Some helpful documents for understanding the development and status of GRiST
Final newsletter from the GRaCE-AGE project that sums up its successes and lays out the sustainable activities from it.
|AMIA Annual Symposium Proceedings, 2016||
When assessors evaluate a person's risk of completing suicide, the person's expressed current intention is one of the most influential factors. However, if people say they have no intention, this may not be true for a number of reasons. This paper explores the reliability of negative intention in data provided by mental-health services using the GRiST decision support system in England. It identifies features within a risk assessment record that can classify a negative statement regarding current intention of suicide as being reliable or unreliable. The algorithm is tested on previously conducted assessments, where outcomes found in later assessments do or do not match the initially stated intention. Test results show significant separation between the two classes. It means suicide predictions could be made more accurate by modifying the assessment process and associated risk judgement in accordance with a better understanding of the person's true intention.
REFERENCE AS: Zaher, N. A., & Buckingham, C. D. (2016). Moderating the Influence of Current Intention to Improve Suicide Risk Prediction. AMIA Annual Symposium Proceedings, 2016, 1274–1282.
|Innovation in Medicine and Healthcare||
One of the main challenges of classifying clinical data is determining how to handle missing features. Most research favours imputing of missing values or neglecting records that include missing data, both of which can degrade accuracy when missing values exceed a certain level. In this research we propose a methodology to handle data sets with a large percentage of missing values and with high variability in which particular data are missing. Feature selection is effected by picking variables sequentially in order of maximum correlation with the dependent variable and minimum correlation with variables already selected. Classification models are generated individually for each test case based on its particular feature set and the matching data values available in the training population. The method was applied to real patients’ anonymous mental-health data where the task was to predict the suicide risk judgement clinicians would give for each patient’s data, with eleven possible outcome classes: zero to ten, representing no risk to maximum risk. The results compare favourably with alternative methods and have the advantage of ensuring explanations of risk are based only on the data given, not imputed data. This is important for clinical decision support systems using human expertise for modelling and explaining predictions.
This is the first newsletter from the GRaCE-AGE project. We will release another one in a similar format to this at the end of the year. In the meantime, we will release news as and when it comes using articles posted into the GRaCE-AGE group.
|Mental Health in Family Medicine||
Aim To explore current risk assessment processes in general practice and Improving Access to Psychological Therapies (IAPT) services, and to consider whether the Galatean Risk and Safety Tool (GRiST) can help support improved patient care.