This post was written by Brendan DUNPHY, CEO of C-BIA Consulting and first published on AIMed
Use of IT systems today varies across specialities with some more open to and active users of systems and data to inform decision-making, others more reticent. Just because systems are available does not mean they are consistently by all clinicians or used well.
Current IT systems are mostly labour-intensive, clunky and disconnected, underwhelming examples of mediocre office technology poorly adapted to a more important and stressful environment. According to one expert, an average UK hospital may have 1,700 IT systems and 1,000 medical devices.
Worse, the systems obtrusively insert itself into the space between patient and clinician. Though their remit is to improve, in practice, they simply provide a recording of a data-light and potentially subjective clinical interaction. The need for better less intrusive systems to augment, inform and challenge clinical decision making is clear, even if the solutions are not yet available.
There is progress on open data and other healthcare-specific standards including system interoperability, a major challenge today. The development and widescale adoption of these standards would make a practical difference to clinicians, reducing data entry, re-using existing data and moving towards a ‘single view of the truth’. But vendors (sometimes against their commercial interest) and providers need to agree and implement these, so their impact will be gradual and limited. In some ways interoperability will only serve to re-enforce the current order, raise barriers to innovation and extend the life of the sub-optimal systems in place today.
To be used and loved, systems need to be above all user-friendly and fit easily into established workflows. Individual system log-on is a major annoyance.
Systems need to be believable, trusted and conform to expectations. Traditional IT does a poor job on usability demanding keyboard (real or virtual) and too much input – but voice and robotics promise to change this. Neither are sufficiently developed today for use beyond narrow niche applications. The high reliability required in medical scenarios means advanced user interfaces such as voice (and AI) will lag in healthcare relative to retail, finance and manufacturing where these technologies are already widely used.
When asked, most clinicians simply want a single space to record their notes, not fields to fill or boxes to tick. Even better, they want to a robot to follow-them around and record and interpret what they say, when they say it; Natural Language Understanding (NLU) is still a long way off so this will remain a dream for the time being!
This is a review of AIMed Europe 2018, learn more about the leading global event for clinicians in AI in healthcare here.