AI Task Force Develops Unified Strategy to Integrate Tech Innovations

Artificial intelligence (AI) initiatives abound in hospital systems, from speeding drug discovery to streamlining trial identification. With AI becoming ubiquitous in cancer care, there is a need to incorporate these innovations in a responsible way.

To that end, Cleveland Clinic Cancer Institute formalized a strategic approach to evaluating emerging technologies to streamline AI adoption in a safe and deliberate manner. Led by Dr. Paul Thambi and Dr. Louis Williams, the new Cancer Institute Office of Technology and Digital Strategy assesses potential AI technologies to ensure alignment with enterprise strategy.

A plethora of AI companies have approached the Cancer Institute with partnership opportunities. The task force vets these proposals, their potential impact on operations and care delivery and alignment with enterprise goals and vision. With this structure in place, the Cancer Institute will be well positioned to integrate AI models into day-to-day care as well as the enterprise business model.

“Vendors don’t just want to deploy their models but do so in an environment where there’s a high degree of quality as well as robust data sets,” explains Dr. Thambi. “The richness of our clinical data allows AI models to make accurate predictions – and has the potential for many new discoveries.”

Current use cases

AI is already being used in the Cancer Institute to:

• Personalize radiation doses for patients with lung cancer
• Reduce notetaking time by up to 75%, giving clinicians more one-on-one time with patients
• Streamline the processing of patient referrals

The task force serves as an entry point for these initiatives. Instead of evaluating AI proposals on a one-off basis, they’ve developed a comprehensive strategy around the use of AI. “AI will be fundamental to all aspects of running the Cancer Institute in the future,” says Dr. Williams. “That extends from operations to scheduling to clinical decision-making. We need to be able to scale well. This means both reducing barriers to adopting AI and ensuring proper governance structures.”

Analyzing new AI applications and providing safeguards

The enterprise has an AI Task Force that evaluates models and AI technologies that are not research based. The AI Task Force is analogous to an institutional review board, but for non-research-based applications, ensuring safety, monitoring for bias, data security and measures to monitor the models for changes in performance.

“It’s crucial to have a system to continuously re-evaluate the effectiveness of each tool,” says Dr. Williams. “You need to analyze the output of models on an ongoing basis. There can be drift in a model that may influence its accuracy over time or there may be bias in the model that isn’t apparent upfront. That’s why it’s necessary to have a unified strategy that includes training models to identify mistakes before a model goes into production as well as longitudinal risk assessment to ensure safety.”

Improving the patient experience

Ultimately, the hope is that these models will offload certain functions to free up clinicians to focus on personal attention for patients and families facing cancer. “We may have tools to do things like giving a better differential diagnosis but there’s a human element necessary to deliver the news of that diagnosis and support patients through their journey,” says Dr. Thambi. “All of these tools can help usbe more present and less distracted by the computer as we are seeing patients.”