Deployment is King
Even the most accurate algorithm is useless if it can’t be seamlessly implemented into a hospital’s workflow. Here’s a way to make that happen.
By John Halamka, M.D., President, Mayo Clinic Platform and Paul Cerrato, MA, senior research analyst and communications specialist, Mayo Clinic Platform
“And so, castles made of sand fall into the sea, eventually.” Jimi Hendrix’s lyrics are an appropriate metaphor for the failure of many AI-enabled models to have an impact. Developers can create the most sophisticated digital solution to a difficult clinical problem, but it would be little more than a sandcastle if they were unable to find a way for clinicians to easily access it and put it to immediate use at the bedside.
If you have been reading our blog on digital health over the last few years, you’re familiar with many of the innovative AI-driven programs being developed to diagnose atrial fibrillation, prediabetes, colorectal cancer, and many other conditions. But many of these solutions lack an easy way to connect with a health care provider’s ecosystem. For a solution to work within a health system, it needs to conform to specific governance protocols, integrate with its IT systems, and win the support of clinicians and administrators who will be using it on a daily basis.
Mayo Clinic Platform recently developed the capability to integrate a solution seamlessly into clinical workflows named Deploy to accomplish these tasks. With a strong emphasis on data privacy, security, and compliance, Deploy can be used by healthcare organizations seeking to leverage AI advancements while maintaining the highest standards of regulatory adherence. In practical terms, that means bridging the gap between solution developers and healthcare providers by evaluating for intended use, proposed value, and clinical and algorithmic performance and speeding up a model’s integration into the provider’s workflow with pre-built hooks into their EHR.
All solutions undergo a comprehensive qualification process encompassing regulatory compliance, technological, clinical, and algorithmic performance. This ensures clear and transparent communication of each solution's maturity, efficacy, safety, and bias mitigation to potential end-users.
A detailed explanation of the technology behind Deploy is beyond this scope of our review, but briefly, it involves two fundamental pillars, namely: data connection/authentication and end user workflow integration. Data connection/integration removes the complexity of developing new interfaces by facilitating a single place of connection for solution developers through appropriate key and connection provisioning to data sources available through standard terminologies. Workflow Integration enables solution developers and hospital systems to leverage a common integration infrastructure to quickly surface new solutions in workflow through simple, web-enabled tools and capabilities. This process removes the technical complexities of onboarding new solutions and relegates it to a simple approval process handled by the health system. Data remains secure as it fundamentally does not leave the provider’s cloud environment.
Mayo Clinic Platform_Deploy’s technical and security protocols assure healthcare providers that their digital solutions won’t wash away in the sand.
Recent Posts
By John Halamka and Paul Cerrato — Mayo Clinic’s Tapestry Study has demonstrated that next generation genetic analysis can have a significant impact on patient care.
By John Halamka and Paul Cerrato — Imagine if you could create a digital clone of yourself that can be used to test various treatment options to determine which one is best for your real self.
By John Halamka and Paul Cerrato — Data scientists use a variety of coding languages to create AI-driven models, but the real “secret sauce” that helps them identify the best algorithms are the weights the coding generates.