Racial-Ethnic Differences in Childhood Asthma Treatment (Text Version)

Slide presentation from the AHRQ 2011 conference.

On September 20, 2011, Eric Sarpong made this presentation at the 2011 Annual Conference. Select to access the PowerPoint® presentation (1.3 MB). Plugin Software Help.


Slide 1

 Racial-Ethnic Differences in Childhood Asthma Treatment

Racial-Ethnic Differences in Childhood Asthma Treatment

Eric M. Sarpong and G. Edward Miller

AHRQ Conference
September 20, 2011

Slide 2

 Introduction

Introduction

  • Asthma—chronic, complex and costly health condition:
    • Estimated costs of asthma in the U.S.—$19.7 billion (NHLBI, 2007).
  • Why childhood asthma medication use?
    • Recent significant increases in:
      • Treated prevalence.
      • Health care and prescribed asthma drug expenditures (Miller & Sarpong).
    • Recent changes in asthma care:
      • Treatment guidelines (NAEPP-EPR3, 2007)—Inhaled corticosteroid (ICS); preferred first-line therapy for persistent asthma.
    • Changes in use and availability of new pharmacotherapies:
      • Increased use of controller medications.
      • Decreased use of mono-therapy with relievers.
  • Differences in asthma treatment persists.

Slide 3

 Pharmaceutical Treatment of Asthma

Pharmaceutical Treatment of Asthma

  • Recommended treatment depends on asthma severity (NAEPP-EPR3, 2007):
    • Controllers (e.g., inhaled corticosteroid):
      • Used in managing asthma symptoms, by minimizing inflammation and reducing the risk of serious exacerbations.
      • Recommended for all children with persistent asthma.
    • Relievers (e.g., inhaled short acting beta agonists):
      • Used in managing moderate or severe asthma attacks by promptly relaxing airway muscles.
      • Recommended for all children with intermittent asthma.

Slide 4

Treated Prevalence of Childhood Asthma by Race-Ethnicity, 2005-2008  

Treated Prevalence of Childhood Asthma by Race-Ethnicity, 2005-2008

Image: A bar graph presents the following data:

  • All Children: 6.5.
  • Non-Hispanic (NH) White: 6.2.
  • NH Black: 8.7.
  • Hispanic: 5.8.

Source: Medical Expenditure Panel Survey (MEPS), 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites significant at p <.05.

Slide 5

 Use of Controllers Among Children With Treatment for Asthma, 2005-2008

Use of Controllers Among Children With Treatment for Asthma, 2005-2008

Image: A bar graph presents the following data:

  • All Children: 59.5.
  • NH White: 98.4.
  • NH Black: 44.2.
  • Hispanic: 49.5.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites significant at p <.05.

Slide 6

 Use of Relievers Only , Among Children with Treatment for Asthma, 2005-2008

Use of Relievers Only, Among Children with Treatment for Asthma, 2005-2008

Image: A bar graph presents the following data:

  • All Children: 30.5.
  • NH White: 22.9.
  • NH Black: 42.8.
  • Hispanic: 38.7.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites significant at p <.05.

Slide 7

 Differences in Treated Prevalence and Use of Asthma Medications

Differences in Treated Prevalence and Use of Asthma Medications

  • NH Black children—more likely than NH White and Hispanic children to be treated for asthma.
  • NH Black and Hispanic children—less likely than NH White children to use controllers.
  • NH Black and Hispanic children—more likely than NH White children to use relievers only.

Slide 8

 Previous Research

Previous Research

  • Large body of literature with mixed evidence on differences in children's use of asthma medication.
  • Some studies find no differences by race-ethnicity, others do find a difference.
  • Studies differ on a number of dimensions:
    • Time period.
    • Population (e.g., Medicaid, private claims, nationally representative).
    • Degree to which they control for differences in underlying characteristics across groups.

Slide 9

 Research Objective and Contribution

Research Objective and Contribution

  • Research Objective:
    • Examine differential use of asthma medication by race-ethnicity.
    • Examine extent to which differences in mean predisposing, enabling and need characteristics explain differences in use.
  • New Contribution:
    • Previous literature limited—Medicaid data, administrative data or community samples, key variables unavailable.
    • Comprehensive look at differences using nationally representative data (MEPS).
    • Provide descriptive information on reasons for differences and possible approaches to addressing these differences.

Slide 10

 Analytic Approach

Analytic Approach

  • Describe differences in controllers and relievers only use by race-ethnicity.
  • Estimate pooled regression models with binary outcomes (i.e., controllers and relievers only):
    • Explanatory variables:
      • Predisposing (e.g., socio-demographics, geographic).
      • Enabling (e.g., health insurance, family income/structure and parental education/employment status).
      • Need (e.g., health status, co-occurring conditions).
  • Use Oaxaca-Blinder method—decompose differences into:
    • Explained—differences due to mean differences in explanatory variables.
    • Unexplained—differences due to differences in estimated coefficients.

Slide 11

 Data

Data

  • Data:
    • 2005-2008 Medical Expenditure Panel Survey (MEPS).
  • Population studied:
    • Insured children (ages 0-17) with reported treatment for asthma:
      • Treatment = health service use associated with asthma.
    • Sample size:
      • N = 813 NH White children.
      • N = 608 NH Black children.
      • N = 644 Hispanic children.
  • Drugs:
    • Link MEPS drug data by NDC to the Multum Lexicon.
    • Use generic names to categorize drugs as controllers or relievers.
    • Measures: any use of controllers and relievers only.

Slide 12

 Descriptive Results-Differences in Mean Characteristics

Descriptive Results—Differences in Mean Characteristics

  • NH Black children more likely than NH White children to:
    • Be covered by public insurance, live in a metropolitan statistical area (MSA) and in the south (predisposing).
    • Live in families with low levels of family income, low parental education, unmarried parent and unemployed parent (enabling).
    • Be in fair/poor physical health (need).
  • Hispanic children more likely than NH White children to:
    • Be covered by public insurance, live in an MSA, in the west, have non-native parents and parents with risky attitudes towards health (predisposing).
    • Live in families with low levels of family income, low parental education, unmarried parent, unemployed parent and live in a larger family (enabling).
    • Be in fair/poor physical health (need).

Slide 13

 Multivariate Results-Predictors of Controller and Reliever Only Use

Multivariate Results—Predictors of Controller and Reliever Only Use

  • Positive effects on controller use:
    • Age 5-11, native parents, married parent, fair/poor physical health, and having treatment for allergies.
  • Negative effects on controller use:
    • Female, MSA, western region, low family income, parental education, and large family.
  • Positive effects on reliever only use:
    • Female, MSA, western region, native parents, low family income, and large family.
  • Negative effects on reliever only use:
    • Age 5-11, married parent, fair/poor physical health, and having treatment for allergies.

Results are based on coefficient estimates from pooled linear probability models.

Slide 14

 Multivariate Results-Predictors of Controller and Reliever Only Use

Oaxaca-Blinder Decomposition—NH Whites vs. NH Blacks, Controllers

Image: A bar graph presents the following data:

  • Enabling, 8.6 (Explained).
  • Need, 2.2 (Explained).
  • Unexplained, 12.8.

Total Difference in Controller Use = 24.5%.

Important variables:

Enabling

  • Low income.
  • ≤High school.
  • Family of >4.
  • Unmarried.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites, significant at p <.10.

Slide 15

 Oaxaca-Blinder Decomposition-NH Whites vs. NH Blacks, Controllers

Oaxaca-Blinder Decomposition Results—NH Whites vs. NH Blacks, Relievers Only

Image: A bar graph presents the following data:

  • Enabling, 5.7 (Explained).
  • Need, 1.2 (Explained).
  • Unexplained, 12.2.

Total Difference in Reliever Only Use = 19.8%

Important variables:

Enabling

  • Low income.
  • ≤High school.
  • Family of >4.
  • Unmarried.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites, significant at p <.10.

Slide 16

 Oaxaca-Blinder Decomposition Results-NH Whites vs. NH Blacks, Relievers Only

Oaxaca-Blinder Decomposition Results—NH Whites vs. Hispanics, Controllers

Image: A bar graph presents the following data:

  • Predisposing, 6.6 (Explained).
  • Enabling, 6.3 (Explained).
  • Need, 1.7 (Explained).
  • Unexplained, 4.4.

Total Difference in Controller Use = 19.1%

Important variables:

Predisposing

  • MSA.
  • West.
  • Non-native parents.

Enabling

  • Low income.
  • ≤High school.
  • Family of >4.
  • Unmarried.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites, significant at p <.10.

Slide 17

 Oaxaca-Blinder Decomposition Results-NH Whites vs. Hispanics, Relievers only

Oaxaca-Blinder Decomposition Results—NH Whites vs. Hispanics, Relievers only

Image: A bar graph presents the following data:

  • Enabling, 4.4 (Explained).
  • Need, 1.2 (Explained).
  • Unexplained, 7.8.

Total Difference in Reliever Only Use = 15.8%

Important variables:

Enabling

  • Low income.
  • ≤High school.
  • Family of >4.
  • Unmarried.

Source: MEPS, 2005-2008. Estimates for insured children with reported treatment for asthma. Differences from NH Whites, significant at p <.10.

Slide 18

Interpretation of Oaxaca-Blinder Decomposition Results  

Interpretation of Oaxaca-Blinder Decomposition Results

  • Several characteristics in the domains of the behavioral model were associated with:
    • Controller use.
    • Reliever only use.
  • Our model:
    • Explained most differences in:
      • Controller use for NH Blacks and Hispanics.
      • Reliever only use for Hispanics.
    • Differences in reliever only use for NH Blacks, largely unexplained.
  • Unobservable factors—differences in responses to characteristics, may be important.

Slide 19

 Limitations and Future Research

Limitations and Future Research

  • Limitations:
    • No measure of asthma severity:
      • Results may change if severity differs across groups.
    • Non-causal descriptive model.
  • Future research:
    • Depart from linear probability models.
    • Use non-parametric approach:
      • Raking (Pylypchuk and Selden, 2008, JHE).
    • Follow previous approaches:
      • Kirby et al, 2010 MCRR and Hudson et al, 2007.

Slide 20

 Conclusions

Conclusions

  • Non-Hispanic whites: low asthma treated prevalence and reliever only use and, higher controller use.
  • Enabling factors explained some, not all, of the differences in controller and reliever only use for NH Blacks and Hispanics.
  • Predisposing factors explained some of the differences in controller use for Hispanics.
  • Some unobservable characteristics may have also played a role.
  • Results are consistent with studies in other therapeutic classes of drugs and disease areas.

Slide 21

 References

References

  • Miller G.E. and Sarpong E.M. Trends in the Pharmaceutical Treatment of Children's Asthma, 1997 to 2008. Research Findings No. 31. September 2011. Agency for Healthcare Research and Quality, Rockville, MD. http://meps.ahrq.gov/mepsweb/data_files/publications/.
  • Kirby JB, Hudson J, Miller GE. (2010). Explaining Racial and Ethnic Differences in Antidepressant Use Among Adolescents. Med Care Res Rev 67:342-363.
  • Crocker D, Brown C, Moolenaar R, Moorman J, Bailey C, Mannino D, Holguin F. (2009). Racial and ethnic disparities in asthma medication usage and health-care utilization: data from the National Asthma Survey. Chest 136(4):1063-71. Epub 2009 Jun 30.
  • Pylypchuk, Y. and T. M. Selden. (2008). A discrete choice decomposition analysis of racial and ethnic differences in children's health insurance coverage. Journal of Health Economics 27:1109-1128.
  • Chen AY, Escarce JJ. (2008). Family Structure and the Treatment of Childhood Asthma. Medical Care 46:174-184.
  • Hudson, J. L., Miller, G. E., & Kirby, J. B. (2007). Explaining racial and ethnic differences in children's use of stimulant medications. Medical Care 45, 1068-1075.
  • National Asthma Education and Prevention Program. Expert Panel Report 3: Guidelines for the Diagnosis and Management of Asthma: Full Report 2007. Bethesda, MD: National Institutes of Health, US Dept of Health and Human Services, National Heart, Lung, and Blood Institute; 2007.
  • Shields A, Comstock C, Weiss KB. (2004). Variations in asthma care by race/ethnicity among children enrolled in a state Medicaid program. Pediatrics 113:496-504.
  • Finkelstein JA, Lozano P, Farber HJ, et al. (2002). Underuse of controller medications among Medicaid-insured children with asthma. Arch Pediatr Adolesc Med 56(6):562-7.
  • Lieu TA, Lozano P, Finkelstein J, et al. (2002). Racial/ethnic variation in asthma status and management practices among children in managed Medicaid. Pediatrics 109:857-865.
  • Blinder, A. (1973). Wage discrimination: Reduced form and structural estimates. Journal of Human Resources 8, 436-455.
  • Oaxaca, R. L. (1973). Male-female wage differentials in urban labor markets. International Economic Review 14, 693-709.
Page last reviewed March 2012
Internet Citation: Racial-Ethnic Differences in Childhood Asthma Treatment (Text Version). March 2012. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/news/events/conference/2011/sarpong/index.html