Current Collaborative Projects 
| Current Collaborative Projects |
|
1. Computation of a SNOMED CT Subset for Intensive Care Services and ED Clinical Notes Partners ICU: Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia ED: Concord Hospital, Emergency Department, Concord, NSW, Australia University of Dresden, Germany Project Description This project aims to assemble a subset of SNOMED CT suitable for use in Intensive Care Units and ED. A collection of clinical notes taken from the RPAH-ICU and Concord-ED have been analysed for SNOMED (SCT) codes. This set of codes contains both false positives, probably 20-30%, and false negatives, possibly around 10%. Using the corpus a subset of the SCT ontology can be computed representing a complete but closed portion (the transitive closure) of the total ontology. Other processes both manual and computational can then be applied to cleaning the closed portion to make it smaller and ontologically cleaner. The result will be a subset suited for use in ICUs in the case and EDs in the other case. Prelimenary work indicates that in a corpus of 64 million words of ICU progress notes, that 90% of the SNOMED terms found are covered by about 2000 SNOMED concepts, 96% by about 5000 concepts and 100% by 20,000 concepts. 2. Predictive Detection of Acute Respiratory Distress Syndrome from the Patient Record Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia Mayo Epidemiological and Translational Research in Intensive Care (METRIC), Mayo Clinic, Rochester, Minnesota, USA Project Description The early detection of Acute Respiratory Distress Syndrome (ARDS) is important in critical care settings. This project has the objective of demonstrating that computer based analysis of both the text and quantitative components of the clinical records can feed a computational model of patients at risk. The model needs to be accurate enough to predict correctly patients at high risk without raising too many false alarms that will unnecessarily distract clinicians from other important tasks. The two institutions will share data in real-time so as to consult with each other to apply multiple expertises to assess patient cases. 3. Intensive Care Real-time Audit Information System Partners Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia Project Description This project has the aim of providing an audit of care in the ICU in real-time. The notion is that a computational model of the patient needs to extracted from the clinical information system including the patient progress notes. Then a similar computational model needs to be built of the care protocol for the patient's condition. An audit will be achieved when the patient model and the care model are folded against each other and scrutinised for differences. The differences can represent weaknesses in the care of the patient or local knowledge important for modifying the care plan protocol. 4. Handovers Information Retrieval System Partners Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia Children's Hospital, Paediatric Intensive Care Unit, Westmead, Australia Project Description The process of Handovers needs to ensure that sufficient, necessary and only timely information is passed from a carer or group of carers to another. This research involves devising methods for specifying the data required for a Handover and then automating the generation of a customised Handovers information system, which may vary according to ward, clinician role, patient type, or disease class.
5. Nursing - Clinical Information System Partners Sydney West Area Health Service Project Description This project involves using the Laboratory's information system generator software to create a nursing clinical information system (N-CIS) for experimental work at the SWAHS. The N-CIS will be based around delivering the Nursing Interventions Classification (NIC) and Nursing Outcomes Classification (NOC) to the nursing station.
6. Mapping SNOMED CT to ICD 10AM for ICU Clinical Notes Partners National Centre for Classification In Health Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia Project Description The purpose of this project is to assess the accuracy to which ICD 10AM codes can be computed with SNOMED CT (SCT) encoded clinical notes. A subset of records will be drawn from the ICU and be manually encoded for SCT codes followed by ICD 10AM encoding. Then automatic classifiers will be built to try to replicate the manual codings. 7. Web based Visualisations of very large medical ontologies Partners Department of Biomedical Engineering, Tehran University of Technology, Iran Understanding very large ontologies is very difficult for the experienced researcher let alone a novice to the filed. We have developed a terminology server to deliver any large clinical classification or ontology via web service. This has limited presentation functionality and in particular is poor at showing multiple pathways through a network to a given node. The aim of this project is to use third party software to create an attractive and useful presentation of any ontology or classification delivered by the terminology server as web pages. 8. Workflow System for Intensive Care Unit Partners Royal Prince Alfred Hospital, Intensive Care Service, Sydney, Australia A system for generating workflow systems for hospital settings has been developed in previous projects and now needs to be used to create an implementation for the Intensive Care Unit. This work requires a study of the processes in the unit and then using the software to create an operational system. 9. Data Mining Clinical Information Systems Partners South West Area Pathology Service Royal Prince Alfred Hospital Hospitals produce very large quantities of data which represents a great deal of their work activities, however little is understood about the structural information this data can provide. The aim of this project is to e mine the data in a number of clinical information systems in an attempt to understand more about systemic features in the care of patients.
|
