Applying and Improving a Publicly Available Medication NER Pipeline in a Clinical Cancer EMR.

Clinical NLP can be applied to extract medication information from free-text notes in EMRs, using NER pipelines. Publicly available annotated data for clinical NLP are scarce, and research annotation budgets are often low. Fine-tuning pre-trained pipelines containing a Transformer layer can produce quality results with relatively small training corpora. We examine the transferability of a publicly available, pre-trained NER pipeline with a Transformer layer for medication targets. The pipeline performs poorly when directly validated but achieves an F1-score of 92% for drug names after fine-tuning with 1,565 annotated samples from a clinical cancer EMR - highlighting the benefits of the Transformer architecture in this setting. Performance was largely influenced by inconsistent annotation - reinforcing the need for innovative annotation processes in clinical NLP applications.

Visualising Variation in the Real-World Clinical Delivery of Chemotherapy Protocols.

Typical univariate measures of variation in chemotherapy protocols fail to capture and describe the full multi-dimensional complexity of treatment adjustments in real-world data. In this preliminary work, we propose novel visualisations of observed treatment events, as well as treatment-as-delivered relative to initial prescriptions, as a means of gaining insights into complex patterns of treatment variation in cancer patients. Simple clustering techniques were also used to confirm the utility of these visualisations and our ability to correlate observed variations with historical events.

Rates, types and associations of sexual risk behaviours and sexually transmitted infections in those with severe mental illness: a scoping review.

The aim of this scoping review was to: (i) determine rates and types of sexual risk behaviours and sexually-transmitted infections (STIs) in those with severe mental illness (SMI); and (ii) delineate correlates of poor sexual health outcome. The online databases OVID MedLine and PsycINFO were searched from databases inception to February 2018 for any literature with a focus on sexual risk behaviours (inconsistent condom use, multiple sexual partners, substance use and transactional sexual acts) or STIs in SMI populations. Fourteen studies were identified; the quality of these studies ranged from poor to moderate. Outcome definitions were heterogeneous, precluding meta-analysis. We found rates of sexual risk behaviours and STIs to be more common among those with SMI than the rates in the general Australian population. Current studies do not acknowledge the relationship of sexual risk behaviours and STI risk and hence do not provide a full model of sexual health outcomes in those with SMI. In order to improve sexual health outcomes in SMI populations, further research is required of greater methodological rigor, with consensus in the definition of sexual risk behaviours, clarifying causal relationships and where in the course of SMI that these outcomes emerge.