Improving research and patient outcomes by linking public and private health datasets at the Kolling Institute of Medical Research

The Kolling Institute of Medical Research – one of the leading centres of health and medical research in Australia – hypothesised that the integration of data can improve the appropriateness, specificity and efficiency of healthcare delivery through interdisciplinary research. Alcidion helped the Kolling Institute collect and integrate data from a number of different public and private health datasets to create a single research database. Now clinicians and research staff have an integrated view of the data across different systems, which helps them to see variation in care and make informed decisions about their patients during their episode of care.
  • 900K
    patient episodes
    consolidated into a single research database
  • 500K
    pathology/radiology results
    consolidated into a single research database

Challenge

The Australian healthcare system generates large amounts of data from public and private clinical and non-clinical sources.  The lack of connectivity across healthcare services has resulted in significant impacts on the quality of care and patient safety.

The Kolling Institute hypothesised that the integration of data can improve the appropriateness, specificity and efficiency of healthcare delivery through interdisciplinary research.

Alcidion was engaged to extract data from multiple sources (public and private) and to create an SQL data store to link the extracted data items.

The objective for one group of stakeholders is to improve patient care by answering clinical questions based on integrated data.  For others, the goal is to provide a dashboard of linked data to support multi-disciplinary meetings by consolidating all relevant data for a set of nominated patients.

Solution

Data was collected and integrated across three exemplars: Head and Neck Cancer, Maternity and Cardiology. Alcidion provided technical and solution skills to link the information available across a number of sources of patient information such as Cerner eMR, McKesson’s Horizon Cardiology System and the ObstetriX maternity system.  Pathology results have also been incorporated by linking to private pathology providers and this data is matched with existing data sets.

Alcidion used its Cerner application and Cerner Command Language (CCL) skills to extract data from the eMR solution.  This data is stored in a central Microsoft SQL repository and is matched to data that has been extracted from other systems.  Extracts have been created using a combination of direct SQL calls, SSIS load routines and real time HL7 messaging.

The integrated datasets are provided to the clinical stakeholders via a number of QlikView applications developed by Alcidion.  Our consultants worked closely with clinicians and user groups to define the data requirements and visualisation techniques.  These applications are then made available to the user base via a web browser using a comprehensive security and access model.

Results

The three exemplars now have access to a research database that contains more than 900,000 patient episodes and 500,000 associated pathology/radiology results.  This data is used to support Multi-Disciplinary Team (MDT) meetings and to answer specific clinical questions. For example:

  • How many of our patients are anaemic just prior to birth and who are they?
  • Of our patients who are discharged how many are not breast feeding and where do they live?
  • How many of our patients re-present following birth?
  • How great is the variation in care or women in their first pregnancy?

The clinicians and Kolling Institute research staff now have an integrated view of the data across different systems which helps them to see variation in care and make informed decisions about their patients during their episode of care.  Individual dashboards and search criteria have been created to suit each exemplar.

The project is currently working with other interested parties to integrate additional datasets, for example, external ultrasound providers and private hospital providers.

An ongoing operational model was implemented to enable the agreed datasets to be refreshed from the production systems.  It is expected that alternative analytical tools will be investigated and other research groups will become involved.  For example, a group at the Institute has been engaged to conduct natural language processing on the textual patient results.  It is expected that when successful, the results of this research project will be extended to include data from other New South Wales Local Health Districts.

Customer Success Stories