Racial-Ethnic Disparities in Childhood Asthma Treatment and Outcomes: The Role of Disease Severity

Slide Presentation from the AHRQ 2008 Annual Conference

---

On September 8, 2008, G. Edward Miller, Julie L. Hudson, and Jim Kirby, made this presentation at the 2008 Annual Conference. Select to access the PowerPoint® presentation (776 KB).

---

Slide 1

Racial-Ethnic Disparities in Childhood Asthma Treatment and Outcomes: The Role of Disease Severity

G. Edward Miller, Julie L. Hudson, and Jim Kirby
AHRQ Conference
September 8, 2008

Notes:

The title of my presentation today is Racial and Ethnic Disparities in Childhood Asthma Treatment and Outcomes, the Role of Disease Severity. My co-authors in this research Julie Hudson and Jim Kirby are my colleagues at AHRQ.

Slide 2

Percent of Children (5-17) with Asthma by Race-Ethnicity

The bar graphs show:

• Children with Treatment for Asthma.
  • White: 7.1* percent.
  • Black: 10.2 percent.
  • Hispanic: 7.1* percent.
• Children with 'Asthma Now.'
  • White: 8.1* percent.
  • Black: 11.9 percent.
  • Hispanic: 7.3 percent.
• Source: MEPS, 2000-2005.
• Note: *P < .05 for difference with black children.

Notes:

• Asthma affects quality of life for an estimated 5 to 6 million school-aged children in the United States.
• In this research, we use two methods of identifying children with asthma. As this slide shows, black children are more likely than white and Hispanic children to have parent-reported treatment for asthma during the year and are also more likely than other children to be identified as currently having asthma, or to have 'asthma now'.
• I will give more detail on these measures later in the presentation.

Slide 3

Percent of Children (5-17) with Poorly Controlled Asthma by Race-Ethnicity

The bar graphs show:

• Any Missed School for Asthma?
  • White: 34.9 percent "yes."
  • Black: 54.3* percent "yes."
  • Hispanic: 54.7* percent "yes."
• Any Emergency Room (ER)/IP Visits for Asthma?
  • White: 4.4 percent "yes."
  • Black: 13.8* percent "yes."
  • Hispanic: 8.2* percent "yes."
• Source: MEPS, 2000-2005.
• Note: *P < .05 for difference with white children.

Notes:

• The focus of this research is treatment and outcomes among children who are identified as having asthma.
• Among children with asthma, Black and Hispanic children are both more likely than white children to have poorly controlled asthma which results in a number of negative outcomes.
• As this slide shows, among children who received treatment for asthma, black and Hispanic children are both about 50% more likely than white children to miss school because of the condition. Further, Hispanic children were almost twice as likely and black children were more than three times as likely as white children to have an emergency room visit or inpatient stay to treat asthma.
• When we examine these same outcomes for conditions other than asthma we find that white children are more likely to miss school for conditions other than asthma and we find no differences across groups in ER / IP use.
• The findings for conditions other than asthma further highlight the impact of asthma on minority children and suggest that differences in ER / IP use for asthma are not solely due to issues such as overall access to care.
• ** Black (23.7%) and Hispanic (19.6%) children were much more likely than white (9.4%) children to say that asthma had a "very serious" effect on their overall health.

Slide 4

Research Objective

• Previous research.
  • Black and Hispanic children lower use of prescription drugs than white children.
    • All drugs.
    • Specific classes.
  • Mixed evidence on disparities for asthma drugs.
• Research Objective:
  • We examine pharmaceutical treatment of asthma in school-aged children to determine whether disparities in use may contribute to poorer asthma outcomes for minority children.
• Severity:
  • A central issue for this research is that differences in asthma severity may affect the need for care, appropriate levels of care and outcomes.

Notes:

• There is mixed evidence on disparities for asthma drugs. For example, a central issue is whether there are differences in the use of controllers or preventive asthma medications. Some previous research finds no racial-ethnic differences while other studies find that black and Hispanic children are less likely to use these preventive medications than white children.
• For example: Central Issue: do children use controllers (preventive meds)?
• **Chen: 1996-2003 MEPS: no racial-ethnic differences in controllers.
• **Lieu: Medicaid data: 2002: black and Latino children less use of controllers.
• **Ferris: NAMCS: 1989-90 minority children 50% less likely to use ICS controllers. By 1995-96, no difference resolved for black, but not for Hispanic children.
• **Much of the previous literature on racial-ethnic disparities in asthma medication use has used Medicaid data or community samples. Also, NAMCS-MEPS and other data sources show rapid change in asthma drug use over time. Contribution of this study: take a comprehensive look at asthma treatments using recent nationally representative person-level data.**

Slide 5

Literature

• National Heart Lung and Blood Institute; National Asthma Education and Prevention Program. Expert Panel Report 3. Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Institutes of Health; 2007.
• Chen AY, Escarce JJ. Family Structure and the Treatment of Childhood Asthma. Medical Care 2008; 46: 174-184.

Notes:

• There is a large previous literature on children and asthma and I will not attempt to summarize it.
• However, I do need to mention two sources in particular:
  • First, we made extensive use of the 3rd report on Guidelines for the Diagnosis and Management of Asthma which was put out in 2007 by the National Heart Lung and Blood Institute. We primarily used this as a resource for categorizing asthma drugs into classes.
  • Second, Chen and Escarce recently published Family Structure and the Treatment of Childhood Asthma. Like us, they used the MEPS data, and to the extent that our research overlaps, our findings are very similar.

Slide 6

Definition of Asthma

• Chronic inflammation of airways causes.
  • Recurring episodes of coughing, wheezing, etc.
  • Airflow obstruction.
• Failure to control underlying inflammation leads to.
  • More symptoms, exacerbations.
  • Excessive use of relief medications.
  • Long term effects on lung functioning and development.
• Causes of asthma.
  • Genetic.
  • Environmental exposures: airborne allergens, viral infections, tobacco smoke, pollution.

Notes:

• A potentially misunderstood aspect of asthma is that it is caused by chronic inflammation of the airways. This chronic inflammation causes the acute symptoms of coughing, wheezing and airflow obstruction that most people associate with asthma.
• Failure to control the underlying inflammation may result in more acute symptoms, excessive use of medications used to treat acute symptoms and long term effects on lung functioning and development.

Slide 7

Pharmaceutical Treatment and Monitoring of Asthma

• Office-based visits for asthma.
• Peak flow meters.
  • Tool for self-monitoring.
  • Important for patients who do not perceive symptoms until they are severe.
• Controllers.
  • Prevent symptoms, but do not provide prompt relief.
  • Must be taken daily.
• Relievers.
  • Provide prompt relief, but do not control symptoms.
  • Excessive reliever use → need to increase use of controllers.
• Oral steroids.
  • Used long term to treat severe persistent asthma/exacerbations.
  • Side effects can be severe.

Notes:

• The focus of our study is the monitoring and pharmaceutical treatment of asthma.
• Peak flow meters are used to measure a child's ability to push air out of their lungs. Monitoring with peak flow meters provides information that helps asthma patients make appropriate use of medications and avoid ER and inpatient hospital visits.
• Controller medications, such as inhaled corticosteroids, are medications that prevent symptoms by reducing the underlying airway inflammation. They do not provide prompt relief and must be taken daily.
• Reliever medications, such as inhaled albuterol, provide prompt relief, but do not control symptoms. Excessive use of relievers (more than 2X per week) may indicate a need to increase the use of controller medications.
• Finally, oral steroids are used long term to treat the most severe asthma and are used in short bursts to control exacerbations. Side effects of these medications can be severe.

Slide 8

Asthma Treatment Recommendations and Severity

The table shows:

• Classification of Asthma Severity.
  • Intermittent.
  • Persistent.
  • Mild to Moderate to Severe.
• Preferred Medication.
  • SABA PRN.
  • Low dose ICS.
  • Increase dose of ICS and/or add LABA or oral steroid.
• Quick Relief Medication.
  • SABA as needed for symptoms/intensity of treatment depends on the severity of symptoms.
• Recommended drug treatment depends on asthma severity.
• No controller use is recommended for children with intermittent asthma.
• Controllers are recommended for all children with persistent asthma with dosages increasing with the level of severity.
• Source: Guidelines for the Diagnosis and Management of Asthma. NAEPP, US DHHS, 2007.
• Note: SABA = short acting beta2-agonists (reliever).
  ICS = inhaled corticosteroids (controller).
  LABA = long acting beta2-agonists (controller).

Notes:

• As I noted earlier, a central challenge for this research is that recommended drug treatment depends on asthma severity.
• No controller use is recommended for children with intermittent asthma. Instead, asthma for these children is treated as necessary with an SABA which is a reliever medication.
• Controllers, esp. inhaled corticosteroids, are recommended for all children with persistent asthma, but children with more severe asthma may be given higher doses of an ICS and/or they may be prescribed other controller medications.

Slide 9

Data: Study Sample 1: Children with Treatment for Asthma

• Data: 2000-2005 Medical Expenditure Panel Survey (MEPS).
• Population studied:
  • School aged children (ages 5-17) with parent-reported treatment for asthma.
• Service use associated with asthma.
  • Drugs:
    • Link MEPS drug data by NDC to the Multum Lexicon.
    • Use generic names to categorize as controllers, relievers or oral corticosteroids.
    • Measures: any use, number of prescriptions.
• Other service use: outpatient/office-based and ER/inpatient hospital.
• Sample size:
  • N = 1,370 white non-Hispanic children.
  • N = 802 black non-Hispanic children.
  • N = 918 Hispanic children.

Notes:

• The primary data for our study come from the MEPS.
• For our first sample, we use the 2000-2005 MEPS and select school-aged children with any parent-reported treatment for asthma during the year. This gives us a sample 1370 white non-Hispanic children, 802 black non-Hispanic children and 918 Hispanic children.
• To get information on drug use, we linked the MEPS drug data to the Multum Lexicon to get standardized generic names for each drug, then we used these generic names to categorize drugs used for asthma as controllers, relievers or oral steroids. For each class of drugs we constructed 0/1 variables for any use and count variables for the number of prescriptions purchased. We also constructed 0/1 variables indicating whether each child had an office-based visit or an ER or inpatient visit for asthma.

Slide 10

Data: Study Sample 2: Children with 'Asthma Now'

• Data: MEPS 2003-2005
  • Beginning in 2003, MEPS respondents were asked the following questions about each person in their household.
• Has (PERSON) ever been told by a doctor or other health professional that (PERSON) has asthma?
• If yes:
  • Still have asthma?
  • Any attacks in the last 12 months?
• If yes to 'Asthma Now':
  • Peak flow meters: any use?
  • Controllers: any use? daily use?
  • Relievers: any use? > 3 canisters in the last 3 months?
• Sample size:
  • N = 740 white non-Hispanic children.
  • N = 505 black non-Hispanic children.
  • N = 507 Hispanic children.

Notes:

• Our second sample uses the 2003-2005 MEPS. Beginning in 2003, MEPS respondents were asked a series of direct questions regarding the experience of each household member with asthma.
• They were asked if each person had ever been told by a doctor that they had asthma. If yes, they were asked whether the person still had asthma or if they had an attack in the last 12 months. Persons answering yes to either of these questions were classified as having 'Asthma Now' and were asked a series of questions regarding their use of peak flow meters and controller and reliever medications.
• A nice feature of these questions is that they provide useful measures of the intensity of use. In particular: do you use controllers daily (as recommended for all persons with persistent asthma); do you make excessive use of reliever medications (e.g., more than 3 canisters in the last 3 months?).

Slide 11

Modeling Approach

• Estimate models with 0/1 race-ethnicity variables.
  • Logistic models for 0/1 outcomes (e.g., any controller use).
  • Negative binomials for count variables (e.g., # of controller scrips).
• Control variables.
  • Socio-demographic: age, sex, insurance status, family income, family structure, parental education, region, MSA, year.
  • Health: perceived health status, co-morbidities related to asthma, limitations.
  • County/Census Block: housing and community characteristics, pollution.
• Marginal effects.
  • Method of recycled predictions.
  • BRR estimates of standard errors.
  • Test differences between white children and other groups.

Notes:

• We use logistic regressions to model 0/1 outcomes such as any controller use. We use negative binomial regressions to model outcomes such as the number of controller prescriptions which are characterized by small, non-negative numbers. The key variables in these models are 0/1 race-ethnicity variables.
• In our main models we examine differences in asthma treatment across racial and ethnic groups controlling for socio-demographic characteristics and health status and co-morbidities that may be associated with asthma severity to. In extended models we also control for county and census block characteristics. There are few qualitative differences between these sets of models, so I will just present results for the main models.
• To estimate the marginal effects of race-ethnicity, we use results from our models to calculate what the mean outcomes for our sample would be if all children were white, if all children were black and if all children were Hispanic. Marginal effects are the differences between these mean outcomes. We use BRR (balanced repeated replication) to estimate standard errors of marginal effects and we test for differences between white children and other groups.

Slide 12

Office Visits for Asthma, Use of Peak Flow Meters by Race-Ethnicity

The bar graphs show:

• Any Office-Based/Outpatient Visits for Asthma?
  • Children with Asthma Treatment.
    • White: 39.3 percent "yes."
    • Black: 39.7 percent "yes."
    • Hispanic: 46.4* percent "yes."
• Any Use of Peak Flow Meters?
  • Children with Asthma Now.
    • White: 28.0 percent "yes."
    • Black: 33.6 percent "yes."
    • Hispanic: 34.9* percent "yes."
• Note: *P <.05 for difference with white children.

Notes:

• The next several slides show our results:
• This first slide focuses on two measures of preventive care.
• We find that among children with treatment for asthma, Hispanic children are more likely than white children to have an office-based visit for asthma. That result is shown in the chart on the left. The chart on the right shows that among children identified as currently having asthma, Hispanic children were also more likely than white children to use peak flow meters.

Slide 13

Use of Controllers by Race-Ethnicity: Children with Asthma Now

The bar graphs show:

• Any Controller Use?
  • White: 47.1 percent "yes."
  • Black: 46.9 percent "yes."
  • Hispanic: 42.4 percent "yes."
• Daily Controller Use?
  • White: 29.9 percent "yes."
  • Black: 28.7 percent "yes."
  • Hispanic: 25.4 percent "yes."
• Note: No statistically significant differences in use.

Notes:

• Next, we examine use of controller medications among children identified as having asthma now.
• In the chart on the left we see that although point estimates are a little higher for white children, there are no statistically significant differences across racial-ethnic groups in the percentage of children with any controller use. The chart on the right shows that there were also no statistically significant differences in the percentage of children who used controllers daily—as is recommended for all children with persistent asthma.

Slide 14

Use of Controllers by Race-Ethnicity: Children with Treatment for Asthma

The bar graphs show:

• Any Controller Use?
  • White: 41.5 percent "yes."
  • Black: 34.4* percent "yes."
  • Hispanic: 35.2+ percent "yes."
• Total Prescriptions.
  • White: 2.7.
  • Black: 2.6.
  • Hispanic: 2.3.
• Note: *P <.05; +P <.10 for difference with white children.

Notes:

• Next, we examine controller use among children with any treatment for asthma during the year.
• Among this group, we find that white children are more likely than black and Hispanic children to use controllers. For example, 41.5% of white children used at least one controller during the year compared to 34.4% of black children.
• We don't find differences, however, in the total number of controller prescriptions purchased. This results, in part, from the fact that when minority children use controllers, they tend to use them more intensively.

Slide 15

Use of Controllers by Race-Ethnicity: Children with Persistent Treatment¹

The bar graphs show:

• Any Controller Use?
  • White: 83.3 percent "yes."
  • Black: 68.2* percent "yes."
  • Hispanic: 72.6* percent "yes."
• Total Prescriptions.
  • White: 6.2.
  • Black: 5.4.
  • Hispanic: 5.1.
• Note: *P <.05 for difference with white children.
  ¹ Sample restricted to children with 3 or more prescriptions for asthma.

Notes:

• We extend our examination of controllers by limiting the sample to children with persistent treatment for asthma. We categorize a child as having persistent treatment if they purchased three or more prescriptions to treat asthma.
• Among this group we find, again, that white children are more likely than others to use at least one controller during the year and we find that the magnitude of these differences is larger than when we examine treatment for all children with treatment for asthma.

Slide 16

Use of Reliever Medications by Race-Ethnicity

The bar graphs show:

• Total Prescriptions.
  • Children with Treatment for Asthma.
    • White: 1.3.
    • Black: 1.9.*
    • Hispanic: 1.7.*
• Use > 3 Canisters in 3 Months?
  • Children with Asthma Now.
    • White: 3.9 percent.
    • Black: 8.0* percent.
    • Hispanic: 11.0* percent.

• Note: *P <.05 for difference with white children.

Notes:

• Moving on to other classes of medications:
• The chart on the left shows that among children with treatment for asthma black and Hispanic children both had more prescriptions for relievers, on average, than white children.
• This result is somewhat ambiguous: on the one hand it may be indicative of good access to care for minority children. On the other hand, it may be indicative of higher rates of poorly controlled asthma.
• The result in the chart on the right, however, unambiguously points to excessive use of relievers by minority children. In particular, these results show that among children identified as having asthma now, black and Hispanic children are both more than twice as likely as white children to have used >3 canisters of relievers in the previous 3 months.

Slide 17

Use of Oral Steroids by Race-Ethnicity

The bar graphs show:

• Any Oral Steroid Use.
  • Children with Treatment for Asthma.
    • White: 8.8.
    • Black: 11.3.
    • Hispanic: 10.7.
• Total Prescriptions.
  • Children with Treatment for Asthma.
    • White: 0.14.
    • Black: 0.25.*
    • Hispanic: 0.18.

• Note: *P <.05 for difference with white children.

Slide 18

Controller Ratio by Race-Ethnicity: Children w Persistent Treatment²

The bar graph shows:

• Controller ratio = (# of controller prescriptions) / (total # of prescriptions for asthma)
• Quality of care measure.
• Indicator variable = 1 if controller ratio ? 0.5.
• Any Controller Ratio.
• White: 70.0 percent >0.5.
• Black: 48.9* percent >0.5.
• Hispanic: 49.1* percent >0.5.

• Note: *P <.05 for difference with white children.
  ²Sample restricted to children with 3 or more prescriptions for asthma.

Slide 19

Summary of Results

• Preventive Measures.
  • Hispanic children were more likely than white children to use peak flow meters and to have an OB/OP visit for asthma.
  • There were no statistically significant differences in the use of controller medications among children with 'asthma now'.
  • Among children with treatment for asthma, white children were more likely to use controller medications.
• Drugs for Acute Symptoms.
  • Black and Hispanic children (esp. Hisp) were more likely than white children to use relievers, and to use them intensively.
  • Black children also had more prescriptions for oral steroids than white children.

Slide 20

Conclusions

• Conclusions: Black and Hispanic children have higher rates of poorly controlled asthma than white children resulting in more missed school and more ER / IP visits. This may indicate unobserved differences in disease severity that may explain the greater use of relievers by minority children and that complicates evaluation of differences/similarities across groups in controller use.
• Implications: Successful treatment of asthma has the potential to profoundly increase quality of life for millions of U.S. children. As minority children are particularly hard-hit by asthma, racial and ethnic disparities in treatment are important to understand. Further research that includes specific measures of asthma severity are necessary to shed light on the nature of disparities in asthma treatment and outcomes.

Notes:

• Chen and Escarce find that children in single-parent households had fewer office visits, fewer reliever scrips and fewer controller scrips. Increasing numbers of siblings has a similar effect. They recognize differences in severity as a threat to validity (e.g., if children in single parent families have less severe asthma the observed patterns of care may be appropriate), but discount it because: 1.) in MEPS analysis they include health status and other controls that should be associated with severity 2.) in the NSCH data they find that children in single parent households have worse reported symptoms 3.) they know of no clinical reason why intrinsic severity should be associated with asthma after controlling for demographic factors.
• We find evidence of less controller use for minority children, but other differences go in the opposite direction. I am still worried about differences in severity in the case of racial-ethnic disparities.
• **For chronic conditions, drug use is one of the best indicators of severity that you can find.**

Current as of January 2009

---

Internet Citation:

Racial-Ethnic Disparities in Childhood Asthma Treatment and Outcomes: The Role of Disease Severity. Slide Presentation from the AHRQ 2008 Annual Conference (Text Version). January 2009. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/about/annualmtg08/090808slides/Miller2.htm

---