DMN Functionality
Scoring Model: e.g. scorecard
Scoring: assigning a compounded value from various contributing factors
Clinical Interpretation of Observations (including scores): e.g. score to risk
decision table
Decision table normalization: Akin to database normalisation, structuring in order to reduce redundancy and to ensure independence of each input and rule
rank/sorting/MIN-MAX
equation or function
General Predictive Model / AI 'as a function': e.g. using BKM / decision services to mediate
Iterating: applying the same logic to a list of items
Categorizing: assigning a label (category) based on either various contributing factors or a range of a
continuous value
DMN Goals
State/Assessment: e.g. patient has/is/on X (Diagnosis)
Gap Assessment: e.g. current vs ideal state
DMN Best Practices
Decomposition: decompose a decision into smaller sub decisions
Service oriented decomposition: decompose a decision by calling externally defined decisions
Separation of concerns: Identify, encapsulate, and manipulate as independently as possible those parts of the model that are relevant to a particular concern
Point of View: create various DRDs of the same DMN model to highlight particular perspectives of a more complex DRD
Selection: Selecting an attribute of a complex data object
BPMN Models
Key clinical processes:
(1) diagnosis of a problem
(2) management of problem
(3) monitoring, including response to intervention
Common BPMN Sequences
diagnose, then establish risk status (based on severity, based on extent)
therapy: indications, contraindications, alternatives, benefit, risks
Recommendation (Indication -> Recommendation): e.g. gap present -> suggested intervention
monitoring: improved/same/worse (gap analysis under DMN), adverse effects, disease complications
BPMN Methodology
Discriminating (routing): selecting an option out of a collection of possibilities
Conditional branching
Filtering: Removing elements of a certain form from a collection of items
Observation-Assessment-Finding (OAF) Pattern (R. Lario)
Meant to communicate that the task of making an Observation(s) was to be performed by the clinician and a value recorded. Further, a clinical rule is applied to the observation and respective values to be assessed asserting a new clinical Finding(s). These findings, in tandem with the implied observation(s), can be aggregated into a compound observation. Where this compound observation is then Assessed using a rule or clinical algorithm creating yet another new finding. Based upon this new finding, the CPG Branched accordingly progressing to the next procedure of either making a new observation or performing an Intervention. This pattern continues to repeat: Observations that are Assessed resulting in Findings followed by a respective Branch inflow into either a subsequent Intervention or Observation (Figure 2. Observation Assessment Finding (OAF) Pattern).
...
Ways of representing data:
Stand alone value with units
Relative to a reference value or range (above normal; low-normal-high)
Z-score (incorporating mean and SD)
Multiple of mean (MOM)
Multiple of upper limit of normal (ULN)
Classification of Data Items
(1) simple vs complex (structure)
(2) constant vs variable/changing
(3) predictable vs unpredictable
(4) situational
Query: Does the person have a history of heart disease?
Simple response: Yes or No.
Complex data structure: date of occurrence, anatomic distribution, severity
Date of birth is constant.
Age is constantly changing at a rate of 1 per year.
Blood pressure changes but not in a constant manner.
Predictability: if a patient is normal, then the measure should be within a defined range.Fundamentals are basic, atomic patterns that solve very fine-grained problems, common across a variety of healthcare models. Fundamentals can be used independently, or combined into more complex patterns.
Data Items - Patterns on
EHR Data - Integration of
DMN Functionality
DMN Goals and Best Practices
BPMN Models and Methodology