Data science to improve clinical practice and clinical trials
Progress of data science and data use is essential to improve clinical trials and real world evidence through regulated and optimized drugs. The quest for inclusive trials seeks fair representation of patient groups in all lived experiences – including people from low and intermediate income countries – which may be more likely to have a earlier start to a Wide range of medical conditions and to be more at risk of having several long -term conditions.
Electronic health records are a key data source to better understand patients, both individually and within the populations. As the electronic data of health files are better captured, linked and organized, the opportunities to understand the risks and trajectories of the disease improve.1-3 These data also allow optimized processing of data that could be reused to improve clinical trials feasibility, recruitment, safety monitoring, economic assessments, generalization studies and long -term results monitoring.
Recent progress in Causal Automatic Learning has the potential to improve patient care, public health measures, quality of services, planning and research – including for clinical trials. For example, automatic learning approaches such as convolutional neural networks (CNN), short -term memory networks (LSTM) and generative opponent networks (GAN) are beginning to allow innovations such as estimation of effects of the treatment or generation of synthetically balanced balance synthetically balanced case control populations and “virtual control groups”. Using automatic learning methods (and using data from previous clinical trials, natural history studies, electronic health records, complaint data or disease registers) to create control groups Virtual, we can further remove our study conceptions from placebo control weapons. With less dependence on human controls, more participants receive innovative treatment rather than placebo or standard care.