The current outbreak of coronavirus disease (COVID-19) was first reported from Wuhan, China, on 31 December 2019. The World Health Organization (WHO) is working closely with global experts, governments and partners to rapidly expand scientific knowledge on this new virus, to track its spread and virulence, and to develop measures to contain the outbreak. In this session we will demonstrate a set of clinical knowledge artifacts (KNART) that are meant to help clinicians to make decisions regarding the need for quarantine and to guide the care flow of individuals that are symptomatic. These COVID-19 clinical knowledge artifacts can act as a clinical guideline that is both human readable and machine automatable. It includes care flow models captured in BPMN, clinical decision support models captured in DMN, along with Predictive Models captured in PMML and Clinical Measures captured in CQL.

1. Novel Coronavirus
(COVID-19)
Integrating Predictive Models (PMML), Clinical Quality Measures
(CQL) and Knowledge Models with BPM+ Models
Dr. John Svirbely, MD
Chief Medical Informatics Officer (CMIO),
jsvirbely@Trisotech.com
Denis Gagne
Chief Executive Officer (CEO),
dgagne@Trisotech.com

2. Coronaviruses (CoV)
• Are zoonotic, meaning they are transmitted between animals and people.
• Is a large family of viruses that cause illness ranging from the common
cold to more severe diseases such as:
• Middle East Respiratory Syndrome (MERS-CoV)
• Severe Acute Respiratory Syndrome (SARS-CoV)
• A novel coronavirus (nCoV) is a new strain that has not been previously
identified in humans.
https://www.who.int/health-topics/coronavirus

3. https://covid19.who.int/

4. Introduction of PMML, CQL and
Knowledge Models
We have three goals today:
• To introduce these seamlessly integrated within the BPM+
stack
• To provide separation of concerns best practices
• To showcase the art of the possible on these topics using
COVID examples

5. Desired characteristics of
our knowledge artefacts
(In a crisis context, these characteristics are must haves)
• To provide unambiguous:
• Knowledge Models
• Decision Models
• Course of Action Models
• That are both human consumable and machine automatable
• That are rapidly modifiable and deployable
Some more COVID-19 examples available here:
https://www.trisotech.com/covid-19

6. API
Container
Automation Marchitecture
CMMN
Case
BPMN
Workflow
DMN
Decision
Predictive Models (PMML)
Clinical Measures (CQL)
FHIR Data
CDS Hooks
seamlessly integrated within the BPM+ stack

7. Separation of Concerns
Perspective Interrogative Purpose Type of Model
Informative What should I do? Decision Support DMN model
Prescriptive What will be done? Decision Automation BPMN and DMN model
Predictive What will happen? Prediction PMML model
Diagnostic How did we do? Quality Measure CQL model
Knowledge What does it mean? Disambiguation Knowledge model

8. Decision Support
DMN 👤
User acts based on the
recommendation from
the DMN decision
DMN: Decision Model and Notation
A standard from the Object Management Group
https://omg.org

9. COVID Decision Support
COVID19 Adult Mortality Model Zhou
Zhou et al developed a simple model for predicting
in-patient mortality of a patient with COVID-19. This
can help to identify a patient who may benefit from
more aggressive management. The authors are from
multiple institutions in China.
References
Zhou F, Yu T, et al. Clinical course and risk factors for mortality of adult inpatients with COVID-19
in Wuhan, China: a retrospective cohort study. Lancet. 2020. doi: 10.1016/S0140-
6736(20)30566-3.

10. Decision Automation
DMN: Decision Model and Notation
A standard from the Object Management Group
DMN
System acts based on
the DMN decision
https://omg.org
BPMN: Business Process Model and Notation
A standard from the Object Management Group
DMN provides the decision
BPMN acts upon the decision

11. COVID Decision Automation
Asymptomatic Suggested Follow Up
A patient who is asymptomatic after a possible exposure
to the novel coronavirus should be monitored for signs
and symptoms during the possible incubation period (up
to 14 days after the exposure).
References
World Health Organization. Clinical management of severe acute respiratory infection
when Novel coronavirus (2019-nCoV) is suspected. Interim guidance. 2020.
WHO/bCoV/Clinical/2020.2

12. Prediction
PMML: Predictive Model Markup Language
A standard from the Data Mining Group
System acts based on
the PMML Prediction
PMML
https://dmg.org

13. COVID Prediction
Risk of transmission on a plane
Making predictions requires historical or prior data.
We used SARS data for reference.
References
Olsen SJ, Chang HL, Cheung TY, et al. Transmission of severe acute respiratory syndrome on
aircraft. N Engl J Med 2003; 349:2416–22
https://www.trisotech.com/blog/modeling-virus-transmission-on-an-airplane

14. Quality Measure
CQL: Clinical Quality Language
A standard from HL7
System acts based on
the result of the CQL measure
CQ
L
https://HL7.org

15. COVID Measure
COVID19 Patient Risk for Fatal Disease simplified
The risk for mortality associated with a COVID-19
infection increases with age and comorbid conditions.
References
Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150
patients from WRuan Q, Yang K, et al. uhan, China. Intensive Care Medicine. doi:
10.1007/s00134-020-05991-x

16. Disambiguation
A meaning-centric and unambiguous communications across the organization
Expression of meaning via:
• Business vocabularies containing terms with their exact meanings
• Concept maps for expressing knowledge
• Deontic (business) rules specifying what is obligated and permitted
• Logical information structures to relate to data in HIT systems
A knowledge model results in a semantically rich common business language that is
reusable in narrative communications, business process, case and decision models, as
well as in software applications specifications.

17. COVID Knowledge Model
References
Surviving Sepsis Campaign: Guidelines on the Management of Critically Ill Adults with
Coronavirus Disease 2019 (COVID-19) ncbi.nlm.nih.gov/pubmed/32222812

18. Conclusion
Today we presented:
• PMML and CQL seamlessly integrated within BPM+ stack
• Disambiguation knowledge models that permeates the BPM+ stack
Questions?