Health Data Management & Semantics
Aim: to generate an accurate cancer stage from histo-pathology reports, and to use this information to help improve cancer management, both for individual patients and at a population-level.
The “stage” of a cancer is a categorisation of its progression in the body, and describes the extent of the primary tumour and any spread to local or distant body sites. While staging has a fundamental role in cancer management, due to the expertise and time required and the multi–disciplinary nature of the task, a definitive stage for cancer patients is not always collected. By automating the collation, analysis, summarising and classification of relevant patient data, the reliance on expert clinical staff can be reduced, improving the efficiency and availability of cancer staging.
The CSIS Software
The Cancer Stage Interpretation System (CSIS) technology uses the same guidelines as those used by clinicians in assigning a stage based on histo-pathology reports. The pathology report is divided into statements (sentences or part of a sentence) while the guidelines are divided into factors. Each statement in the report is then evaluated against each of the guideline factors for relevance and whether it is a positive or negative reference. Machine learning techniques are then used to teach the CSIS engine the set of statements in a pathology report which indicate that the report is describing a particular guideline factor.
In addition, an extract is produced consisting of sentences that were found to contribute to the final staging decision, and their relationship to criteria from the formal staging guidelines for lung cancer.
Lung Cancer Clinical Trial
In collaboration with the Queensland Cancer Control Analysis Team (QCCAT), the software prototype system was used within a clinical trial context. The CSIS engine was trained on a set of 710 pathology reports describing surgical resections of the lung. The aim was to produce an accurate pathological T and N stage. The system was then formally trialled in a clinical setting on a previously unseen set of 179 lung cancer cases. The trial compared the automatic stage decisions from CSIS to the stages assigned by two expert pathologists.
The results of the trial showed that the automated stages produced were accurate enough for the purposes of population level research and for indicative staging of pathology reports prior to multi–disciplinary team meetings and have been published in the Journal of American Medical Informatics Association.
The CSIS software is now installed at QCCAT.
This work has been extended to include other sorts of medical free text, such as radiology reports (for M staging) as well as automatic population of synoptic reports for lung cancer, and continues in the Medical Free Text Processing project.
Aim: to develop a trusted health data integration service delivering better health and research outcomes through novel data linkage mechanisms.
Research into the management and delivery of healthcare is critically dependent on access to data, however much of this data resides across many data repositories and organisations, and is often highly protected and private. Australia has a rich collection of health and community data repositories that could potentially be linked to help find answers to important health and social questions. Bringing these data repositories together would enhance greatly our ability to tackle diseases and understand complex issues.
HDI™ is a data integration tool developed at the Australian e-Health Research Centre (AEHRC) to provide private and secure access to an integrated virtual data repository, enabling research and analysis on a larger scale than would be otherwise be possible. It provides sophisticated infrastructure for publishing, locating retrieving and analysing data in large-scale, diverse and distributed information systems.
HDI will lead to benefits for all Australians by enabling:
Through HDI, CSIRO is enabling a network-based research infrastructure to develop the equivalent of a major virtual data repository that links individual data repositories. This will:
One tool, many data sources
clean, linked data can be extracted, analysed and delivered from disparate databases, delivering all the functionality in a single, easy to use tool.
Data privacy, integrity & security
Virtual data repository
a virtual repository of data, rather than the warehousing of data, ensures that control of data stays with the data custodian.
a service-oriented approach that builds on leading-edge web services technology.
industry standards-based protocols, algorithms and software.
implemented in Java for platform independence.
Security & Privacy
Increasing concerns over privacy and confidentiality, coupled with a growing body of legislation and codes of practice governing the use of personal and health data, means that sharing health data for research purposes across health data custodian boundaries poses technical, organisational and ethical challenges.
HDI uses privacy-preserving linking algorithms to protect and identity and personal information.
Integration & Linking
HDI supports the networking of health data repositories by:
HDI takes a federated approach to integrating data repositories, with web services used to query the data repositories and retrieve results for further analysis and reporting. Metadata is used to describe the data repositories and present a view of the data to users and is used in the planning and executing of complex queries across the data services.
Linking of patient records in different data repositories, while maintaining the privacy of patients, is core HDI functionality. Matching of patients across these data repositories is made possible using encrypted demographic data, meaning identifying information remains protected.
Aim: to develop new mathematical and algorithmic techniques for medical image watermarking and experimental validation of their suitability.
Medical image protection and authentication are becoming increasingly important in an e-Health environment where images are readily distributed over electronic networks. Research has shown that medical image watermarking is a relevant process for enhancing data security, content verification and image fidelity. At the same time, it is necessary to preserve as much original information in the image data as possible, to avoid causing performance loss for human viewers.
A widely accepted fact in generic image watermarking is that not all watermarking methods are suitable for all image types and all applications. Although this topic has been explored for generic image watermarking, it has received very little attention in medical image watermarking. Most medical image watermarking research focuses on developing watermarking systems that preserve image fidelity and/or robustness, under typical non-medical image degradation processes (for example, data communication losses). However, they do not provide tailored solutions for specific medical image types, or for typical degradation processes arising from typical medical uses involving image manipulations.
Given the range of medical image types that exist, as well as the need to protect images without undue loss of data, and the knowledge that not all watermarking methods are suitable for all image types, an important gap in medical image watermarking research was been identified. If medical images are to be protected appropriately as they travel from one site to another, a suitably tailored watermarking scheme must be selected for each image type.
This project investigated the development and pilot implementation of a novel technique for embedding hidden ‘watermark’ information in medical image data files which would be useful for enhancing security and privacy protection, and at the same time would allow estimation of the amount of data loss (if any) that has occurred to an image due to successive manipulation, transmission and storage operations. This is important in ensuring that safety and quality standards are maintained, by protecting patient data from misuse by allowing ownership and access information to be embedded, and protecting health sector users from unknowingly using degraded data for critical clinical purposes.
The project was run as an e-Health Research Centre partner project, in conjunction with NICTA Qld Laboratory and with collaborating investigators based at the Queensland University of Technology and the University of Queensland.
Aim: to establish the relative efficacy of personal and ambulatory devices to detect movement and precursors, particularly in aged patients, that may indicate a likely fall, with the aim of providing intervention to prevent the actual fall.
Factors such as ageing population, rising health costs and the increasing incidence of long-term chronic disease are generating healthcare challenges of an unprecedented scale. These issues are driving a trend towards increasing levels of care in the home, with early discharge from hospitals, or ‘ageing in place’ initiatives, in which the elderly are encouraged to maintain independent living for as long as possible.
As part of the program of ambulatory monitoring of stroke and elderly patients, a Smart-State initiative of the Queensland Government, the e-Health Research Centre conducted clinical trials to establish the relative efficacy of personal and ambulatory devices to detect movement and precursors, particularly in aged patients, that may indicate a likely fall, with the aim of providing intervention to prevent the actual fall.
In addition to movement monitoring, the e-Health Research Centre evaluated systems capable of continuously monitoring and recording patient vital signs information, such as heart rate. While the trials were primarily focused on patients in clinical settings, it is anticipated that similar approaches could be extended to the home environment.
Falls in the aged community have been a significant problem for many years and often result in major injury leading to disability or death. In Queensland alone, around 300 deaths and 16,000 hospital admissions per year are attributed to falls in older people .
In particular, falls in older people with stroke, who’s conditions are often managed with medications such as vasodilators, anti-arrhythmics and diuretics are a concern. Many of these drugs result in a significant lowering of blood pressure and can lead to orthostatic or cardiac arrhythmia syncope, conditions that result in markedly increased rates of mortality and sudden death.
In responding to the needs of this group of vulnerable people, aged care clinicians and nursing staff would be significantly assisted by the use of personal and ambulatory monitoring devices to monitor patient vital signs and activity. The ability to monitor vital signs remotely would also facilitate the remote management of medications.
Devices that detect human movement are available on the market and are gaining acceptance in non-clinical environments such as sports medicine.
The continuous personal monitoring of patients may provide a number of advantages:
Early detection & intervention
Events such as heart arrhythmias, falls and stumbles are detected, alarmed and recorded as they occur allowing clinicians, nursing staff or carers to intervene quickly.
Analysis & interpretation
Clinicians can draw upon accurate data for analysis and interpretation.
Personal monitoring reduces the burden placed on already over-extended clinical staff and carers.
Patients receiving in-home care are able to maintain a degree of independence in surroundings that are familiar to them.
Patients living in geographically remote communities who require medical monitoring without the need for hospitalisation can be monitored.
The cost of providing continuous personal monitoring is lower than the costs of transporting and accommodating patients.
Two categories of monitors were trialed:
Communications Systems Options
Personal monitoring technology can use standard wireless transmission of data to a PC/Laptop/PDA when indoors or via mobile phone/conventional phone when outdoors/home. Devices can also record data to an internal memory card in situations where wireless protocol cannot be used.
Data System Possibilities
Information recorded could be accessed by carers using a standard Web browser. In an emergency, the system software could have the ability to send text messages to alert family, physicians, or other carers or emergency services directly.
Options for Applications
Use of personal monitoring systems with wireless technology will enable real-time, remote monitoring while the patient can continue with normal daily activities. Patient information can be stored in a secure database allowing access by clinicians or carers.
Aim: to develop a decision support tool to assist patients to self-manage their own care and assist clinicians and allied health workers to remotely assess and enhance the progress of patients in rehabilitation.
Most chronic disease patients undergo a period of rehabilitation at their home following treatment in a hospital environment. A large portion of the rehabilitation is intended to aid in the recovery of functional ability and to enable normal daily living activities. The current practice of clinical assessment of functional rehabilitation remains mainly a qualitative and subjective assessment process by the clinician, physiotherapist or occupational therapist and not a reliable, quantitative assessment. Patients must typically travel to the hospital or clinic on a regular basis for such an assessment.
Currently, there are no tools that can remotely perform accurate, quantitative assessments of mobility and physical activity status in rehabilitation. The e-Health Research Centre (eHRC) is developing a decision support tool to assist patients to self-manage their own care and assist clinicians and allied health workers to remotely assess and enhance the progress of patients in rehabilitation. To develop this tool the eHRC will partner with device companies and healthcare providers (for example, community health hub; aged care centre, nursing home or hospital department) to acquire movement and vital signs data and evaluate the tool as part of a clinical trial.
The pilot phase in this research project is the development of a Rehabilitation Assessment Decision Support (RADS) Tool. The tool or model receives inputs from various devices. This might provide data for the monitoring of individuals and/or provide evidence over a group or cohort of patients. Ultimately it can be provided as a service for patients in their home (for example, in-home monitoring by a community and residential care service provider). It is intended to link this tool to various personal measurement devices that detect movements and vital signs.
In a subsequent phase this technology will be adapted and enhanced: