
Figure 1. CogStack – using NLP to unlock health records
CogStack retrieves and extracts information from both structured and unstructured health data and utilises embedded analytics and visualisation technologies to assist in unlocking the untapped value of data from routine clinical care.
Data housed within electronic medical records (EMRs) are often incomplete, contain large quantities of unstructured data stored in proprietary systems and are in different formats (heterogeneity of data sources).
The challenge is most of the information is written in a form which is difficult to retrieve, analyse and learn from. Hence, the wealth of information potentially available within electronic heath records is often inaccessible and underused.
CogStack (Figure 1) applies natural language processing (NLP) and artificial intelligence (AI) technology that allows extraction of information regardless of the format, whether it’s structured information or unstructured within a scanned document or image (e.g. PDFs, free text documents).

Figure 2. CogStack health service use cases
CogStack can facilitate the effective use of EMR data in several ways:
- Enabling free-text search for EMRs and relationships between entities in the text
- Timely risk monitoring and alerting
- Presenting a visual timeline of patient documents to clinicians
- Improving patient recruitment into clinical trials.
CogStack was developed by Bioinformatics and Health Informatics Researchers at Kings College London. The platform has been successfully deployed and implemented in multiple NHS services in the UK and its implementation is being scaled across the NHS.
Monash University have established a collaboration with Kings College London to adapt CogStack to the Australian context.
Project: National Learning Health System (LHS) Data Management Platform
This innovative project received $1.9M in MRFF funding to implement CogStack within our Health services. This allowed Healthcare Leaders and Researchers to harness the power of Health information in electronic Health records to improve Healthcare quality, health outcomes and enable innovative research.
The project was a joint collaboration between:
- Alfred Health
- Monash University
- The National Centre of Health Ageing (Peninsula Health)
- Outcome Health
- Sydney Health Partners
- Health Translation South Australia
- Kings College London
- Brisbane Diamantina Health Partners
- Western Australia Health Translation Network
- Australian Digital Health Agency
- Digital Health CRC
- Safer Care Victoria
In this project, CogStack was adapted for the Australian context with the aim to deploy this platform within the IT environment of our Health Service Partners.
The Monash Partners Learning Health System (LHS) guided three project work streams for the implementation of CogStack (Figure 2). The first Health Centres to pilot the system included:
Alfred Health utilised CogStack and its NLP/AI algorithms for clinical trial participant identification.
Peninsula Health (National Centre for Healthy Ageing) used the CogStack tool to help detect conditions such as dementia that are often poorly documented and difficult to identify in structured electronic health records.
Outcome Health explored alternatives to their current data management system and implemented CogStack to further investigate opportunities to securely use unstructured data and unlock the analytical benefits through the implementation of NLP.
Their work was supported by and developed in conjunction with Monash University.