Virtual Medical Record Implementation for Enhancing Clinical … · Timișoara, Romania Valentin...
Transcript of Virtual Medical Record Implementation for Enhancing Clinical … · Timișoara, Romania Valentin...
Virtual Medical Record Implementation for Enhancing Clinical Decision Support
Faculty of Automation and Computers University ”Politehnica” Timişoara
Timișoara, Romania
Valentin Gomoi, PhD student Daniel Dragu, PhD student
Vasile Stoicu-Tivadar, Prof. dr. ing.
Goal:
Based on:
- HL7 Standards (CDA, vMR)
- a topic map as knowledge source
Increase access to data for Clinical Decision Support Systems
Content:
1. Introduction, Motivation
2. System Architecture
2.1 Tools and standards
2.2 Inference engine
2.3 HL7 CDA Component (Retrieve data)
2.4 Data manager
3. TM-VMR
3.1 Implementation
3.2 Communication
3.3 Results
Medical Guideline:
- systematically developed statements to assist practitioner and patient decisions about appropriate healthcare for specific clinical circumstances.
Medical Protocol:
- provides the information on duration, dosages, procedures which were omitted in the guideline.
Guidelines? Protocols? Recommendations?
Medical Recommendations:
- suggestions representing the result of the inference based on medical guidelines.
1
Why using Clinical Decision Support Systems ?
• local adaptation of the guidelines • compliance with the protocols and guidelines • standardization of medical practice • faster implementation of new practice • better visualisation of medical information
1
Approaches regarding the implementation of the
computer-based guidelines and protocols: • Asbru, • PROforma, • DeGeL, • Arden Syntax, • GUIDE. • Egadss
Existing solutions for Clinical Decisions Support (CDS)
1
Local adaptation and data retrieving problems
ü Existing solution present low adaptability
Adaptable solutions need standards for:
- patient data transfer and representation – HL7 CDA,
vMR - medical rules representation – Arden Syntax
1
2
Data Manager
Inference engine Egadss
TM-vMR
HL7 CDA Components
HL7 CDA Components
System Architecture for increasing CDS interoperability
Data Manager Is used to ensure the communication between the other
system components being called and calling different web
services
Has the roles to respond at different requests from:
• interface,
• medical data sources (HL7 CDA Components),
• inference engine
• vMR-TM
2.4
2
Data Manager
Inference engine Egadss
TM-vMR
HL7 CDA Components
HL7 CDA Components
System Architecture for increasing CDS interoperability
Inference (Egadss):
Egadss- is a clinical decision support system
Uses: - Arden Syntax (MLM) for the representation of medical rules; - HL7 CDA level 3 messages as input; - HL7 CDA level 2 messages as output; - Clips as inference engine;
2.2
2
Data Manager
Inference engine Egadss
TM-vMR
HL7 CDA Components
HL7 CDA Components
System Architecture for increasing CDS interoperability
HL7 CDA Component
- extracts information from different databases;
- represents information in HL7 CDA format;
- implementation: Visual Studio .Net 2008, C# language;
- access to more complex information;
2.3
2
Data Manager
Inference engine Egadss
TM-vMR
HL7 CDA Components
HL7 CDA Components
System Architecture for increasing CDS interoperability
TM-vMR representing vMR with TM
• TM – encoding knowledge and connecting this encoded
knowledge to relevant information resources
• vMR – standard for the representation of medical
knowledge used in different CDS systems
3
vMR
• standard model
• appeared as a necessity for the interoperability
between different CDS and data sources
• contains 131 medical data elements
• improves communication between CDS systems
and other medical systems
3
TM Ø ISO/IEC 13250:2003; Ø a semantic technology; Ø knowledge representation;
3
Ø to qualify the content of topics;
Ø to link topics together;
Ø to filter information;
Ø to structure unstructured information sets;
Ø to merge topics and topic maps;
TM-vMR Implementation
Is realized by using Topincs open source software
Steps
o create a topic type for every vMR class within the vMR
atomic terms;
o model all relationships within the vMR
o define the serialization names for all terms within the
schema
3.1
TM-vMR with CDS Comunication through “tobjects”
• topic map objects in “Topincs” are called “tobject”
• tobjects – represent an data element from the vMR
• tobjects – allow the insertion, deletion, modification and
many other types of special functions ()
3.2
• realized with the help of the web services
• The web services interact with the vMR through
“tobjects”
• client server architecture
• services are consumed from the Data Manager
• NuSOAP – PHP technology
• communication over HTTPS
3.2
TM-vMR with CDS Comunication through “tobjects”
TM-vMR - benefits
o connection of any “vMR compatible” CDS
o extensibility for vMR DAM
o (web) services especially designed for CDS developers
o easy to use knowledge base
o development of a collaborative solution for the capturing of
medical information
3.3
TESTING TM-vMR
3.3
The test was done for the management of diabetes used in
Timișoara Emergency County Clinical Hospital, Timisoara:
- MLMs – where created – containing the medical rules
- PHP Web Services to extract the needed patient data from
the TM-vMR
Patient data needed for the
TESTING TM-vMR
3.3
The test was done for the management of diabetes used in
Timișoara Emergency County Clinical Hospital, Timisoara:
- MLMs – where created – containing the medical rules
- PHP Web Services to extract the needed patient data from
the TM-vMR
Patient data needed for the
PHP web services to access patient data trough tobjects
3.3
$obiectul = Tobject::get('LaboratoryObservationCode'); $val = $obiectul->get_value(); return $val;
Creating to the needed tobject
Get the needed value
Patient data used
3.3
Data Type Values Units
Glicemia 40 – 350 mg/dl
Alkaline Reserve 6 – 26 mmol/L
pH 6.8 – 7.45 -
Sodium 120 – 150 mmol/L
Potassium 2 – 7 mmol/L
weight 70 – 120 Kg
Urea 20 – 70 mg/dl
Ketones 1 / 0 Boolean
Results
3.3
* The system was tested for these 30 patients data sets offering the expected recommendations.
Recommendation: Administrate 1000 ml NaCl
Concentration 0.9 %, 8 units nsulin
Patient data
Conclusions A method was developed to make medical recommendations more:
=>Complex =>Accurate
by allowing access to more complex data and increasing CDS interoperability • connection of any “vMR compatible” CDS systems • CDS connection with various types of data sources
- Advantages resulted using these tools: ü increase in quality of medical care ü providing more efficient treatments ü using new medical knowledge in current clinical practice
Thank You for Your attention !
This work was partially supported by the strategic grant POSDRU/88/1.5/S/50783, Project ID50783 (2009), co-financed by the European Social Fund – Investing in People, within the Sectorial Operational Programme Human Resources Development 2007-2013