Patterns for EHR/Data
EHR Integration (aka "Case"): All and only the data pertinent to X.
Is available data relevant? (can I use this or do I need to order a new test?)
How is missing data handled?
EHR to Situational data (aka "Case Features"): e.g. (Any) Diagnosis of Diabetes -> Diabetes Present yes/no
Binding of Predictive Models (aka "Model Features"): e.g. Mapping quantitative to qualitative features (e.g. true -> 1.0, false -> 0.0)
Strong typing: Setting type for allowable values and assigning types to all names
normalization
Use of FHIR questionnaire
Patterns for DMN Models
DMN Functionality
Scoring Model: e.g. scorecard
...