Unity in diversity: electronic patient record use in multidisciplinary practice

In this paper we examine the use of electronic patient records (EPR) by clinical specialists in their development of multidisciplinary care for diagnosis and treatment of breast cancer. We develop a practice theory lens to investigate EPR use across multidisciplinary team practice. Our findings suggest that there are oppositional tendencies towards diversity in EPR use and unity which emerges across multidisciplinary work, and this influences the outcomes of EPR use. The value of this perspective is illustrated through the analysis of a year-long, longitudinal case study of a multidisciplinary team of surgeons, oncologists, pathologists, radiologists, and nurse specialists adopting a new EPR. Each group adapted their use of the EPR to their diverse specialist practices, but they nonetheless orientated their use of the EPR to each others' practices sufficiently to support unity in multidisciplinary teamwork. Multidisciplinary practice elements were also reconfigured in an episode of explicit negotiations, resulting in significant changes in EPR use within team meetings. Our study contributes to the growing literature that questions the feasibility and necessity of achieving high levels of standardized, uniform health information technology use in healthcare.

Bayesian Active Learning for Classification and Preference Learning

Information theoretic active learning has been widely studied for probabilistic models. For simple regression an optimal myopic policy is easily tractable. However, for other tasks and with more complex models, such as classification with nonparametric models, the optimal solution is harder to compute. Current approaches make approximations to achieve tractability. We propose an approach that expresses information gain in terms of predictive entropies, and apply this method to the Gaussian Process Classifier (GPC). Our approach makes minimal approximations to the full information theoretic objective. Our experimental performance compares favourably to many popular active learning algorithms, and has equal or lower computational complexity. We compare well to decision theoretic approaches also, which are privy to more information and require much more computational time. Secondly, by developing further a reformulation of binary preference learning to a classification problem, we extend our algorithm to Gaussian Process preference learning.