Using real-life patient records to help identify predictors of future exacerbation risk (Asthma risk predictors)

Study type

• Study topic:
  Disease /health condition
• Study type:
  Non-interventional study
• Scope of the study:
  Disease epidemiology
• Data collection methods:
  Secondary use of data

Non-interventional study

• Non-interventional study design:
  Cohort

  Cross-sectional

Study drug and medical condition

• Medical condition to be studied:
  Asthma

Population studied

• Short description of the study population:
  Patients aged 12–80 years who have physician diagnosed asthma and ≥3 years of continuous medical records.
• Age groups:
  Adolescents (12 to < 18 years)

  Adults (18 to < 46 years)

  Adults (46 to < 65 years)

  Adults (65 to < 75 years)

  Adults (75 to < 85 years)
• Special population of interest:
  Other
• Special population of interest, other:
  Asthma patients
• Estimated number of subjects:
  50000

Study design details

• Main study objective:
  To identify patient characteristics recorded within routine primary care datasets that are associated with increased risk of frequent asthma exacerbations with a view to building a risk assessment model to "Score" patients in terms of their future exacerbation risk.
• Outcomes:
  Co-primary outcomes:• Moderate-to-severe exacerbations: based on the ATS/ERS taskforce definition, any of:(i) Asthma-related: a. Hospitalisations (inpatient admissions) OR b. A&E attendance OR(ii) Use of acute oral steroids• Clinical exacerbations: As above, but including asthma-related out-patient-department attendance and antibiotics for lower respiratory tract infections. Disaggregate components of exacerbation definitions:• Oral steroid prescriptions• Hospitalisations for asthma or lower respiratory conditions• A&E attendance for asthma or lower respiratory conditions• Antibiotic prescriptions for lower respiratory tract infections (LRTIs)
• Data analysis plan:
  Phase 1. Autocorrelation plots will be examined to assess seasonality and timedependent relationships in the rate of exacerbations as described in the methods section.Phase 2 & 3. Univariate (phase 2) and multivariate (phase 3) associations will be estimated using:(i) Negative binomial regression models: will be used to determine predictors offuture risk in terms of severe exacerbation rates over subsequent 1 & 2yrs.(ii) Ordinal logistic regression models: will be used when annual exacerbationsare categorised 0, 1 and ≥2.(iii) Logistic regression models: will be used when severe exacerbations isdefined as a binary outcome. Here, population attributable risks will beestimated to provide an estimation of the proportion of “frequentexacerbation” trait that is explained by each statistically significant factor inthe multivariate models.Phase 4. Hierarchical multi-level modelling will be used to estimate associationswith exacerbation rates.