The Importance of Collaboration When Solving Problems With AI
Time and again, clinical users have expressed that when they’re expected to adopt a new technology or process, they’re less likely to do so if the solution was handed to them as mainly an IT directive rather than a collaborative project. The question should not be, “How can we adopt the latest technology?” but rather, “How can we understand a common problem, with all the right context in a clinical environment, that must be solved?” That’s a key part of AI governance: understanding stakeholder pain points, risk and the impact to workforce.
To build trust, clinicians must be involved in governance. If an organization wants to solve a clinical problem such as documentation burden, physicians and nurses should be at the table. IT teams should be embedded with a nurse to understand where the bottlenecks are that can be solved with technology.
In my experience, when organizations don’t have clinicians — especially nurses — involved in governance, there is resistance to AI adoption. A project is then viewed as a cost-cutting measure or a tool for replacement. Those concerns must be addressed immediately. An AI tool supports a clinician in their work, reducing the number of tasks or mental workload so that their energy and time can go toward direct patient care. That must be communicated at the very beginning to establish and build trust that clinician burnout, which has been a longstanding issue in healthcare, is a top priority that organizations want to address.
Organizations have room to grow when it comes to workforce training and communication of AI programs. For instance, through the governance process, leadership should ensure that education and additional help desk support are baked into a solution’s deployment.
RELATED: Clinical workflow redesign drives healthcare optimization.
How Healthcare Organizations Can Measure AI Success
Another piece of governance includes clear measurements for ROI. Has your organization defined how it will measure success, and is that definition understood and clear to all involved? Those measurements could be in better clinical outcomes, reduced workflow complexity or improved data processes.
That’s why the most common AI use cases I’ve seen in healthcare in recent years have been around ambient clinical documentation, back-office administrative automation, and medical imaging and clinical decision-making support. These use cases have measurable outcomes in time saved, tasks reduced or changed, or patients supported.
Even before healthcare organizations can shop around for an AI solution, they must also address data quality. Data readiness is foundational when layering on an AI tool to eliminate inefficient processes and workflows, rather than exacerbating them. Improving data governance will also help to address clinicians’ concerns around trust.
A Strategic Partner Can Help Clinicians Build Trust in AI
AI adoption should not be about having the most solutions possible, which users may not even glance at. It should be about addressing inefficient processes, establishing governance, modernizing data as needed and targeting high-value problems.
At CDW, we often start with listening to nursing and other clinical leadership about the priorities they want to address. Then, through our Transformation Centers, we can help clinical teams understand how a solution — such as an ambient listening tool or a voice-enabled smart assistant — might fit into their environment by trying it in a functional testing room. They can explore how a solution interacts with their workflows and see how a problem can be addressed from a clinical viewpoint first, rather than from an IT perspective. This experience supports clinician-driven implementation.
A strategic partner can help organizations measure success and ensure that there’s agreement on definitions and goals. Change management is also a key area where a partner can provide support, building out training and a network of AI champions that will improve adoption.
This article is part of HealthTech’s MonITor blog series.
