Using Natural Language Processing to Predict Risk in Electronic Health Records.

Clinical narratives recording behaviours and emotions of patients are available from EHRs in a forensic psychiatric centre located in Tasmania. This rich data has not been used in risk prediction. Prior work demonstrates natural language processing can be used to identify patient symptoms in these free-text records and can then be used to predict risk. Four dictionaries containing descriptive words of harm were created using the Diagnostic and Statistical Manual of Mental Disorders, the Unified Medical Language System repository, English negative and positive sentiment words, and high-frequency words from the Corpus of Contemporary American English. However, a model based only on these keywords is limited in predictive power. In this study, we introduce an improved NLP approach with a social interaction component to extract additional information about the behavioural and emotional state of patients. These social interactions are subsequently used in a machine-learning model to enhance risk prediction performance.

Bridging the Gap: Enhancing Maritime Vessel Cyber Resilience through Security Operation Centers.

Increasingly disruptive cyber-attacks in the maritime domain have led to more efforts being focused on enhancing cyber resilience. From a regulatory perspective, there is a requirement that maritime stakeholders implement measures that would enable the timely detection of cyber events, leading to the adoption of Maritime Security Operation Centers (M-SOCs). At the same time, Remote Operation Centers (ROCs) are also being discussed to enable increased adoption of highly automated and autonomous technologies, which could further impact the attack surface of vessels. The main objective of this research was therefore to better understand both enabling factors and challenges impacting the effectiveness of M-SOC operations. Semi-structured interviews were conducted with nine M-SOC experts. Informed by grounded theory, incident management emerged as the core category. By focusing on the factors that make M-SOC operations a unique undertaking, the main contribution of this study is that it highlights how maritime connectivity challenges and domain knowledge impact the M-SOC incident management process. Additionally, we have related the findings to a future where M-SOC and ROC operations could be converged.

Risk prediction using natural language processing of electronic mental health records in an inpatient forensic psychiatry setting.

OBJECTIVE: Instruments rating risk of harm to self and others are widely used in inpatient forensic psychiatry settings. A potential alternate or supplementary means of risk prediction is from the automated analysis of case notes in Electronic Health Records (EHRs) using Natural Language Processing (NLP). This exploratory study rated presence or absence and frequency of words in a forensic EHR dataset, comparing four reference dictionaries. Seven machine learning algorithms and different time periods of EHR analysis were used to probe which dictionary and which time period were most predictive of risk assessment scores on validated instruments. MATERIALS AND METHODS: The EHR dataset comprised de-identified forensic inpatient notes from the Wilfred Lopes Centre in Tasmania. The data comprised unstructured free-text case note entries and serial ratings of three risk assessment scales: Historical Clinical Risk Management-20 (HCR-20), Short-Term Assessment of Risk and Treatability (START) and Dynamic Appraisal of Situational Aggression (DASA). Four NLP dictionary word lists were selected: 6865 mental health symptom words from the Unified Medical Language System (UMLS), 455 DSM-IV diagnoses from UMLS repository, 6790 English positive and negative sentiment words, and 1837 high frequency words from the Corpus of Contemporary American English (COCA). Seven machine learning methods Bagging, J48, Jrip, Logistic Model Trees (LMT), Logistic Regression, Linear Regression and Support Vector Machine (SVM) were used to identify the combination of dictionaries and algorithms that best predicted risk assessment scores. RESULTS: The most accurate prediction was attained on the DASA dataset using the sentiment dictionary and the LMT and SVM algorithms. CONCLUSIONS: NLP, used in conjunction with NLP dictionaries and machine learning, predicted risk ratings on the HCR-20, START, and DASA, based on EHR content. Further research is required to ascertain the utility of NLP approaches in predicting endpoints of actual self-harm, harm to others or victimisation.

Online computer-aided vicarious exposure for OCD symptoms: A pilot study.

BACKGROUND AND OBJECTIVES: Computer-aided vicarious exposure (CAVE) for obsessive-compulsive disorder (OCD) is an intervention in which participants learn and rehearse exposure with response prevention (ERP) by directing a character around a virtual world. This study aimed to pilot an online CAVE program for OCD in a community sample with high OCD symptomatology. METHODS: Participants (n = 78) were allocated to an intervention group (three 45-min weekly CAVE sessions) or to a waitlist control group. The treatment group were asked to complete three 45-min sessions over a four week period. RESULTS: Those who completed at least one CAVE session showed greater improvement on measures of OCD symptomatology at one-month post-treatment (d = 0.49-0.81) compared to waitlist (d = 0.01-0.1). Older age, past treatment and higher symptom severity were associated with non-adherence. LIMITATIONS: These findings should be considered preliminary due to sample size limitations and an absence of an active control group. However, the findings suggest that further development and evaluation of the program is warranted. CONCLUSIONS: Preliminary findings suggest that online CAVE programs have potential to bridge treatment gaps among those reluctant to attend treatment or engage with in vivo exposure exercises. These programs may also have potential applications as an adjunct to face-to-face or online cognitive behavioural therapy.

Internet delivered images for exposure in specific phobia, design considerations for self-directed therapy.

Exposure to phobic stimuli in subjects with specific phobia typically results in increased anxiety, ranging from mild to severe, followed by gradual habituation. The Internet is a candidate medium for the delivery of phobic stimuli to phobic subjects, such as pictures, video clips or computer animations. Delivery of such images in home settings warrants careful attention to the range and time course of anxiety responses elicited, and to tailoring of progression through hierarchies of images. The agency of the user is paramount, they need to have the final say at all stages of exposure as to whether to proceed or not. We have incorporated solutions to these requirements in the design of an internet-based exposure program (FEARDROP). This employs a database repository of pictures and videos of phobic stimuli. Images are called up by the user engaging a tracking circle with their mouse and following it around the screen. The image fades out if the circle is not followed, a form of 'dead man's brake'. Anxiety responses are measured at intervals on a visual analogue scale and graphed for the user. Initial results show substantial habituation to spider pictures within minutes, with a controlled comparison to video images in progress.