Health and wellbeing products and services for individuals are becoming increasingly popular as people realise the benefits provided by lifelogging or quantified-self platforms in such areas as exercise, diet management and mood. However, in addition to the data that users record using these platforms, all user interactions and events can be elusively logged to represent usage. Such user interaction or event logs provide rich and large datasets that can fuel applied artificial intelligence. As products and services based on these digital interaction technologies are taken up across public healthcare provision, should healthcare policy and practice take more cognisance of the opportunities and risks in gathering interaction data? Is 'healthcare' ignorant that there is knowledge in such data? Are there differences between event logging in healthcare and other areas such as commerce, media and industry? In order to realise benefits in analysing such data, methods that help ensure consistency, accuracy, data protection, as well as reproducibility of knowledge derived from log data need to be examined. This paper presents methods to explore usage log data and a process workflow followed by a presentation of two real world case studies. The workflow has been coined Health Interaction Log Data Analysis (HILDA) and focuses on data prospecting and machine learning stages to show the opportunities realisable in analysing interactional or event data automatically recorded by digital healthcare services.
展开▼
