Trends in Outpatient Influenza Antiviral Use Among Children and Adolescents in the United States

We conducted a cross-sectional study using the MarketScan Commercial Claims and Encounters Database (IBM Watson Health, Armonk, NY), which captures data on paid commercial insurance claims for all covered health care encounters, including both outpatient and inpatient settings in all 50 US states. We included data collected from those aged 0 to 18 years between July 1, 2010 and June 30, 2019. For children >1 year of age, we required continuous enrollment during the 12 months before antiviral dispensing to comprehensively capture health care utilization and accurately classify comorbidities. To assess the overall real-world usage of antiviral agents, we assessed all eligible individuals throughout the calendar year and did not restrict data capture to influenza season, individuals with ILI symptoms, or those with comorbidities. This study of deidentified data was considered exempt from review by the Vanderbilt University Medical Center and the University of Illinois Chicago Institutional Review Boards.

Antiviral exposures were identified by using outpatient pharmacy dispensing claims for oseltamivir, zanamivir, and baloxavir. Antiviral use in the inpatient setting was not assessed as hospital dispensings and is not captured in the database. Peramivir was excluded because it is an intravenous medication with limited use in the nonhospitalized population. To match national guidelines during the study period in which acute treatment duration was 5 days and recommended prophylactic duration ranged from 7 to 10 days for oseltamivir and 10 days for zanamivir,^12,13  treatment of acute infection was defined as a dispensing for ≤5 days and prophylactic use was defined as treatment for >5 days. Baloxavir is administered as a single dose for acute and prophylactic use. Given the likely limited use of baloxavir for prophylactic use in children, any dispensing of these medications was defined as acute use. For antiviral dispensing associated with influenza infection, an antiviral was required to be dispensed within 2 days of influenza diagnosis date and only during influenza season.

Qualifying influenza diagnoses were identified on the basis of outpatient influenza International Classification of Diseases, Ninth Edition and International Classification of Diseases, 10th Edition (ICD-10) primary diagnoses codes.^14,15,36  Recent studies reveal that ICD-10 codes have greater validity than the International Classification of Diseases, Ninth Edition codes for influenza infection,^37–40  and influenza ICD-10 codes have been successfully implemented within influenza surveillance systems.^40,41  ILI was defined by using the Armed Services ILI diagnosis code definition.⁴²  We evaluated both the influenza-specific and ILI definitions because of AAP, CDC, and IDSA recommendations for the treatment of individuals with ILI during influenza season regardless of laboratory-confirmed influenza infection.

Information on weekly activity for each defined influenza season was from the CDC Flu Activity & Surveillance program, including the Influenza-Like-Illness Surveillance Network (ILINet).⁴³  Each influenza season was defined as the period from October 1 of one year through March 30 of the next year, so our study included 2010 to 2011 through 2018 to 2019 influenza seasons. Geographic regions in the United States were categorized by geographic census division. Influenza activity was reported as the weighted percentage of visits for ILI for the census division as reported by ILINet. Influenza season severity for the 2010 to 2011 through 2018 to 2019 seasons was reported by using previous CDC seasonal severity classifications.^44,45 

The primary outcome was the rate of antiviral dispensing defined as the number of dispensed pharmacy claims for any influenza antiviral (oseltamivir, zanamivir, or baloxavir) divided by the total number of enrolled children. Antivirals were identified by using National Drug Codes. The rate of acute influenza treatment was calculated by dividing the number of dispensings with a supply of ≤5 days by the total number of enrolled children. The rate of prophylactic use was calculated by dividing the number of dispensings with a supply of >5 days by the total number of enrolled children. As a secondary outcome, we calculated the rate of treatment among those with an influenza diagnosis by dividing the number of antiviral dispensings by the total number of children with an influenza diagnosis. Children at a high risk of influenza complications were identified by using previously published ICD-10 diagnoses.^{14,28,46–50} 

The secondary outcome of antiviral costs was defined as the cost per dispensing claim, adjusted to 2019 US dollars by using the Medical Care Component of the Consumer Price Index.^51,52  Paid dispensing claims were identified, and adjusted costs were summed to estimate the total antiviral costs per influenza season. Cost data were unavailable for the 2018 to 2019 influenza season and were excluded from the analysis.

Descriptive statistics are presented as mean and SD, as well as median and interquartile range (IQR) for continuous variables and as counts and percentages for categorical variables. Results were stratified by age (<2, 2–5, 6–11, 12–17 years), geographic region, treatment indication (acute or prophylactic), and influenza season.

The analysis included 1 416 764 unique antiviral dispensings between 2010 and 2019. More than 97% of antiviral dispensing occurred during influenza seasons, and antiviral exposure trended closely with CDC influenza surveillance activity (Fig 1). The median age of children treated with antiviral agents was 9 years (IQR 5–13), and antiviral dispensing was more common in boys and older children. The most common location of care was outpatient office visit followed by urgent care (Table 1). Individuals with risk factors for influenza complications comprised 63.3% of the study population.

Oseltamivir was the most frequently prescribed antiviral (>99%), followed by zanamivir and baloxavir (Table 1). Antiviral treatment (n = 1 008 099; 71.2%) was more common than prophylaxis (n = 408 665; 28.8%). Most antiviral dispensing had an influenza (68.1%) or ILI (81.4%) diagnosis within 2 days of antiviral dispensing. Antiviral dispensing varied by season and was highest during the 2017 to 2018 season and lowest during the 2011 to 2012 season. The most common durations of antiviral dispensings for oseltamivir were 5 (71.6%) and 10 (14.6%) days, followed by 6 days (6.0%) and 7 days (1.5%). The most common diagnoses associated with antiviral dispensing were influenza, fever, and acute pharyngitis (Supplemental Table 3).

The overall dispensing rate during the entire study period was 20.6 antiviral dispensings per 1000 children per season, ranging from 4.35 to 48.6 dispensings per 1000 children per season, and generally increased over the study period (Table 2). The overall dispensing rate per 1000 influenza diagnoses was 606, ranging from 494 to 673 per influenza season (Table 2). Antiviral treatment among those with an influenza diagnosis (Fig 2A) and prophylactic treatment among enrolled children (Fig 2B) rates varied by age, with children ≥12 receiving more acute antiviral treatment and children 2 to 11 years of age receiving more prophylactic dispensing. The rate of antiviral treatment among children <2 years of age with an influenza diagnosis was 367 dispensings per 1000 influenza diagnoses. In children <2 years of age with an influenza diagnosis, prophylactic dispensings were 233 per 1000 influenza diagnoses (Supplemental Fig 4). Variations in age-based rates of antiviral treatment per 1000 influenza diagnoses (Supplemental Fig 5) and prophylaxis per 1000 enrolled children (Supplemental Fig 6) were similar across influenza seasons.

Geographic variation in dispensing rates was illustrated by using the most recent influenza season, 2018 to 2019 (Fig 3). During that season, there was a threefold variation in dispensing rates, ranging from 25 per 1000 enrolled children in the Pacific region to 78 per 1000 enrolled children in the East South Central region. The median weighted percentage of visits for ILI in these regions, as reported by the CDC ILINet during the same season, was 2.48 (Pacific) and 3.39 (East South Central), a ∼37% difference. Geographic dispensing rates varied by season, and the differences between the highest and lowest regional rates were substantial, ranging from a 3.1-fold difference (2018–2019) to a 21-fold difference (2010–2011; Supplemental Fig 7A–F).

The overall adjusted prescription costs during the study period totaled $208 458 979 (median adjusted cost $138). Inflation-adjusted costs for individual prescriptions varied and ranged from $111 (IQR 92–113) during the 2010 to 2011 season to $151 (IQR 123–299) during the 2016 to 2017 influenza season. The 2017 to 2018 season had the highest costs ($46 345 486), and the 2011 to 2012 season had the lowest costs ($7 838 740; Supplemental Table 4).

This large, cross-sectional study describes national trends in influenza antiviral treatment among children and adolescents in the United States. There is a wide variability in the use of influenza antiviral agents in the pediatric population, including by age, geographic region, presence of high-risk conditions, and across influenza seasons. These findings suggest that antiviral prescribing may be a high-value target for care improvement among children and adolescents with influenza.

The AAP, IDSA, and CDC strongly recommend antiviral treatment in children <5 years of age, especially in children <2 years of age, regardless of the duration of their symptoms.^13,23  Antiviral treatment decreases the duration of influenza symptoms when it is started <48 hours from symptom onset, and it is associated with fewer influenza complications, such as otitis media, pneumonia, and hospitalization in children.^19,20,22  Despite the fact that symptomatic influenza was more common in children <6 years of age than those 6 to 17 years old during the study period,⁵³  antiviral treatment was more common among children >12 years of age. In our study, 37% of children <2 years of age and 34% of children 2 to 5 years of age with an influenza diagnosis received antiviral treatment. A 2019 study of prospectively enrolled individuals during the 2009 to 2016 influenza seasons revealed that ∼10% of children <2 years of age and ∼20% of children 2 to 17 years of age received antiviral treatment.⁵⁴  Our finding of low rates of antiviral use among outpatient children <2 years of age reflects low rates of antiviral use among hospitalized children <2 years of age, as well. A 2022 study revealed that from 2007 to 2020, 62.2% of hospitalized children <2 years of age received oseltamivir treatment.²⁹  The reasons for under-prescribing in this high-risk population of children <2 years of age are unclear and likely multifactorial. Qualitative work by the CDC revealed a number of factors in under-prescribing influenza antiviral, including wide range perception of efficacy, differences in interpretation of testing, and misunderstanding of the national guidelines.⁵⁵  Finally, there may be a concern for adverse drug events associated with oseltamivir use in children, such as neuropsychiatric events or vomiting that are listed on the manufacturer label. However, recent evidence does not support an association between oseltamivir and suicide attempts.³³  In addition, a 2023 study evaluating influenza-associated serious neuropsychiatric events revealed similar rates of neuropsychiatric events among those treated and not treated with oseltamivir.¹⁴ 

These findings are especially concerning given that influenza disproportionally affects young children and results in a greater number of clinic visits, emergency department visits, influenza complications, and hospitalizations in children <5 years of age compared with older children and adolescents.¹¹  The low rate of antiviral use in young children in this study, along with recent evidence of low guideline-concordant antiviral treatment in children at a high risk of influenza complications,²⁷  highlights the need for improved influenza management in the most vulnerable children in the United States.

There is limited information on geographic patterns in influenza antiviral use. A study of antiviral use during the 2008 to 2009 and 2009 to 2010 influenza seasons revealed higher rates of oseltamivir use in the South and lower rates on the West Coast.²⁶  That same study also revealed the highest regional growth in antiviral use among the West and Northeast. Our study of 9 influenza seasons revealed similar geographic differences in antiviral dispensing, with the highest rate in the East South Central region and the lowest in the Pacific region. We also found consistent geographic variation in antiviral prescribing across influenza seasons, but these differences varied markedly by influenza season. The magnitude of these differences in antiviral use, ranging from 3- to 21-fold, cannot be explained by geographic differences in influenza incidence. For example, in the 2010 to 2011 season, there was a 21-fold difference in antiviral dispensing rates between New England (2 per 1000) and East South Central (42 per 1000). However, during the same season, the median weighted percentage of visits for ILI for the region, as reported by the CDC ILI Surveillance Network, were more similar (New England 1.0% and East South Central 2.4%) Interestingly, the relative differences in antiviral dispensing rates across regions is similar to that of antibiotic rates over the same study periods, with the lowest rates of community antibiotic prescriptions in the West and highest in the East South Central.⁵⁶  Regional trends do not seem directly correlated with influenza vaccine use, in which the East South Central has the lowest rate of vaccine uptake and the Pacific region has the median rate of influenza vaccination.⁵⁷  A better understanding of clinical and societal factors that drive adherence to national treatment guidelines is needed to reduce unnecessary variation in influenza treatment in the pediatric population.

After adjusting for inflation, the cost of influenza antiviral agents has increased over the past 10 years with a concomitant increase in utilization. There were intermittent oseltamivir/Tamiflu shortages over the study period. Drug shortages typically suggest either a lack of manufacturers to meet demand or supply chain issues resulting in decreased production by manufacturers. The first generic version of oseltamivir was FDA-approved in August 2016, and the marginally decreased cost of antiviral agents in the final year of available financial data (2017–2018) may reflect generic manufacturers entering the market.⁵⁸  Our data, in conjunction with medication shortages during multiple influenza seasons, suggest there is not enough generic competition to change the price of oseltamivir. As more generic options become available, it is possible that there may be more substantial decreases in antiviral costs. In addition, better availability of other antiviral agents, including baloxavir and zanamivir, and the development of newer antiviral agents to increase treatment options for providers may remove barriers to appropriate antiviral use in children. Additional study is needed to determine if the cost of oseltamivir will continue to downtrend or if rising baloxavir use, which entered the market in 2018 and does not have a generic version, will emerge as a mitigating factor. Baloxavir may be an appealing option in the pediatric population because of its single-dose administration.

This study has several limitations. The IBM Marketscan database only contains medication dispensing with an insurance pharmacy claim. It is possible that some individuals may have supplemental insurance or pay out of pocket for influenza antiviral agents. For the same reason, we could not identify prescriptions that were written but never filled. There were intermittent oseltamivir shortages over the study period, and we could not assess the role of its availability in its dispensing data. The database does not contain laboratory results, and we were not able to determine individual-level laboratory-confirmed influenza infections. This limitation is mitigated by the real-world nature of the study design and national guidelines recommending empirical antiviral treatment in those with influenza symptoms during influenza season, even without laboratory confirmation. We were also unable to identify some children at a high risk of influenza complications, including those with prolonged aspirin use, those of Alaskan or Indigenous heritage, and hospitalized individuals. Finally, there may be misclassification of some cases of prophylactic use for 2 reasons. First, we assigned all baloxavir use to the acute treatment group and likely underestimated prophylactic use among baloxavir users. Second, we classified all oseltamivir use greater than the 5-day recommended treatment course in the prophylaxis category. Some uses of >5 days of oseltamivir may have been for acute treatment, resulting in an overestimation of prophylactic use. Despite these limitations, our findings highlight the need to improve care for children with influenza through the judicious use of available antiviral agents, especially for children at a high risk of influenza complications.

There is wide variability in the use and rising cost of influenza antiviral agents in the pediatric population. The rates of guideline-concordant treatment of young children at a high risk of influenza complications are low, and there is marked variation in influenza antiviral treatment by geographic region. These findings suggest a target area for care improvement in the treatment of children with influenza.

AAP

American Academy of Pediatrics

CDC

Centers for Disease Control and Prevention

FDA

Food and Drug Administration

ICD-10

International Classification of Diseases, 10th Edition

IDSA

Infectious Disease Society of America

ILI

influenza-like illness

ILINet

Influenza-Like-Illness Surveillance Network

IQR

interquartile range