This is the second consecutive workshop within ACSW to deal with the topics around health knowledge management and associated systems. Last year we had styled the workshop as the Australasian Workshop on Health Knowledge Management and Discovery (HKMD). This year we've migrated the focus slightly more toward the supporting systems and management of data that underpins a health knowledge management capability --- hence the name Health Data and Knowledge Management (HDKM). >The papers, while small in number, illustrate some of the major foci for HDKM. As the range and ubiquity of electronic health data sets expands, the interest in record linkage continues to grow. Addressing the demand for practical and reliable record linkage, we have Peter Christen's paper, Febrl - A Freely Available Record Linkage System with a Graphical User Interface, which provides an open and user friendly package wherein both students and practitioners can experiment with a range of record linkage techniques. >Issues around security and trust are persistent themes in HDKM. We need reliable systems that deliver the information to those who need it, and yet manage reliably not to deliver information beyond the requirements of legitimate clinical, research and administrative functions. Addressing this theme, Security Enhanced Linux to Enforce Mandatory Access Control in Health Information Systems, by Luis Franco, Tony Sahama and Peter Croll, introduces a flexible and fine-grained access control architecture. Continuing the theme, Open Trusted Health Informatics Structure (OTHIS), by Vicky Liu, William Caelli, Lauren May and Peter Croll, provides a proposal for a viable, trusted architecture for health information systems based on a set of separate but connected trusted modules. Completing our treatment of this theme, Gerald Webers paper, A Platform-Independent Approach for Auditing Information Systems, provides a model for establishing an audit trail of both data input and presentation. >The last theme includes two articles from my own laboratory and looks at knowledge discovery and machine learning applications. In Towards an Architecture for Quality Audit Reporting to Improve Hypertension Management, Thusitha Mabotuwana, myself, Rekha Gaikwad, John Kenelly and Timothy Kenealy report ongoing work to achieve reliable clinical quality improvement data for chronic disease management from existing General Practice electronic medical records. Jaree Thongkam, Guandong Xu, Yanchun Zhang and Fuchun Huang, in Breast Cancer Survivability via AdaBoost Algorithms compare the performance of a data mining pre-processing method and different versions of the popular AdaBoost ensemble learning method to predict 60-month breast cancer survival from a Thai hospital database. Finally, in Automatically Generated Consumer Health Metadata Using Semantic Spaces, Guocai Chen, myself and Joanne Evans describe results in deriving the genre of consumer health webpages from their patterns of language use, with the eventual aim of better matching consumers to their desired online resources.
展开▼
