Slide 13. Methods: Meta-Analysis

• Observations: Each study outcome (child/adult, hospital/etc.) = 1
• Dependent variable: Percent change in measure
• Explanatory variables/indicators:
  • Type of measure (e.g., hospital, ED, productivity)
  • Age group (children, adults, both)
  • Whether most patients in the study were Medicaid/SCHIP
  • Study patients' asthma severity (persistent, intermittent, or all)
  • Study design used (randomized, all controlled, uncontrolled)
  • If intervention primarily focused on physicians (not patients)
  • Whether outcome measured after 2 years
  • Sample size and sample size squared
• Regressions: Weighted linear, clustered SE
• Impact: Percent change predicted from coefficients

Contents