’s pregnancy, a history of premature birth, and other congenital conditions not listed elsewhere (Supplemental Table 9). Asthma was an underlying condition in children aged ≥2 years.
Person-years were calculated by using total months each child was enrolled in a health plan supplying data to MarketScan each year. We determined factors associated with severe LRTI compared with nonsevere LRTI hospitalization. Variables significant in bivariate analysis and gender were included in multivariable generalized linear models. We did not include individual medical conditions in multivariable models; the type and frequency of conditions varied by age group and age group specific models were beyond the scope of this article (Supplemental Table 10). Given differences in how data are collected for the commercial and Medicaid databases, results for the bivariate and multivariable analyses are stratified by insurance status.
We combined the rates from the commercial and Medicaid databases to estimate the total annual number of children in the United States with LRTI hospitalizations. Rates were weighted by the percent of children in the United States with private versus government-sponsored insurance.15 We assumed children without private or government sponsored insurance (<10%) were most similar to those with government sponsored insurance, and included them with the Medicaid population. Rates were weighted by age based on US census data.16 Using the combined weighted rates, we estimated the number of LRTI hospitalizations, severe LRTI hospitalizations, and severe LRTI hospitalizations requiring mechanical ventilation in the United States. Estimates were calculated for the entire year, and for January through March, the 3 months with the highest numbers of LRTI and severe LRTI hospitalizations, since resource needs for LRTI fluctuate throughout the year.
This analysis was determined to be research not involving human subjects.
A total of 55.9 million enrollment records of children 0 to 18 years old were included in the commercial database from 2007 to 2011 (7% <1 year old, 4% 1 to <2 years old, 14% 2–4 years old, 25% 5–9 years old, and 50% 10–18 years old), representing 29 million unique enrollees. Among a total of 2.2 million all-cause hospitalizations, 120 516 (5.5%) were LRTI hospitalizations, 16 797 (13.9%) were severe LRTI, and 1845 (1.5%) of children with LRTI hospitalization required mechanical ventilation (Table 1); 209 children hospitalized with LRTI with mechanical ventilation did not have ICU admission recorded. Among all LRTI hospitalizations, 64% had hospital ICD-9 discharge codes for pneumonia (480–486), 38% had bronchiolitis (466.1), and 7.7% influenza (487–488) codes; 60% of severe LRTIs were coded as pneumonia.
A total of 17.8 million enrollment records of children aged 0 to 18 years were included in the Medicaid database from 2007 to 2011 (7% <1 year old, 8% 1 to <2 years old, 19% 2–4 years old, 27% 5–9 years old, and 40% 10–18 years old), representing 6.4 million unique enrollees. Among a total of 1.6 million all-cause hospitalizations, 94 006 (5.7%) were LRTI hospitalizations, 12 053 (12.8%) were severe LRTI, and 1959 (2.1%) LRTI hospitalizations required mechanical ventilation (240 children with mechanical ventilation did not have ICU admission recorded). Among all LRTI hospitalizations, 60% had pneumonia ICD-9 discharge codes, 46% had bronchiolitis, and 7% had influenza codes; 54% of severe LRTIs were coded as pneumonia.
In both databases, 60% to 67% of all LRTI hospitalizations were among children <1 year old (Table 1). Additionally, among all cause hospitalizations, the proportion due to LRTI was higher among children aged 1 to <2 years (30% in commercial and 28% in Medicaid) and 2 to 4 years (21% in commercial and 21% in Medicaid).
The median age for children hospitalized for LRTI was 1 year for the commercial database and <1 for Medicaid (Table 2). More boys were hospitalized for LRTI than girls. Median length of stay was 2 days; length of ICU admission was not available. In both databases, <1% of LRTI hospitalizations had an in-hospital death recorded. Only 4.2% (commercial) and 7.4% (Medicaid) of children were hospitalized with LRTI more than once in the previous year. Approximately half (50% commercial; 58% Medicaid) of the children hospitalized with a LRTI had at least 1 underlying medical condition.
Among all children, rates of hospitalization for LRTI, severe LRTI, and in hospital deaths during LRTI hospitalization were highest in children aged <1 year and decreased with age (Table 3). Rates of hospitalization for LRTI and severe LRTI were higher in boys than girls. Total and age-specific rates of LRTI and severe LRTI hospitalizations were higher among children enrolled in Medicaid compared with those enrolled in commercial insurance. Among all children, rates of hospitalization for LRTI and severe LRTI per 100 000 person-years were 2.4- and 2.2-fold higher in the Medicaid children compared with children with commercial insurance, respectively. Among children enrolled in the Medicaid database during 2007 through 2011, 12.9% were eligible (eg, Basis of Eligibility code) due to a disability or blindness, 71% because they were a child, and 10% had unknown criteria.
During 2007–2011, rates of hospitalization for severe LRTI among children differed little year to year for all age groups (Fig 1). A similar finding was noted for rates of LRTI hospitalizations among children during the study period (Supplemental Fig 2).
We explored risk factors for severe LRTI compared with nonsevere LRTI hospitalizations (Table 4). Among children hospitalized with LRTI, children of younger age were relatively less likely to have severe LRTI compared with older children aged 10 to 18 years. An LRTI admission during the previous year increased the risk of severe versus nonsevere LRTI by 40% to 50% compared with children without a previous admission. Pulmonary diseases other than asthma and cardiac diseases were the most common medical conditions among children with severe and nonsevere LRTI hospitalizations; both were more common in children with severe LRTI. All the other underlying conditions examined were associated with varying degrees of increased risk for severe LRTI.
After adjusting for potential confounders with multivariable models, severe LRTI hospitalization, compared with nonsevere LRTI hospitalization, was associated with ≥1 underlying medical condition in children with both commercial or Medicaid insurance (Table 5). The presence of 1 underlying medical condition more than doubled the risk of severe LRTI compared with children with no conditions, and the risk of severe LRTI increased with increasing number of medical conditions. Children with ≥3 medical conditions had risk of severe LRTI approximately fivefold higher than those with no underlying conditions. Among the children with Medicaid, Hispanic race/ethnicity was associated with a 24% higher risk for severe LRTI compared with white race. We did not include specific underlying medical conditions in our multivariable models because the frequency and type of underlying medical conditions varied by age (Supplemental Table 10); age group specific models were beyond the scope of this article.
Combined age and insurance-adjusted rates of hospitalization for LRTI, severe LRTI, and LRTI hospitalizations requiring mechanical ventilation are shown in Table 6. From 2007 to 2011, the adjusted rate of hospitalization for severe LRTI was 53 per 100 000 person-years, and ranged from 16 to 243 per 100 000 person-years among children aged 10 to 18 and <1 year olds, respectively. The rates of hospitalization for LRTI (404 per 100 000 person-years) and LRTI hospitalizations requiring mechanical ventilation (8 per 100 000 person-years) also varied by age. During 2007–2011, we estimated 235 825 LRTI hospitalizations among children aged 0 to 18 years in the United States each year, 31 289 (13%) were for severe LRTI, and 4350 (1.8%) required mechanical ventilation.
Because the number of hospitalizations for severe LRTI varied by season (Supplemental Fig 2), we expected intensive care and mechanical ventilation needs to be higher during the winter seasons. During the peak months for severe LRTI hospitalization, January through March, 42% of severe LRTI hospitalizations and 45% of severe LRTI hospitalizations requiring mechanical ventilation occurred; 43% of children requiring mechanical ventilation during these peak months were <2 years old.
During 2007 through 2011, 13% to 14% of hospitalizations for LRTI among children in the United States were severe enough to require ICU or mechanical ventilation. We estimated that the overall adjusted rate of hospitalization for severe LRTI among children was 53 per 100 000 person-years; each year, 31 289 pediatric hospitalizations for severe LRTI occurred. Among all children, 30% of hospitalizations for severe LRTI were detected among children aged <1 year. For this age group, the adjusted rate of hospitalization for severe pneumonia was 243 per 100 000 person-years, or 1 in every 411 infants. The rate of hospitalization for severe LRTI decreased with increasing age. Children with underlying medical conditions were at an increased risk for severe LRTI. Having 1 medical condition doubled a child’s risk for severe LRTI compared with children with no medical conditions, and an increasing number of underlying medical conditions further increased this risk. Children hospitalized with LRTI with ≥3 underlying conditions had an approximately fivefold higher risk of severe LRTI, compared with children with no underlying medical conditions. Finally, we found that for children of all ages, those enrolled in Medicaid had higher rates of hospitalization for severe LRTI than children with commercial insurance; the proportion of children with underlying conditions was higher in the children with Medicaid.
Infants, children <1 year of age, had the highest rates of hospitalization for severe LRTI, LRTI hospitalizations requiring mechanical ventilation, and in-hospital deaths during a LRTI hospitalization. The rates of hospitalization for severe LRTI decreased with increasing age. Children aged 10 to 18 years had a rate of hospitalization for severe LRTI of 16 per 100 000 person-years, and children aged 1 to <2 years had a rate of 112 per 100 000 person-years. Other studies have revealed higher rates of LRTI hospitalization among children <1 year of age, compared with older children, in the United States.4,14,17–19 We are the first group to report rates of hospitalization for severe LRTI. Our results suggest that infants and young children will likely require substantial hospital resources during a pandemic or epidemic associated with severe LRTI.
Rates of hospitalization for LRTI and severe LRTI were higher among children enrolled in Medicaid than children with commercial insurance. Medicaid provides health insurance coverage for children in low-income families and for some children with certain disabilities or medical conditions. Low income, low education levels, younger maternal age, crowded living conditions, and race are well-described risk factors for low birth weight, prematurity, and for respiratory infections and could contribute to differences in outcome by insurance status.20 Also, a higher prevalence of pregnancy complications, prematurity, and other neonatal complications could contribute to the high rates of LRTI in children <1 year old in this population.21 We also identified a higher frequency of underlying medical conditions and multiple medical conditions among children hospitalized for LRTI with Medicaid versus commercial insurance, a finding supported by other reports and likely contributing to the higher rates of hospitalization for LRTI and severe LRTI.22 Approximately 13% of Medicaid children were eligible based on a disability. To more completely explore the differences in rates between Medicaid and commercially insured children, a better understanding of additional socioeconomic characteristics and other factors contributing to underlying medical conditions would be necessary.
The presence of underlying medical conditions significantly increased the risk of severe LRTI among children hospitalized with LRTI, a finding consistent with other reports.7,19 We found that 1 underlying medical condition more than doubled the risk for severe LRTI compared with no conditions, and each additional medical condition was associated with an increased risk of severe LRTI. Children with ≥3 medical conditions were at the highest risk for severe LRTI, compared with children with no medical conditions. In our study, children with cardiac disease and pulmonary conditions other than asthma had threefold and fivefold increased risk for severe LRTI, respectively. Several underlying medical conditions, such as neurologic and neuromuscular disorders, and a history of premature birth were common among children hospitalized for severe LRTI23,24 and were associated with an increased risk for severe versus nonsevere LRTI.
We used our adjusted rates to project estimated numbers of LRTI hospitalizations requiring ICU-level care and mechanical ventilation among children in the United States during 2007–2011. We estimated ∼236 000 LRTI hospitalizations and 31 000 severe LRTI hospitalizations among children in the United States each year. During 2007–2011, 42% of severe LRTI hospitalizations and 45% of severe LRTI hospitalizations requiring mechanical ventilation occurred during January through March. During these 3 months, ∼4300 hospitalizations for severe LRTI occurred each month. Randolph et al25 estimated an availability of 3899 PICU beds in the United States in 2001. Our national estimates are useful for planning for public health emergencies for infections or diseases that may result in respiratory distress or failure. Age-stratified projections are vital because mechanical ventilation equipment, personnel training, and drug dosing may differ for younger ages.26
There are several limitations that may affect our conclusions. First, all data are from administrative records, and coding practices could affect our results. Recently, changes in coding practices for pneumonia among adults were suggested to account for trends in pneumonia mortality.27 We did not find changes in trends over time during our study period and included both primary and secondary listed LRTI discharge codes; thus, we captured children with sepsis or respiratory failure as first listed codes, similar to Lindenauer et al.27 Second, age was not included in the Medicaid database. Thus, we likely overestimated the number of children in the <1-year age group and underestimated the number of children in the 1- to <2-year age group from the Medicaid data; this would be reflected in rates for these age groups. However, when we compared age calculated with the Medicaid method to the reported age in the commercial database, calculated and reported age was the same for ∼90% of patients. We are not able to differentiate hospital-acquired from community-acquired LRTI in our data sets. The children included in MarketScan databases may not be representative of the US population, although the number of enrollees is large and the rates for hospitalization are similar to other reports.7 MarketScan provides information on ICU use that is not available in other data sets. Finally, in our combined age and insurance adjusted rates, we assumed that children without insurance were more similar to those covered by Medicaid; this may not be true. However, during the study period, the proportion of uninsured children was <10% of all children.15
We report rates of hospitalization for severe LRTI among children in the United States, the first reported rates for severe LRTI in the United States that we are aware of. The large data sets used to estimate our rates, MarketScan commercial and Medicaid databases, cover a wide geographic area and include children with private and government-sponsored insurance. During 2007 through 2011, young children had the highest rate of hospitalization for severe LRTI. Among children hospitalized for LRTI, children with underlying medical conditions were at greatest risk for severe LRTI. Given these findings, prevention strategies to reduce LRTI among children should focus on young children (and people who may transmit infections to them), children with underlying medical conditions, and children in socioeconomic settings requiring public insurance, including expanded coverage for recommended vaccines to meet Healthy People 2020 goals. Preparedness planning for respiratory related emergencies should account for the large need for ICU and mechanical ventilation among infants and young children.
- Accepted June 16, 2014.
- Address correspondence to Alicia M. Fry, MD, MPH, Centers for Disease Control and Prevention, 1600 Clifton Rd, Mailstop A-32, Atlanta, GA 30333. E-mail:
Dr Greenbaum designed the study, carried out the analysis, and drafted the manuscript; Dr Chen assisted with the study design and data processing and analysis; Dr Reed designed the study, interpreted the data, and critically reviewed the manuscript; Drs Beavers and Callahan critically reviewed the manuscript; Dr Christensen helped to design the study and critically reviewed the manuscript; Dr Finelli interpreted the data and critically reviewed the manuscript; Dr Fry conceptualized and designed the study, interpreted the data, drafted the manuscript, and critically reviewed the manuscript; and all authors approved the final manuscript as submitted.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Dr Greenbaum’s current affiliation is Infectious Diseases Division, Johns Hopkins Hospital, Baltimore, MD.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Heron M
- Iwane MK,
- Chaves SS,
- Szilagyi PG,
- et al
- ↵Danielson E. White Paper: Health Research Data for the Real World: Truven Health Analytics, 2014 January 2014. Available at: http://truvenhealth.com/Portals/0/Users/031/31/31/PH_13434%200314_MarketScan_WP_web.pdf. Accessed July 1, 2014
- ↵Truven Health Analytics. MarketScan research. Available at: http://marketscan.truvenhealth.com/marketscanportal/. Accessed July 1, 2014
- Carmen DeNavas-Walkt BP,
- Smith J
- US Census Bureau
- Lee GE,
- Lorch SA,
- Sheffler-Collins S,
- Kronman MP,
- Shah SS
- Brandon GD,
- Adeniyi-Jones S,
- Kirkby S,
- Webb D,
- Culhane JF,
- Greenspan JS
- van Woensel JB,
- van Aalderen WM,
- Kimpen JL
- Copyright © 2014 by the American Academy of Pediatrics