Applications

Project Osler

Developing AI-based OR scheduling algorithms to maximize throughput and reduce surgical backlogs. Now accepting a limited number of trial participants.


A Global Opportunity

COVID-19 paralyzed healthcare systems worldwide, delaying millions of procedures and creating an unprecedented surgical backlog.

One promising approach to reducing this backlog involves using AI models to seek and resolve scheduling inefficiencies that lower OR throughput. Far better at identifying effective resource combinations than humans, these models can significantly aid decision-makers if properly integrated.

1,000,000

Backlogged joint and spinal surgeries by mid-2022 in the U.S.


3,000

Deaths in the U.K. due to delayed cancer treatments by 2026


$1.3 Billion

Estimated cost to clear just part of Canada’s backlog

The Perioperative Chokepoint

Surgical Programs are often the largest cost and revenue sources for hospitals. Despite this, most hospitals still manage their operating room schedules with a patchwork of manual tools.

Even modern, all-encompassing Health Informatics systems require focus and technical expertise which is often lacking in budget- and time-constrained departments. This results in longer turnaround times, fewer surgeries, and a lower total procedural capacity. It is time to change this.

Project Osler | A Joint Initiative

Project Osler is a collaborative effort to build a comprehensive, AI-driven scheduling system for the perioperative department. As a multi-hospital project coordinated by KDS, it will explore where data and constraint-solving AI models can improve KPIs such as wait time, cancellations, and resource utilization.

By joining this initiative, healthcare institutions can gain a front row seat to a landmark study that may have significant implications for their own OR efficiency.

Project Objectives

Project Osler’s software will help decision support specialists optimize surgical scheduling. To do this, it will focus on improving three main objectives of the perioperative process itself:

Peaks and valleys in surgical demand cause stress hospital resources and create downstream challenges. Project Osler will help planners optimize their OR block schedules to reduce variability and create a more uniform pattern of resource utilization.

Multiple ORs often share materials, equipment, and personnel. Osler will help planners take this into account when scheduling surgeries to realize economies of scale and reduce PACU bottlenecks.

Factors like turnaround time, equipment delivery, and seasonal hours can significantly impact patient flow. Osler will advise planners about the potential impact of these external influences, and help schedule surgeries accordingly.

Success Criteria

OR managers are usually content to get utilization rates above 60%1. This is a significant waste, and makes it difficult to reduce large backlogs.

Project Osler will be judged by its improvements in factors that influence the throughput, utilization, and consistent performance of the perioperative department. The changes in these metrics will be analyzed one year after the implementation of its software MVP at participating hospitals.

With patient lives and billions of dollars on the line, reaching these KPIs will be a significant achievement for any accountable perioperative department.


1 2020, Healthcare Financial Management Association

Target Metrics

5-10% Door-To-Door Minutes

25-50% On-Time Starts & Ends

20-50% PACU On-Hold Delays

10-20% Same-Day Cancellations

25-75% Backlog of Surgeries

Secure Local Deployment

The AI models used in Project Osler requires significant computing power. However, public clouds like AWS are often not an option.

Project Osler has identified the Distributed Compute Protocol as an alternative computing environment. By creating scalable, on-site compute networks from existing PCs, DCP will provide hospitals with cloud-like scalability for Osler's models while keeping all data on-premises.

Project Osler's partners may also use DCP for other AI and medical applications. For innovative hospitals, it provides private yet scalable compute that no cloud can rival.

Learn About DCP -->

Roadmap

Project Osler is a multi-stage collaboration between KDS, healthcare providers, and experts in the fields of perioperative management, data science, and computational intelligence. Beginning in Q1 2021, each phase will welcome new partners while expanding the core software solution.

  • Phase I

    • Proof-of-concept AI scheduling algorithm
    • Benchmarking against historical data
    • Deployment in hospital testbeds

    To August, 2021
  • Phase II

    • Full-featured AI scheduling application
    • Realtime forecasting in hospital testbeds
    • Expansion of participating hospitals

    Starts September, 2021
  • Phase III

    • Integration with hospital ERP systems
    • Inter-hospital and regional scheduling
    • Multisite medical compute networks

    Starts May, 2022

Outcome of Phase I

The first phase of the project will provide a web-based diagnostic tool for hospital administrators to determine the impact of different weekly schedules on their OR metrics. This enables OR block times to be optimally assigned to different departments, based on efficient patterns only AI can identify.

As a pilot in data-based scheduling, Phase I will be insightful for hospitals at all stages of their journey toward digital adoption. Its dashboard will work alongside existing workflows, while integrating disparate sources of data that will be required by more advanced healthcare applications.

Join The Project

Project Osler is seeking hospitals or health authorities to become early-stage participants. By joining now, these participants will gain:

Access to software developed in Phase 1


Insights into the development and preliminary trials


The ability to provide input and guide development goals


A license to a DCP private cloud for future healthcare applications