Predictive Analytics

Steps to complete:

  1. Evaluate existing instruments used to measure outcomes
  2. Modify or amend existing instruments (if required)
  3. Combine surveys if redundant
  4. Execute measurement of outcome
  5. Power analysis conducted to evaluate number of patients needed for initial study (alpha = .05, power = .95, effect size = .15)
  6. Match research and control group sample sizes
  7. Initial analysis of outcome as dependent variable, all others as independent using correlations and hierarchical regression
  8. Presentation of results focusing on coefficients for the purpose of prioritizing identification of causal variables
  9. Move to logistic regression for probability analysis
  10. Final stage, move to odds ratio for management of risk based on forecasting, using data trajectories

Quote Required