Keywords gestational age - hospitalization - immunization, passive - infant - infant, premature
- observational study - practice guideline as topic - respiratory syncytial viruses
Respiratory syncytial virus (RSV) is a common cause of lower respiratory tract infection
and the leading cause of hospitalization among infants in the United States.[1 ]
[2 ] Birth at ≤ 35 weeks gestational age (wGA) is a strong risk factor for RSV hospitalization
(RSVH),[2 ] and young chronological age (CA) increases RSVH risk. From 1997 through 2006, an
estimated 132,000 to 172,000 RSV-associated hospitalizations occurred annually among
U.S. children aged < 5 years, with those aged < 1 year having the highest hospitalization
rates.[1 ]
[3 ] Other conditions associated with increased RSVH risk include pulmonary disease (especially
chronic lung disease of prematurity [CLDP]), hemodynamically significant congenital
heart disease (CHD), immunodeficiency, and neuromuscular diseases.[4 ] Preterm infants hospitalized for RSV disease have a high morbidity burden, including
high rates of intensive care unit (ICU) admission and use of mechanical ventilation,[5 ]
[6 ]
[7 ] particularly among younger infants.[5 ]
[8 ]
[9 ]
Although there are no recommended treatments or approved vaccines for RSV, passive
immunization with the monoclonal antibody palivizumab has been shown to reduce hospitalizations
due to RSV[10 ] and is currently indicated for preventing serious lower respiratory tract disease
caused by RSV in high-risk infants aged < 2 years.[11 ] Since the initial approval of palivizumab by the U.S. Food and Drug Administration
in 1998, the American Academy of Pediatrics (AAP) Committee on Infectious Diseases
(COID) has recommended immunoprophylaxis (IP) with palivizumab for infants at increased
risk of RSV disease (summarized in [Table 1 ]).[4 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[17 ] In 2014, the AAP stated that palivizumab IP was no longer recommended for preterm
infants > 29 wGA without CLDP or CHD,[4 ]
[18 ] asserting that the rates of RSVH among these infants were similar to those of term
infants.[19 ]
Table 1
American Academy of Pediatrics recommendations[a ] for respiratory syncytial virus immunoprophylaxis
Risk factor
1998[b ]
2003[c ]
2009[d ]
2014[e ]
< 29 wGA
< 12 months of age at season start
< 12 months of age at season start
< 12 months of age at season start
< 12 months of age at season start; no second season
29–31 wGA
< 6 months of age at season start
< 6 months of age at season start
< 6 months of age at season start
Not recommended
32–35 wGA
< 6 months of age at season start with additional risk factors[f ]
< 6 months of age at season start with two of five risk factors[g ]
32–34 wGA and < 3 months of age at season start with one of two risk factors[h ]; dose until 3 months CA
Not recommended
Abbreviations: CA, chronological age; wGA, weeks gestational age.
a Recommendations for infants with no other underlying reason for receiving RSV IP
(i.e., chronic lung disease of prematurity or congenital heart disease).
b American Academy of Pediatrics Committee on Infectious Diseases and Committee on
Fetus and Newborn.[12 ]
c
Red Book: 2003 Report of the Committee on Infectious Diseases. 26th ed.[14 ]
d
Red Book: 2009 Report of the Committee on Infectious Diseases. 28th ed.[16 ]
e American Academy of Pediatrics Committee on Infectious Diseases, American Academy
of Pediatrics Bronchiolitis Guidelines Committee.[4 ]
f Risk factors include neurologic disease in very-low-birth-weight infants, number
of young siblings, child care center attendance, exposure to tobacco smoke in the
home, anticipated cardiac surgery, and distance to and availability of hospital care
for severe respiratory illness.
g Risk factors include child care attendance, school-aged siblings, exposure to environmental
air pollutants, congenital abnormalities of the airways, and severe neuromuscular
disease.
h Risk factors include child care attendance and siblings < 5 years of age.
The impact of the 2014 policy change on RSV IP and outcomes is not fully understood.
The 2014 to 2017 RSV seasons were the first since 1998 during which otherwise healthy
29 to 34 wGA infants were ineligible for RSV IP according to the AAP policy. Several
U.S. studies have demonstrated an increase in RSVH risk among infants 29 to 34 wGA
and < 6 months CA under the more restrictive recommendations.[9 ]
[20 ]
[21 ]
[22 ] These findings would benefit from validation in other large datasets. In addition,
the effect of the policy change on RSVH severity and costs has not been evaluated.
The present study addresses these gaps by using a large U.S. administrative claims
database to examine RSV IP use and the rates, severity, and costs associated with
hospitalizations for RSV before and after the 2014 AAP policy change for 29 to 34
wGA preterm infants versus term infants.
Materials and Methods
Study Design and Data Source
This was an observational retrospective cohort study conducted using administrative
claims data from the Optum Research Database (ORD). The ORD contains medical and pharmacy
claims with linked enrollment information covering 59.5 million lives from 1993 to
the present and is geographically diverse across the United States. Medical claims
in the ORD include diagnosis codes and procedure codes from the International Classification
of Diseases, 9th/10th Revisions, Clinical Modification (ICD-9/10-CM); Current Procedural
Terminology codes; Diagnosis Related Group codes; Healthcare Common Procedure Coding
System codes; site of service codes; provider specialty codes; and paid amounts. Pharmacy
claims in the ORD are for outpatient pharmacy services and include drug name, dosage
form, drug strength, and fill date. Inpatient pharmacy claims are not captured in
the ORD. Institutional Review Board approval or waiver of authorization was not required,
as no identifiable protected health information was accessed during this study.
Study Population
The study included otherwise healthy 29 to 34 wGA preterm infants (the group affected
by the 2014 AAP policy change) and ≥ 37 wGA term infants (a group unaffected by the
policy change). Included infants were required to have commercial health insurance;
to have at least one medical claim with a diagnosis code in any position for birth
status and date of birth from July 1, 2011, through March 31, 2017 (study period);
to have been discharged alive from their birth hospitalization during the study period;
to have at least 1 day of continuous health plan enrollment with medical and pharmacy
benefits after their index date (birth hospitalization discharge date plus 2 days);
to have age, sex, geographic region, and insurance plan type recorded in the claims
database; to be aged < 6 months during the study period; and to have no evidence of
other major health problems (CLDP, CHD, cystic fibrosis, immunodeficiency, congenital
respiratory anomalies, neuromuscular disease, organ transplants, and other neuromuscular,
immunological, or genetic conditions).
Outcomes
Study outcomes were assessed during the follow-up period, which began on the index
date and continued until the earliest of the end of the study period, the end of the
first 6 months of life, health plan disenrollment, or death. Outcomes were captured
by RSV season (November 1 through March 31) during the follow-up period and stratified
by wGA. As a supplemental analysis, outcomes were also stratified by CA at the time
the outcome occurred (< 3 or 3–< 6 months). RSV IP use was calculated as the proportion
of infants with at least one medical or outpatient pharmacy claim for palivizumab
at the specified CA. RSVH included hospitalizations coded as RSV infection, RSV pneumonia,
and RSV bronchiolitis (ICD-9-CM codes 079.6, 466.11, and 480.1 or ICD-10-CM codes
B974, J121, J205, and J210). Bronchiolitis hospitalizations (BHs) included those coded
as RSV bronchiolitis and those coded as unspecified bronchiolitis (ICD-9-CM code 466.19
or ICD-10-CM codes J218 and J219, if there was no evidence of RSV, influenza, bacterial
pneumonia, or other viral pathogen within 3 days of the claim). RSVH and BH were not
mutually exclusive; the percentage of hospitalized infants whose hospitalizations
qualified as both RSVH and BH ranged from 95 to 98% in the RSVH group and from 70
to 82% in the BH group, depending on gestational age and RSV season. The rates of
RSVH and BH were measured per 100 infants per RSV season (infant-seasons, analogous
to person-years). Indicators of RSVH and BH severity included length of stay, ICU
admission, and use of mechanical ventilation (including noninvasive approaches) during
RSVH or BH. The costs of RSVH and BH (combined health plan– and patient-paid amounts
from the claims database) were calculated among infants with at least one applicable
hospital stay and adjusted to 2015 US$ using the medical care component of the Consumer
Price Index.[23 ]
Statistical Analyses
Seasonal variations in RSV circulation and severity were accounted for using preterm/term
rate ratios to compare hospitalization rates in the 2011 to 2014 (pre–policy change)
and 2014 to 2017 (post–policy change) RSV seasons.
Multivariable difference-in-difference (DID) log binomial modeling was performed to
investigate the differential effect of the AAP policy change on hospitalization risk
for preterm infants versus term infants. DID analysis compares differences in outcomes
before and after an intervention (in this case, the AAP policy change) between a group
that is affected by the intervention (preterm infants) and a group that is unaffected
by the intervention but exposed to the same background factors[24 ] (term infants). The DID expression for the post-2014 versus pre-2014 risk estimate
was calculated as:
Hospitalization rate for preterm infants in 2014 to 2017/hospitalization rate for
preterm infants in 2011 to 2014
Hospitalization rate for term infants in 2014 to 2017/hospitalization rate for term
infants in 2011 to 2014.
The ratios in the DID expression remove the influence of variables that are unrelated
to the policy change but may vary over time.[24 ] Examples of such variables are RSV season severity, trends in bronchiolitis admission
practices, and the 2015 transition from ICD-9-CM to ICD-10-CM coding. Adjusted hospitalization
rates derived from regression models that include wGA, time (pre– or post–policy change),
CA, sex, and interaction of wGA and time as explanatory variables were used in the
DID expression to control for possible differences between the groups being compared.[24 ]
Other outcomes were compared among preterm and term infants for the 2011 to 2014 versus
2014 to 2017 RSV seasons using chi-square tests for categorical variables (RSVP IP
use, ICU admission, and mechanical ventilation use) and t -tests for continuous variables (length of stay and costs). Statistical significance
was defined as p < 0.05. Statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC).
Results
Study Population
Of 31,614,116 patients with commercial insurance during the study period, 517,223
met the birth date, health plan enrollment, and age criteria; were discharged alive
from the birth hospitalization; and had no missing demographic data. The final sample
included 12,558 preterm infants and 323,216 term infants ([Fig. 1 ]).
Fig. 1 Patient selection. wGA, weeks gestational age. *A large number of infants (113,671)
had an unknown gestational age and therefore were not eligible to be included in the
final sample. † Major health problems (identified using International Classification of Diseases,
9th/10th Edition, Clinical Modification codes, Diagnosis Related Group codes, and
Current Procedural Terminology codes) include chronic lung disease, congenital heart
disease, cystic fibrosis, immunodeficiency, congenital respiratory anomalies, neuromuscular
disease, organ transplants, and other neuromuscular, immunological, or genetic conditions.
RSV Immunoprophylaxis Use
Outpatient RSV IP use among preterm infants ([Fig. 2 ]) decreased significantly after the AAP policy change (p < 0.001 within each wGA cohort). The percentage of infants receiving at least one
RSV IP dose while aged < 3 months in 2011 to 2014 ranged from 23.6% (for 33–34 wGA
infants) to 48.3% (for 29–30 wGA infants) ([Supplementary Table S1 ], available in the online version). In 2014 to 2017, these percentages fell to 0.2
and 6.3%, respectively. Similarly, the percentage of preterm infants who received
at least one RSV IP dose while aged 3 to < 6 months ([Supplementary Table S1 ], available in the online version) ranged from 10.0% (for 33–34 wGA infants) to 69.8%
(for 29–30 wGA infants) in 2011 to 2014, but fell to 0.2 and 19.0%, respectively,
in 2014 to 2017. The proportion of term infants with RSV IP use was < 1% for all RSV
seasons evaluated (data not shown).
Fig. 2 Preterm infants receiving at least one outpatient dose of palivizumab by RSV season.
Any claim for palivizumab was counted; doses were not required to be administered
according to recommendations. RSV IP use among preterm infants within each gestational
age group decreased in 2014 to 2017 compared with 2011 to 2014 (p < 0.001 for all). Error bars represent ± 1 standard error. IP, immunoprophylaxis;
RSV, respiratory syncytial virus; wGA, weeks gestational age.
Rates of RSV Hospitalization and Bronchiolitis Hospitalization
Crude RSVH and BH rates (without adjustment for potential confounders) increased among
preterm infants after the policy change. Unadjusted RSVH rates for preterm infants
in the 2014 to 2017 versus 2011 to 2014 RSV seasons were 3.1 versus 2.1 per 100 infant-seasons
(p = 0.022) ([Table 2 ]). Unadjusted BH rates for preterm infants in the 2014 to 2017 versus 2011 to 2014
RSV seasons were 4.0 versus 2.5 per 100 infant-seasons (p = 0.001) ([Table 2 ]). Infants aged < 3 months had greater rates of both RSVH and BH than those aged
3 to < 6 months ([Supplementary Table S2 ], available in the online version). Among term infants, the rates of RSVH and BH
were unchanged in 2014 to 2017 compared with 2011 to 2014 (p > 0.05 for all CA groups) ([Table 2 ] and [Supplementary Table S2 ], available in the online version).
Table 2
Unadjusted RSV hospitalization and bronchiolitis hospitalization rates for preterm
and term infants
Hospitalizations per 100 infant-seasons (95% CI)
p -Value
2011–2014
2014–2017
RSV hospitalization
Preterm
2.1 (1.7–2.7)
3.1 (2.5–3.8)
0.022
Term
1.1 (1.1–1.2)
1.1 (1.0–1.1)
0.171
Bronchiolitis hospitalization
Preterm
2.5 (2.0–3.1)
4.0 (3.3–4.8)
0.001
Term
1.3 (1.2–1.3)
1.2 (1.2–1.3)
0.618
Abbreviations: CI, confidence interval; RSV, respiratory syncytial virus.
The rate ratios comparing unadjusted preterm infant hospitalization rates with unadjusted
term infant hospitalization rates increased for RSVH and BH after the AAP policy change
([Table 3 ]). From before to after the policy change, the rate ratios for preterm versus term
infants increased from 1.9 to 2.9 for RSVH (p = 0.011) and from 2.0 to 3.3 for BH (p = 0.001). The rate ratios before and after the policy change for infants aged < 3
months were similar to those for infants aged 3 to < 6 months ([Supplementary Table S3 ], available in the online version). When examined by RSV season, hospitalization
rate ratios before the policy change ranged from 1.2 to 2.8 for RSVH and from 1.3
to 2.6 for BH ([Fig. 3 ]), compared with 2.1 to 3.7 for RSVH and 2.6 to 3.9 for BH after the policy change
([Fig. 3 ]). The percentage of preterm infants in each RSV season receiving RSV IP was less
following the AAP policy change ([Fig. 3 ]).
Fig. 3 Unadjusted RSV hospitalization and BH rate ratios by individual RSV season. The rate
ratios indicate preterm infant risk relative to term infant risk. BH, bronchiolitis
hospitalization; IP, immunoprophylaxis; RSV, respiratory syncytial virus; RSVH, RSV
hospitalization; wGA, weeks gestational age.
Table 3
Unadjusted RSV hospitalization and bronchiolitis hospitalization rate ratios for preterm
infants versus term infants
Rate ratio, preterm infants vs. term infants (95% CI)
p -Value[a ] for pre– vs. post–policy change rate ratios
RSV hospitalization
2011–2014
1.9 (1.4–2.4)
0.011
2014–2017
2.9 (2.3–3.7)
Bronchiolitis hospitalization
2011–2014
2.0 (1.5–2.5)
0.001
2014–2017
3.3 (2.6–4.0)
Abbreviations: CI, confidence interval; RSV, respiratory syncytial virus.
a
p -Values are for interaction of gestational age and time period (2011–2014 or 2014–2017),
calculated using a difference-in-difference negative binomial model.
Multivariable DID modeling of rate ratios with adjustment for patient characteristics
(wGA, sex) revealed that the preterm RSVH risk relative to term infants increased
significantly after the AAP policy change. The hospitalization risk associated with
prematurity was 55% greater for RSVH (relative risk = 1.55, 95% confidence interval
[CI]: 1.11–2.17, p = 0.011) and 63% greater for BH (relative risk = 1.63, 95% CI: 1.20–2.22, p = 0.002) in the 2014 to 2017 seasons compared with the 2011 to 2014 seasons.
Severity of RSV Hospitalization and Bronchiolitis Hospitalization
Mean length of stay for RSVH and BH ([Fig. 4A ]) among preterm infants was significantly greater after the AAP policy change. Mean
(median) RSVH length of stay increased from 4.7 (4.0) days in 2011 to 2014 to 7.8
(5.5) days in 2014 to 2017 (p = 0.028). BH length of stay increased from 4.7 (4.0) to 7.5 (5.0) days (p = 0.015). Among term infants, mean length of stay for both RSVH and BH was unchanged
in 2014 to 2017 compared with 2011 to 2014 (range: 4.4–4.5 days; p > 0.05 for all CA groups). Preterm infants aged < 3 months had the highest mean (median)
length of stay overall: 5.4 (5.0) days pre–policy change versus 9.5 (6.0) days post–policy
change (p = 0.062) for RSVH, and 5.2 (4.0) days pre–policy change versus 9.6 (6.0) days post–policy
change (p = 0.022) for BH ([Supplementary Table S4 ], available in the online version).
Fig. 4 Severity indicators for RSV hospitalization and bronchiolitis hospitalization. (A)
Length of hospital stay. (B) ICU admission during hospitalization. (C) MV use during
hospitalization. Error bars represent ± 1 standard error. ICU, intensive care unit;
MV, mechanical ventilation. *p < 0.05 for differences between time periods.
Admission to the ICU during RSVH and BH ([Fig. 4B ]) was significantly more prevalent among preterm infants after the AAP policy change.
In the 2014 to 2017 versus 2011 to 2014 RSV seasons, the proportions of ICU admissions
among hospitalized preterm infants were 48.8 versus 26.6% for RSVH (p = 0.006) and 45.3 versus 25.3% for BH (p = 0.006). The proportions of term infants who were admitted to the ICU during RSVH
and BH in 2014 to 2017 versus 2011 to 2014 were not significantly different: for RSVH,
26.8 versus 23.1% (p = 0.082); for BH, 24.9 versus 21.5% (p = 0.080). Overall, ICU admission was most prevalent among preterm infants aged < 3
months: 38.5% pre–policy change versus 62.3% post–policy change (p = 0.024) for RSVH, and 36.4% pre–policy change versus 63.9% post–policy change (p = 0.005) for BH ([Supplementary Table S4 ], available in the online version).
The use of mechanical ventilation during RSVH and BH ([Fig. 4C ]) was numerically greater for preterm and term infants after the AAP policy change,
but the differences between periods did not always reach statistical significance.
In 2014 to 2017 versus 2011 to 2014, the proportions of infants who required mechanical
ventilation during RSVH were 15.9 versus 6.3% (p = 0.073) for preterm infants and 8.1 versus 4.3% (p = 0.001) for term infants. The proportions requiring mechanical ventilation during
BH were 15.1 versus 5.3% (p = 0.039) for preterm infants and 7.6 versus 4.0% (p < 0.001) for term infants. Preterm infants aged < 3 months had the highest prevalence
of mechanical ventilation use overall: for RSVH, 10.3% pre–policy change versus 20.8%
post–policy change (p = 0.178); for BH, 9.1% pre–policy change versus 23.0% post–policy change (p = 0.063) ([Supplementary Table S4 ], available in the online version).
Costs of RSV Hospitalization and Bronchiolitis Hospitalization
Mean hospitalization costs among preterm infants aged < 6 months were numerically
higher in the 2014 to 2017 RSV seasons compared with the 2011 to 2014 seasons, but
these differences were not statistically significant (p = 0.089 for RSVH and p = 0.070 for BH; [Fig. 5 ]). Mean hospitalization costs among term infants showed negligible changes in 2014
to 2017 versus 2011 to 2014 (p = 0.148 for RSVH and p = 0.126 for BH). Overall, the highest mean (standard deviation) hospitalization costs
were found among preterm infants aged < 3 months: $21,761 ($38,212) during the pre–policy
change period versus $55,731 ($148,752) in the post–policy change period (p = 0.116) for RSVH, and $18,836 ($33,804) pre–policy change versus $52,992 ($139,306)
post–policy change (p = 0.070) for BH ([Supplementary Table S5 ], available in the online version). Median (interquartile range) hospitalization
costs, presented in the footnote of [Fig. 5 ] and in [Supplementary Table S5 ] (available in the online version), showed similar patterns.
Fig. 5 Mean cost of RSV hospitalization and bronchiolitis hospitalization among infants
with at least one hospitalization. Error bars represent ± 1 standard error; n values represent the number of hospitalizations. RSV, respiratory syncytial virus;
SD, standard deviation. *Median (interquartile range) RSV hospitalization costs in
2011 to 2014 versus 2014 to 2017 were $10,742 ($18,682) versus $14,179 ($18,682) for
preterm infants and $9,130 ($9,713) versus $10,164 ($11,171) for term infants. † Median (interquartile range) bronchiolitis hospitalization costs in 2011 to 2014 versus
2014 to 2017 were $9,151 ($10,906) and $12,823 ($18,841) for preterm infants and $9,044
($9,165) versus $10,272 ($11,008) for term infants.
Discussion
This retrospective claims analysis of infants aged < 6 months showed a significant
decrease in outpatient use of RSV IP among 29 to 34 wGA preterm infants after the
2014 AAP policy change. This decreased use of RSV IP among preterm infants was concurrent
with increased RSVH and BH risks compared with term infants; preterm/term rate ratios
for RSVH and BH were greater in each post-2014 season relative to the 2011 to 2012
and 2013 to 2014 seasons. In addition, most RSVH and BH severity metrics increased
in 2014 to 2017 compared with 2011 to 2014 among preterm infants—and were particularly
high among preterm infants aged < 3 months—while remaining similar among term infants.
These findings suggest that the AAP policy change may have resulted in greater rates
of RSVH among preterm infants and that these hospitalizations became more burdensome
in terms of morbidity.
Our results are in line with those of previous analyses that found substantially decreased
outpatient RSV IP use and increased RSVH risk among preterm infants after the AAP
policy change.[9 ]
[20 ]
[21 ]
[22 ] In single-center studies by Rajah et al and Blake et al, RSVH rates for preterm
infants were significantly greater in post– versus pre–policy change RSV seasons.[9 ]
[21 ] Similarly, administrative claims analyses utilizing a patient population comparable
to ours from the Truven Health MarketScan database showed that RSVH risk among preterm
infants increased significantly in the 2014 to 2015 season compared with the 2013
to 2014 season,[20 ] and in the 2014 to 2016 seasons compared with the 2012 to 2014 seasons.[22 ]
A few studies have not found increased RSVH rates after the AAP policy change; however,
each of these has substantial methodological limitations, including failure to consider
CA[25 ] or specifically examine preterm infants,[26 ] insufficient statistical power to detect clinically plausible differences in RSVH
rates,[26 ]
[27 ] and inclusion of premature infants who remained eligible for RSV IP in the 2014
to 2015 season per Texas Medicaid policies.[25 ]
[28 ]
[29 ]
Importantly, our DID analysis—which isolated the impact of the AAP policy change by
adjusting for other temporal factors—showed that the increased RSVH risk associated
with preterm status increased further under the more restrictive RSV IP recommendations,
corroborating the results of a similar analysis by Goldstein et al.[22 ] Thus, contrary to the assertion that RSVH rates are similar among preterm infants
and term infants,[19 ] the predominant evidence confirms that preterm infants are at greater risk of RSVH
compared with term infants[2 ]
[3 ]
[5 ]
[8 ]
[30 ]
[31 ] and indicates that this risk disparity has worsened since the AAP policy change.
Previous studies have shown that RSVH is not only more common among younger infants[1 ]
[2 ]
[3 ]
[5 ]
[8 ]
[30 ] but also more severe, with young preterm infants having the longest hospital stays
and highest rates of ICU admission and mechanical ventilation use.[5 ]
[8 ]
[9 ]
[32 ] For example, in the SENTINEL1 observational study of 29 to 35 wGA infants aged < 12
months who were hospitalized for RSV and did not receive RSV IP during the 2014 to
2015 RSV season, 42% of the overall study population had an ICU admission and 20%
required mechanical ventilation.[5 ] Among infants aged < 3 months, these percentages were 68 and 44%, respectively.[5 ] Younger infants and preterm infants have also been shown to have greater RSVH costs.[5 ]
[22 ]
[32 ]
[33 ] Our results are congruent with these earlier findings and expand upon them by showing
that RSVH severity among preterm infants increased after the AAP policy change. Similar
observations were made in a retrospective study by Rajah et al, who found that several
metrics of RSVH severity increased from the 2013 to 2014 season to the 2014 to 2015
season among 29 to 34 wGA infants aged < 3 months, including ICU admission (from 30%
to 68%), use of mechanical ventilation (from 10% to 53%), length of hospital stay
(from 1.8 days to 8.8 days), and costs (from $19,686 to $30,662).[9 ] Although cost increases were numerically greater for preterm versus term infants
in our study, the relatively low number of preterm infants hospitalized for RSV or
bronchiolitis likely contributed to lack of statistical significance for these differences
from before to after the policy change. A possible explanation for increases in hospitalization
severity and/or costs observed among preterm infants in our study and others is that
infants hospitalized in 2014 to 2017 may have been at greater risk for more severe
RSV disease because of reduced use of RSV IP among higher risk infants in 2014 to
2017 compared with 2011 to 2014. We found that the use of mechanical ventilation increased
across both categories of infants in 2014 to 2017, although it remained more prevalent
among preterm infants. This may have been due to the development of newer mechanical
ventilation devices that are more efficient and easier to use, resulting in increased
acceptance of noninvasive ventilation as a standard therapy for infants with bronchiolitis[34 ] and decreased use of higher flow nasal cannula.
Study Limitations
This study has several limitations. As inpatient medication use is not captured by
claims data, inpatient RSV IP (typically provided at discharge from the birth hospitalization
during the RSV season) could not be examined. The proportion of infants who received
RSV IP may therefore have been underestimated, both because some infants may have
received it in the hospital, and because inpatient RSV IP may not have decreased to
the extent seen for outpatient use because it is not usually subject to payer prior
authorization. In addition, any claim for palivizumab was counted; information regarding
the number and timing of doses per infant was not available, and doses were not required
to be administered according to recommendations. It is possible that dosing as recommended
would have resulted in a greater effect on the study outcomes. RSVH and BH were identified
using ICD-9/10-CM diagnosis codes on inpatient claims because laboratory test results
are not available in the claims database. Although RSV diagnosis codes have previously
been shown to accurately reflect positive RSV test results in the emergency department
setting,[35 ] the number of RSVHs may be underestimated because the AAP does not recommend routine
virologic testing, which could reduce the use of RSV diagnosis codes (although the
effect of this limitation was mitigated by analyzing BH as well as RSVH). Similarly,
identification of ICU admission and mechanical ventilation use may be suboptimal because
of coding inconsistencies or undercoding due to bundling of inpatient hospitalization
services. The low sample sizes of infants hospitalized for RSVH and BH may have contributed
to an inability to detect statistical significance for some metrics of hospitalization
severity and costs. Finally, as no required length of continuous health plan enrollment
was specified, infants may have < 6 months of follow-up, and RSVH or BH occurring
after disenrollment would not have been captured.
Conclusion
The proportion of preterm infants receiving RSV IP decreased significantly after implementation
of the more restrictive policy issued by the AAP COID in 2014. After this policy change,
the already heightened risk of RSVH among preterm infants compared with term infants
increased further, confirming the findings of other nationwide U.S. analyses. In addition
to greater post–policy changes in RSV and BH rates for preterm infants, hospitalized
infants experienced illnesses that were more severe, particularly for infants aged < 3
months. These findings suggest that the youngest preterm infants are bearing a substantial
burden from increased RSV-associated disease after the 2014 AAP policy change.
