Effective, reliable and integrated technology systems are key enablers for the provision of quality healthcare. These systems can help improve quality, consistency and efficiency across the continuum of care; enhance physician and staff experience by improving information pooling and exchange, transparency and accessibility; and facilitate reporting and data mining.
However, at times, technology can also stand in the way of effective delivery of care, particularly when a problem – like slow log times or failing hardware – impacts clinical flow. One of our clients was facing this challenge and engaged The Deetken Group to identify where deficiencies in technical support (“service failures”) were adversely impacting workflow and productivity and – critically – compromising patient care.
The Deetken Group leveraged a number of machine learning techniques to identify and quantify the impact of these service failures. These included the application of topic modelling and data mining algorithms. Using these techniques, we were able to automate the review of over 1 million requests from staff for support (in the form of “call centre tickets”) and to identify specific instances of service failures. We could then accurately determine the error rate of call centre tickets and estimate the additional workload and productivity impacts of these errors. Overall our analysis identified that over 4,000 workdays per year were lost due to these call centre errors. As a result of our work, our client asked us to design the optimal support system (in this case, a “call centre model”) for resolving the challenges uncovered through our research.