Trends in Childhood Asthma: Prevalence, Health Care Utilization, and Mortality

Asthma Prevalence

Asthma prevalence among 0- to 17-year-old children increased from 36 per 1000 children (SE 3.4) to 75 per 1000 (SE 5.1) from 1980 to 1995 but then decreased 17% to 62 per 1000 children (SE 5.9) in 1996 (Fig 1). Asthma prevalence increased by an average of 4.3% per year from 1980 to 1996. The 1997 estimate of childhood asthma attack prevalence from the redesigned questionnaire, 54 per 1000 children, and subsequent estimates cannot be compared directly with previous estimates and should be considered the first points of a new trend.¹¹ Because the redesigned asthma questions measure asthma attack prevalence (individuals who had previously received a diagnosis of asthma and who had 1 or more asthma attacks in the past 12 months) as opposed to asthma prevalence in the past 12 months, it is not surprising that the post-1997 estimates are lower than previous estimates. Asthma attack prevalence from 1997 to 2000 remained level with no statistically significant difference between estimates during this time period (SEs not shown). Table 1 shows childhood asthma prevalence for selected years by race/ethnicity and age group. The gap between black and white non-Hispanic children widened progressively during the period, from 15% higher asthma prevalence among black non-Hispanic children in 1980–1981 to 26% higher in 1995–1996. In 2000, black non-Hispanic children had an asthma attack prevalence rate 44% higher than that of white non-Hispanic children. From 1985/1986 to 1995/1996, asthma prevalence increased dramatically among Hispanic children. However, from 1997 to 2000, asthma attack prevalence among Hispanic children remained below that for white non-Hispanic children. Within the 3 pediatric age groups, prevalence increased over time. Children aged 0 to 4 years had the most rapid growth in asthma prevalence from 1980/1981 to 1995/1996. From 1997 to 2000, asthma attack prevalence remained fairly level among all age groups.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Trends in childhood asthma prevalence, hospitalizations, office visits to private physicians, and mortality, age <18 years.

Ambulatory Health Care Utilization

The rate for annual visits to private physician offices for childhood asthma followed an increasing trend during 1989 to 1999, the period for which annual data are available, and rose by an average of 3.8% per year (Fig 1). However, before this period, the asthma office visit rate declined by a total of 27% during the 9-year period from 1980 to 1989. Compared with white children, the asthma office visit rate in 1998–1999 for black children was 1.2 times higher (Table 2). Asthma office visit rates were higher in younger children and almost doubled among children 0 to 4 years of age between 1980/1981 and 1998/1999 compared with more modest increases among older children. Data for visits to hospital EDs and OPDs are shown in Table 2 for 1995 to 1999. Black children had visit rates to hospital OPDs and EDs approximately 3 times higher than those for white children in 1998–1999. When all sources of data for ambulatory visits for asthma are combined (shown in the far right columns in Table 2), black children had a visit rate 1.6 times higher than white children in 1998–1999. As with office visits to private physician offices, rates of hospital ED and OPD visits in 1995–1999 were higher among younger children compared with those aged 11 to 17 years.

Hospitalizations

The childhood asthma hospitalization rate grew slowly from 1980 to 1999, by an average of 1.4% per year. However, the hospitalization rate seems to have plateaued since the mid-1990s (Fig 1). Asthma hospitalization rates increased to a much greater extent among black children than white children. In 1998–1999, the asthma hospitalization rate among black children was 3.6 times the rate for white children (Table 3). The results of a sensitivity analysis show that if hospitalizations with unknown race are assumed to be for children of white race, then the hospitalization rate for black children in 1998–1999 was 3.2 times that for white children and hospitalization rates for black and white children increased 25% and 11%, respectively, from 1980/1981 to 1998/1999 (data not shown). Although asthma prevalence was lowest among the youngest children, hospitalization rates were substantially higher among 0- to 4-year-olds and increased more rapidly compared with older children. After 1995–1996, asthma hospitalization rates plateaued among all age groups.

Mortality

Asthma death rates increased by an average of 3.4% per year from 1980 to 1998 (Fig 1 and Table 4). After reaching a peak of 3.8 per 1 000 000 children in 1996, the childhood asthma death rate declined 18% in 1997 to 3.1 per 1 000 000 children. However, asthma mortality rose again in 1998 to 3.5 per 1 000 000 children. Black non-Hispanic children had the highest asthma death rates and the greatest increase over time. In 1985–1986, the death rate among black non-Hispanic children was 4.1 times higher than the death rate for white non-Hispanic children and in 1997–1998 was 4.6 times higher. Hispanic children had asthma death rates similar to those of white non-Hispanic children. Asthma mortality trends over time were generally similar among all age groups, but 11- to 17-year-old children had asthma death rates approximately twice those of younger children.

Diagnostic Transfer

A possible explanation for increasing asthma prevalence and health care utilization is diagnostic transfer-an increasing tendency over time of respiratory conditions to be labeled as asthma rather than as other conditions.^26,27 If the increase in asthma prevalence and health care utilization was mirrored by a decline in that for other respiratory conditions, then diagnostic transfer may be a plausible explanation for the trends observed since 1980. We assessed trends for bronchitis, bronchiolitis, and pneumonia because these conditions may be confused with asthma. From 1980 to 1996, the prevalence of chronic bronchitis increased by 2.3% per year on average compared with 4.3% per year for asthma prevalence. Although this pattern does not suggest diagnostic transfer, the pre-1997 NHIS data are based on parental report and not necessarily on physician diagnosis. From 1989 to 1999, the office visit rate for bronchitis and bronchiolitis decreased by 6.3% per year and for pneumonia by 9.9% per year, whereas that for asthma rose 3.8% per year. However, from 1980 to 1989, a period for which annual data are not available, the visit rate to private physician offices for bronchitis and bronchiolitis increased by a total of 81% during the 9- year period and for pneumonia by 91%, whereas the asthma visit rate decreased by 27%. The hospitalization rate for bronchitis and bronchiolitis increased by 0.7% per year from 1980 to 1999, whereas that for asthma increased by 1.4%. However, the pneumonia hospitalization rate decreased by 1.3% per year during this same period. Thus, there is an inverse relationship between trends in asthma office visits and hospitalizations and those for pneumonia, bronchitis, and bronchiolitis.

Dramatic Increases in Childhood Asthma

The burden of asthma on the pediatric population as measured by asthma prevalence, ambulatory visits, and mortality increased dramatically during the past 2 decades. The childhood asthma hospitalization rate increased more slowly but was still rising during a period when pediatric hospitalization rates for other causes were declining.^28–30 The factors behind the increasing asthma burden remain unclear.

The disproportionate increase in asthma prevalence, health care utilization, and mortality for children younger than 5 years, for whom accurate diagnosis may be more difficult, raises the question of changes in diagnosis over time. The majority of episodes of wheezing in infants are related to transitory conditions and are not associated with increased risk of asthma later in life.³¹ It is possible that over time, asthma was more likely to be diagnosed in infants during these transient episodes because of greater awareness of the diagnosis and treatment of asthma among parents and health professionals. However, it is unlikely that changes in diagnostic awareness in younger children completely account for the sustained increase in asthma prevalence in older age groups as measured by the NHIS, especially because the NHIS focused on asthma in the 12 months before the survey.

Diagnostic transfer is another possible change in labeling of respiratory conditions in children that could contribute to the increase in asthma prevalence and health care utilization. Our analysis of trends in bronchitis prevalence cannot be used to definitively support or refute the theory that an increasing tendency to transfer diagnoses from bronchitis to asthma over time has resulted in increased asthma prevalence. In contrast, trends in health care utilization show that rising asthma office visit rates from 1989 to 1999 were roughly mirrored by decreases in those for bronchitis, bronchiolitis, and pneumonia, and increasing asthma hospitalization rates from 1980 to 1999 by decreasing pneumonia hospitalization rates. These patterns suggest that diagnostic transfer has contributed to the increasing trend in asthma health care utilization. However, more detailed investigation is necessary to examine the influence of other factors that could also contribute to these patterns. For example, efforts to reduce hospitalizations for pneumonia and bronchitis may have been more successful than those to reduce asthma hospitalizations.

Another possible explanation for increasing asthma prevalence and morbidity is changing environmental exposure.²⁷ Some have hypothesized that urbanization leads to multiple exposures (both indoors and outdoors) that increase the risk of developing asthma.^{26, 32–35} However, evidence in this area is mixed; 1 study demonstrated that both urban and rural populations underwent similar increases in asthma prevalence during the 1980s.³⁶ Research to elucidate the causes of asthma continues.

Recent Trends Suggest That the Asthma Burden Has Plateaued

Just as puzzling as the prolonged increase in asthma prevalence, health care utilization, and mortality is the recent plateauing of some indicators and the apparent decrease of asthma prevalence after 1995. The interpretation of trends in asthma prevalence has been complicated by the redesign of the NHIS in 1997. Although the redesign will improve the validity of asthma surveillance by requiring a diagnosis by a health professional, it created a break in the trend of asthma prevalence¹¹ and generated uncertainty about how to interpret the drop in asthma prevalence in 1996. The 1996 sample was reduced by 40% as a result of pilot testing of the 1997 redesigned survey, and the impact of this decreased sample size on prevalence estimates is unclear. Although asthma hospitalizations and mortality have plateaued since 1995, these recent trends cannot be used to “ confirm” a change in the prevalence trends because multiple factors affect asthma morbidity and mortality. For example, changes in hospitalization rates may also reflect changes in medical practice, asthma therapy, and access to and utilization of care.^{5, 26,28,37} Mortality may reflect availability and utilization of health care, access to and correct use of medication and prevention strategies, and severity of disease.^7,38 Therefore, additional years of data are necessary to determine whether the pattern of increasing asthma prevalence during the past 2 decades has changed.¹¹

One possible reason for the recent plateauing of hospitalizations and deaths may be the impact of clinical and public health intervention and prevention efforts. Although it is difficult to evaluate the efficacy of specific or local prevention programs with national data, the impact of nationwide programs, such as the State Children’s Health Insurance Program that was passed as part of the Balanced Budget Act of 1997, may be discerned. A large nationwide increase in enrollment of children in insurance programs (2 684 300 children were enrolled by September 1998)³⁹ may be partly responsible for the large jump in asthma office visits in 1998. Another recent nationwide development is the dissemination of the National Asthma Education and Prevention Program Guidelines in 1997.⁶ It has been theorized that adoption of these guidelines by medical practitioners was responsible for the recent decline in asthma mortality among all ages in the United States in 1997.⁷

Disparities in Asthma Burden Among Children

As documented in several studies,^{9,29, 34,38,40–47} prevalence, morbidity, and mortality are higher among black children compared with white children. Although much has been made of racial disparities in asthma prevalence, much larger disparities exist for morbidity and mortality⁴⁶ (L. J. Akinbami and K. C. Schoendorf, National Center for Health Statistics, unpublished manuscript). Risk of asthma hospitalization and mortality has been found to be correlated with minority race and living in poverty.^{25, 27,33,42, 48} Although most poor children have access to a source of asthma care,^{41,49, 50} poor and minority children with asthma are less likely to receive care in a high-quality setting with continuity of care^{41, 49,50} and less likely to be prescribed or to use maximally effective preventive therapy.^{33,49, 51,52} Furthermore, some studies suggest that after adjusting for severity of disease, even greater discrepancies in health care utilization exist between poor and nonpoor children and between black and white children⁴¹ (L. J. Akinbami and K. C. Schoendorf, National Center for Health Statistics, unpublished manuscript). Such differences in utilization of and access to health care are likely to contribute to the racial and income disparities in asthma morbidity and mortality that have been widely described in the literature.

Strategies to Expand Childhood Asthma Surveillance

There are regional and local variations in asthma prevalence, hospitalization, and mortality that national data do not reveal.^{8,9,42, 53} As demonstrated by studies focusing on regional variations in asthma hospitalization trends, ^5,28 examining smaller population groups exposes important differences in practices and outcomes and facilitates translating surveillance findings into policy. Strategies to increase local surveillance include analyzing existing data on hospital discharges or billing information and standardized household surveys.⁸ The Behavioral Risk Factor Surveillance System is a random-digit-dialed telephone survey operated in all 50 states, the District of Columbia, and Puerto Rico. The Behavioral Risk Factor Surveillance System has core questions about health conditions (including asthma) and an optional standardized module beginning in 2001 that states may use to gather more detailed information on asthma, including childhood prevalence. Although asthma prevalence data from the core survey are collected only for adults, some states added questions about childhood asthma. However, the questions were not standardized between states. A standardized asthma module to measure childhood asthma prevalence may be available for states to use in the future (J. Moorman, National Center for Environmental Health, Centers for Disease Control and Prevention, personal communication, January 16, 2001). The NCHS is also launching an asthma survey in the State and Local Area Integrated Survey, a telephone survey that will use a uniform questionnaire for all states and will provide state-specific information about asthma for children and adults.

Limitations

It is not possible using NCHS data to detect multiple hospitalizations and office visits for each patient. Therefore, the extent to which changes over time reflect the changes in the frequency of recurrent asthma office visits or hospital readmissions is unknown. It is unlikely, however, that readmissions are primarily responsible for observed trends given asthma readmission rates reported elsewhere.⁵⁴ In addition, it is unclear how readmissions may affect the observed racial disparity. That is, it is unknown whether children of different race/ethnic backgrounds have a different level of risk for readmission.

Missing race data are an important limitation for the NAMCS, NHAMCS, and NHDS. Race for all of these surveys is obtained from medical records, in contrast to the NHIS for which race is self-reported. For the approximately 9% of the NAMCS and NHAMCS records missing race, NCHS imputed this information.^13,16,17 Given the limitations in the data collected on race, the analysis by race for asthma ambulatory visits must be interpreted with caution. For the NHDS, the 5% to 22% of hospital discharge records with missing race data are most likely to represent white patients.²¹ There is no ideal solution to adjust for underreporting of race in the NHDS; therefore, data by race should be interpreted cautiously. However, there is evidence that a sizable racial disparity in asthma hospitalizations exists given the robustness of the disparity in the sensitivity analysis for unreported race in the NHDS and the racial disparities reported in several other studies using different data sources.^28,30,33,42