Smart Text Analytic Tools (STAT) for analysis of patient-centred communications to strengthen health systems in BC

Principal Investigator: 
University: 
University of British Columbia
Faculty: 
Faculty of Medicine
Department: 
Department of Medicine

Healthcare stakeholders, including health authorities, facilities, pharmaceutical companies and insurers are increasingly acknowledging the importance of big data to enhance understanding of health behaviours and health systems. Existing analytic tools available to navigate the volume of diverse data types at a frequency that can match the speed at which data is generated are in early stages of development, and often lack validation due to limited access to health data. The ability of healthcare stakeholders to make sense of this valuable data is restricted by a lack of capacity and user-friendly analytic tools. 

Dr. Lester leads the UBC mobile health (mHealth) research team, which has been developing a set of smart-text-analytic-tools (STAT) to analyze patient and care provider communication data in the form of open natural language text. The WelTel digital health platform was created by Dr. Lester and has been tested in a diversity of geographic (Kenya, Canada, USA, South Africa) and health settings (HIV, TB, Asthma, maternal and child health) since 2005. The result of twelve years of mHealth research is a dataset consisting of hundreds of thousands of text messages sent by patients and providers. Topics of discussion include advice related to medication side effects, information requests, and the need for access to psychosocial and logistical support services. This data has the potential to identify outpatient self-reported priorities over time, informing patient-centered improvements in health system responsiveness and preparedness. 

The UBC mHealth research group will further develop STAT into a minimum viable product  that can analyze a variety of open natural language text data using natural language processing. This tool will allow both public health systems and private enterprise to streamline approaches to analyzing large volumes of text-based data.

Host Institution: 
University of British Columbia
Research Location: 
Vancouver Coastal Health Research Institute
Year: 
2017