Wednesday 1 Nov | 10:00-11:00

 

Part 1: Artificial intelligence (AI) expedites patient throughout and accelerates growth in Hospital-in-the-Home

A rapid improvement event approach was utilised to develop an automated solution to drive growth in at-home patient care. Clinical and Analytics stakeholders developed and validated a machine learning algorithm which identifies patients within the Emergency Department (ED) and Inpatient Wards who may be suitable for at-home care. Identified patients are ‘flagged’ on live, point-of-care Electronic Patient Journey Boards, alerting the treating team and Hospital in the Home (HITH) of a potentially suitable patient. Expedited referral and review enables more patients to be considered for home sooner, meeting consumer expectations and supporting hospital bed access and flow.

Objectives:

• Understand how a range of health business stakeholders and clinicians came together to solve a business problem using human-centred design and an agile approach


• Understand how AI has been utilised to identify patients with specific attributes


• Understand how AI has been integrated into business operations to expedite patient throughout
  
  

Corinne Howell, St Vincent’s Virtual & Home; Australia

Bede McKenna, St Vincent’s Hospital Melbourne, Australia

 

Part 2: Surgical safety management with AI: A prospective study in a large-scale ophthalmic surgery centre

If medical professionals mistakenly identify patients or confuse left and right, would they honestly report everything? We used artificial intelligence (AI) to verify patient IDs, left and right, and intraocular lens authentication in a large-scale ophthalmic surgery centre for a year. As a result, we discovered errors or near misses at a rate more than 20 times higher than traditional reporting. If only one-twentieth of the mistakes are being reported, the improvement cycle would likely not function properly. We will explain whether the AI system is wrong, or if humans are creatures capable of reporting only one-twentieth of their mistakes.

Objectives:

• Understand human biases related to hiding mistakes


• Understand the capabilities of AI systems


• Understand the principles of psychological safety

Hitoshi Tabuchi, Hiroshima University, Japan

Yasuyuki Nakae, Tsukazaki Hospital, Japan

Masahiro Akada, Tsukazaki Hospital / Kyoto University, Japan