- 1 ACC/AHA Guidelines and Indications for Coronary Artery Bypass Graft Surgery
- 2 Data
- 2.1 Demographics
- 2.2 Clinical status
- 2.3 Distribution and severity of coronary artery disease (CAD)
- 2.4 Left ventricular dysfunction (grade 0 to 4) (more severe)
- 2.5 Aggressiveness of arteriosclerotic process
- 2.6 Surgical Factors
- 3 General Indications for Coronary Artery Bypass Operation (Section VI)
- 4 Patient-specific Guidelines and Indications (Section X)
- 5 Methodology and Equations (Appendix F)
- 6 References
- 7 Glossary
ACC/AHA Guidelines and Indications for Coronary Artery Bypass Graft Surgery
The original report from the ACC/AHA Task Force on Assessment of Diagnostic and Therapeutic Procedures can be found here ACC/AHA Guidelines and Indications for Coronary Artery Bypass Graft Surgery
The ACC and AHA have designated a Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures. That Task Force appointed a Subcommitee to develop guidelines and indications for the coronary artery bypass operation. This is the report of the Subcommittee. It contains a distillation of current information, focused on present indications and practices.
The report includes general information regarding indications for the coronary artery bypass graft (CABG) derived from comparision with initial medical treatment. The paper strongly emphasizes that the general guidelines cannot take into account even the majority of the variables that are involved in most recommendations to patients.
An additional method for calculating time-related outcomes from patient-specific data is suggested as more accurate. Log-linear risk factor equations are used to generate outcome graphs.
This usecase attempts to demonstrate the a 'standard' approach to decision support using the general guidelines and investigate the possiblity of using functions to accomodate the risk factor calculations within the framework of a standard Clinical Pathway Protocol.
The risk factor calculations can be thought of as black-boxes which take (as input) risk factors driven by patient-data which outputs recommendations with accompanying probability for a particular point in time from now.
It's interesting to note that such an approach is an effective way to accomodate uncertainty in logic programming reasoners (such as Euler,Pychinko, and CWM) for semantic web technologies - a well known limitation of model-theoretic systems - by the use of functions which represent models generated from targetted, domain-guided statistical analysis.
The risk factors that drive the general guidelines as well as the outcome calculations are based on specific diseases and symptoms as well as other factors (such as age, etc..).
The symptoms and diseases amongst the contributing factors can generally be modelled as instances of rim:!Observation.
An act that is intended to result in new information about a subject. The main difference between observations and other acts is that it has a value attribute that is used to state the result of the assessment action described in Act.code.
Most of the symptoms/diseases are associated with the finding itself, an anatomic location, and a severity which can correpond to the following HL7 RIM terms:
- Age at operation
- Body size
Age at operation
HL7 RIM has a rim:birthTime attribute on the rim:LivingSubject Entity (of which rim:Person is a subclass). The range of this attribute is a point in time:
A quantity specifying a point on the axis of natural time. A point in time is most often represented as a calendar expression.
There is no cooresponding term in HL7 RIM for this, but a simple Datatype attribute could suffice (whose definition infers a specific unit or the unit can be metadata on the value - see: InterpretationProperties)
A controlled vocabulary (a datatype attribute with a closed list of values) could probably suffice or alternatively disjoint subclasses of rim:Person to represent both genders.
Yet to be reviewed:
- Response to stress testing (more severe)
- Acute myocardial infarction
- Hemodynamic instability (grade 0 to 4) (more severe)
- NYHA functional class (I to IV) (higher)
Angina symptoms can be captured as instances of rim:Observation.
An alternative option (not taken here) is to model them more ontologically than as a controlled vocabulary:
Class(AnginaPectoris complete ChestPain restriction(isSpecificImmediateConsequenceOf someValuesFrom = MyocardialIschaemiaLesion) )
Angina Classifications (Canadian class 0 to IV) (more severe)
Angina classifications could be captured using a rim:value (on rim:Observation) attribute with numeric values corresponding to the class (0,1,2,3,4).
GALEN refers to the unstable/stable angina modifiers as severity levels:
Class(StableAngina complete AnginaPectoris restriction(hasSeverity someValuesFrom = intersectionOf( Severity restriction(hasQuantity someValuesFrom = intersectionOf(Level restriction(hasTrendInState value = stable)) ) ) ) )
The severity levels of 'unstable' and 'stable' can be values of the rim:interpretationCode attribute.
Distribution and severity of coronary artery disease (CAD)
- Left main coronary artery disease
- 1-3 vessel disease
- Myocardial Ischema
Left ventricular dysfunction (grade 0 to 4) (more severe)
Aggressiveness of arteriosclerotic process
- Diffusely diseased coronary arteries
- Peripheral vascular disease
- Cerebrovascular disease
- Hyperlipidemia (more severe)
- Date of operation
- Nonuse of IMA to LAD
- Incomplete revascularization
- Perioperative myocardial infarction (inadequate myocardial management)
A content ontology for this usecase would need to be able to capture these factors as data. Galen already has quite a good coverage as well as the internal Patient Record Ontology the Cleveland Clinic's Cardiothoracic Surgery Research Department is developing (with 'extensions' for the domain of cardiothoracic surgery).