The prevalence of childhood asthma: interpreting falling rates in the context of shifting measurement and the COVID-19 pandemic

Purpose of review

The estimated prevalence of childhood asthma in the United States, as measured by the National Health Information Survey (NHIS), has decreased by 30% since 2017. This review provides context for observed changes in asthma rates by describing recent shifts in NHIS data collection and analysis, and considers whether the COVID-19 pandemic might impact asthma prevalence in years to come.

Recent findings

The NHIS underwent a planned redesign in 2019 with updated sampling weights to better match the U.S. population. In early 2020, the COVID-19 pandemic resulted in unplanned modifications to NHIS implementation, which may have included fewer children from populations at a heightened risk for asthma. Decreasing prevalence estimates in recent years are likely at least in part due to these survey changes rather than true epidemiologic shift. However, pandemic-related changes to risk factors for childhood asthma (including exposure to rhinovirus infections and allergic sensitization) may also influence prevalence in the future.

Summary

Recent changes in estimated rates of childhood asthma in the USA are likely driven by changes to survey methods and implementation, both before and during the COVID-19 pandemic. Additional years of data are needed to determine whether a true shift in disease prevalence is occurring.

INTRODUCTION

Asthma is one of the most common chronic diseases of childhood. Prevalence estimates in the United States are based on weighted responses to the National Health Information Survey (NHIS), an annual cross-sectional assessment of U.S. households conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention (CDC). The percentage of U.S. children with asthma identified by the NHIS gradually increased until 2009. Despite mild variation in estimated prevalence over the past 20 years, however, most surveys since 2001 have reported rates of 8–9% [1]. A review of more recent data suggests that overall prevalence may now be in a state of flux. NHIS estimates have rapidly decreased over the past several years, from 8.4% in 2017 [2] to 7.5% in 2018 [3], 7.0% in 2019 [4] and 5.8% in 2020 [5]. Fewer children with asthma would be an encouraging development, yet caution is warranted when interpreting newer prevalence data. In particular, estimates from 2019 onwards need to be considered within the paired contexts of a major NHIS redesign rolled out in 2019 [6], and the global coronavirus disease 2019 (COVID-19) pandemic, which impacted survey methods starting in 2020. 

Box 1.

no caption available

RECENT CHANGES TO NATIONAL HEALTH INFORMATION SURVEY METHODS, AND SURVEY IMPLEMENTATION DURING THE COVID-19 PANDEMIC

Participation in surveys administered by the U.S. Department of Health and Human Services (which includes the CDC) has fallen over time [7]. The NHIS has been particularly affected by this shift towards lower engagement, with a nearly 20-percentage point decrease in response rates between 1997 and 2014. To address concerns about nonresponse bias affecting the reliability of population-based estimates, the NHIS adjusted sampling weights in 2019 to better match the general U.S. population [8]. Additional changes in the 2019 survey included a shift to randomly selecting one adult and child per household for inclusion (rather than one adult and child per family), and the removal of a question asking respondents to identify a single/primary racial category.

The reported rate of asthma among U.S. children decreased once these changes were implemented. A similar pattern was observed after the previous NHIS redesign in 1997, when survey questions were altered to better capture physician-diagnosed asthma and asthma attacks. Although the prevalence of childhood asthma was lower for a couple of years following the 1997 changes, the CDC advised that newer rates were not directly comparable with historical estimates as a result of changed methods [1,9]. Data collected after the recent redesign are similarly difficult to compare with pre-2019 data [10].

A year after this redesign was completed, established methods for administering the NHIS required rapid adaptation in response to the global COVID-19 pandemic. Data were collected in typical fashion during in-person home visits throughout the first quarter of 2020. After a national emergency was declared in the USA in March 2020, the need to mitigate risk of infection led to the NHIS being conducted exclusively by telephone until halfway through the year. Thereafter, telephone calls were placed first, with home visits conducted as a back-up option as needed [11]. These methodological shifts resulted in greater nonresponse bias and significant changes to the surveyed population. Compared with respondents from the second quarter of 2019 and first quarter of 2020, the population reached by phone starting in the second quarter of 2020 was significantly less likely to report an income less than $35 000; more likely to own their home; and more likely to have lived in their current home for several years [12]. Previous studies have demonstrated an increase in asthma prevalence among children living in poverty [1] and children whose families rent their housing [13]. The changes to NHIS methods in 2020, while necessary, may have therefore inadvertently captured a sample of children at a lower risk of asthma, and could help explain the apparent drop in asthma prevalence from 7.0 to 5.8% in a 1-year period.

COMPARING RECENT PREVALENCE DATA WITH HEALTHCARE UTILIZATION

With the combined impact of the 2019 redesign and 2020 implementation changes due to the pandemic, it may take several years before a reliable trend line for childhood asthma prevalence is available. Prior to the pandemic, healthcare utilization data might have provided an approximate (if imprecise) reference against which to compare observed changes in disease rates and determine whether a true epidemiologic shift was occurring. Historic estimates of asthma prevalence from the NHIS were generally concordant with Medicaid administrative claims data for asthma care [14]. However, this type of analysis will not provide insight into more recent prevalence trends spanning 2020–2021. The marked drop in asthma exacerbations in the USA starting in March 2020 [15] did not occur as a result of changes in underlying prevalence, but rather because of widely implemented pandemic control measures: school closures, mask wearing and social distancing led to decreased exposure to common environmental triggers of asthma exacerbations including respiratory virus infections [16^▪].

Childhood asthma symptoms and associated healthcare use typically fluctuate over the calendar year, with well described spikes in the autumn and early spring [17,18]. Interestingly, after the pandemic started, in the fall of 2020, there was a notable lack of increase in paediatric emergency department (ED) visits for asthma and only a blunted increase in hospitalizations [19^▪]. Other studies observed a similar loss of seasonal variation in asthma visits, and decreased rates of exacerbations requiring treatment with systemic corticosteroids for at least one year after the pandemic started [20^▪,21^▪]. Visits to outpatient paediatric clinics for sick visits and well child checks alike fell during the first months of the pandemic before normalizing in later 2020 [22,23^▪,24]. Collectively, these findings suggest that there will be fewer children with claims for asthma-related care in 2020 administrative datasets, even without changes to underlying prevalence rates. Comparing prevalence estimates with asthma-specific claims data in the coming years will continue to defy easy interpretation as rates of exacerbations and acute care visits normalize to prepandemic levels.

EXPOSURE TO RISK FACTORS FOR DEVELOPING ASTHMA DURING THE PANDEMIC

Although recent shifts in the measured rate of childhood asthma prevalence are likely explained at least in part by survey changes, it is also worth considering whether the global pandemic might have altered exposure to risk factors for developing asthma that could influence the true disease prevalence in years to come. Respiratory tract infections with non-COVID-19 viruses [including influenza, respiratory syncytial virus (RSV) and rhinovirus] are a frequent cause of exacerbations among children who already have asthma [25^▪]. There is compelling evidence that these common respiratory viruses may also contribute to the development of asthma: at-risk children who wheeze in the first 3 years of life due to a rhinovirus infection are significantly more likely to develop asthma later in childhood [26,27].

A marked decrease in paediatric respiratory infections due to non-COVID-19 viruses was observed starting in March 2020, following the widespread implementation of societal lockdowns [21^▪]. With fewer rhinovirus infections in the immediate postlockdown period, the overall burden of rhinovirus-associated wheezing illnesses is likely to have been lower for children who were less than 3 years old at the time. However, although influenza and RSV rates remained low during the first autumn and winter following lockdown, rhinovirus infections rebounded to prepandemic levels by the fall of 2020 [21^▪,25^▪]. The rapid return to normal rhinovirus infection levels, likely related to decreased susceptibility of rhinoviruses to certain infection prevention measures (i.e. hand sanitizer, masking) and a lack of viral interference from influenza [25^▪], suggests that any potential decrease in the risk of asthma development among these youngest children is likely to be small and temporary.

Patterns of atopic disease during the pandemic also need to be considered, as underlying atopy is a critical contributor to allergic asthma development and exacerbations. A history of atopic dermatitis and/or sensitization to aeroallergens or egg by 1 year of age increases a child's likelihood of developing asthma [28], and the risk is higher still among children with allergic sensitization who also experience rhinovirus-associated wheezing prior to age 3 [26]. One study from Ireland found that children born near the start of the pandemic (March–May 2020) were at an increased risk of atopic dermatitis and egg allergen sensitization by age 12 months, in comparison with historical controls [29]. Another study conducted in China found that sensitivity to select aeroallergens and food allergens, as well as elevated total IgE levels, increased within the study cohort in 2020 as compared to prepandemic years [30]. Additional studies following children over many years are needed to determine whether aeroallergen sensitivity and rates of atopy among young children were impacted by the pandemic. In contrast with the temporary decrease in rhinovirus infections, an increase in allergic sensitization and resultant atopic disease early in life could hypothetically increase rates of childhood asthma development in the future.

THE IMPORTANCE OF RELIABLE PREVALENCE DATA

To be certain, prevalence is only one measure of disease burden within the general population. Decisions to prioritize specific health conditions for research and public health efforts often consider other important metrics beyond prevalence alone. Federal research funding awarded by the National Institutes of Health has been associated more with condition-specific disability-adjusted life-years (DALYs) than prevalence [31]. Health systems are likely to prioritize initiatives to address conditions based on cost and patterns of healthcare utilization. Prior to the pandemic, poorly controlled asthma was a leading cause of paediatric ED visits and hospitalizations each year [32,33]. Children who require treatment for asthma in an ED or hospital setting are at increased risk for needing additional acute healthcare services in the future [34–37]; interventions to improve outcomes are thus designed to consider the small group of high utilizers who have more severe disease [36] or use the ED as a preferred source of asthma care [38], rather than the full population of U.S. children with asthma.

Even with other metrics often utilized, the value of reliable, longitudinal prevalence data remains. Monitoring rates of asthma in the general population is essential for understanding disparities in disease rates, severity and health outcomes. In the USA, black children with asthma are at a significantly greater risk than white children of having exacerbations that lead to ED visits, hospitalizations or even death [39–43]. When asthma outcomes are analysed using census-based data representing the full population of U.S. children, disparities continue to worsen over time [1]. Yet, asthma is not evenly distributed across the full population. Children from historically marginalized groups, including those with non-Hispanic Black or Puerto Rican ancestry, are more likely to develop asthma [1]. Differences in asthma prevalence rates between sub-groups of the larger U.S. population are informed by structural racism and inequities in American society [44^▪▪], factors which also contribute to morbidity and confound analyses that rely on census-based population data [45]. When outcomes comparisons are limited to children who are at-risk for asthma morbidity (based on NHIS prevalence data), a different picture emerges, and indicates that disparities for black children may be improving [1]. Analyses that favour at-risk rates over population-risk rates also provide greater insight into the social determinants of health that drive disparities, so that targeted interventions can be developed to promote equity in health outcomes [45]. This is the importance of accurate prevalence information: we cannot appropriately study or address disparities without a clear understanding of who is at the greatest risk.

CONCLUSION

After a nearly 20-year period of relative stability, recent data suggest that the prevalence of childhood asthma is rapidly decreasing in the United States. This apparent trend is likely explained by changes to NHIS design and implementation (both planned and unplanned) rather than a true shift in disease burden, and should be interpreted with caution. Our healthcare community may expect to see additional variation in estimated prevalence. Future shifts in prevalence might be informed by a return to traditional survey implementation practices as pandemic restrictions relax, or by pandemic-driven differences in child exposure to risk factors for developing asthma. The impact of climate change on childhood asthma will also merit additional consideration in the years to come. The rise in global temperatures may herald changes in routine exposure to environmental risk factors for asthma, with worsening outdoor air quality, altered growth patterns of common plant aeroallergens and increased time spent indoors in proximity to home health hazards (i.e. cigarette smoke, cockroaches or mould) [46]. It will be our shared task to ensure that survey methods continue to optimize equitable representation, and that future analyses include careful consideration of the context in which prevalence data are obtained.

Acknowledgements

None.

Financial support and sponsorship

This work was completed without any specific support.

Conflicts of interest

None of the authors have any conflicts of interest to disclose.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest

• ▪▪ of outstanding interest