A team working with an AI solution

Driving Success in Healthcare AI – Best Practices

To drive successful adoption and deployment of AI technologies, and ultimately desired outcomes, it is critical to have a partnership in which customers and vendors are working collaboratively. Healthcare organizations and vendors alike must share their responsibilities to divide and conquer the work. Through our research we discovered several best practices for both providers and vendors, which we broke down in our recent Healthcare AI 2019 report. I’ve chosen a few to discuss below.

Best Practices for Healthcare Organizations

For healthcare providers and their organizations, the success or failure of an AI tool can really come down to change management and operations.

Embed AI in the workflow

Healthcare organizations frequently face information and tool overload. New tools are frequently perceived as extra hoops to jump through. Done right, AI tools should fit within their users’ regular routines

One CIO recommends spending time observing clinician workflows to make sure AI tools fit: “Any CTO, CIO, or CDO should be sensitive to clinicians’ workloads. I watched what different types of clinicians do on a daily basis with each patient and found places in the workflow where we could insert predictive models in the reports. That way, the tool is part of a normal daily routine and is not going to disrupt that at all. The clinician doesn't have to run another report or print anything out; it’s going to show up inside the EMR with this extra information they didn’t have before.”

Take ownership for driving change management and operationalizing insights

The organization’s leadership needs to take a big role to get their staff engaged. In the report, we didn’t really talk about having a clinician champion to drive the changes in workflow because while this bottom-up approach can be useful, we find that the top-down approach with AI is more effective. If an organization wants to create a successful AI project, the leadership needs to have buy-in from the start so they can socialize the idea across different departments.

Leadership then needs to report back on any progress and successes to the staff members. That lets the staff members fully realize the success they have achieved and further encourages them.

How Vendors Can Help Drive Client Success

For healthcare organizations to be successful with their AI solutions, they need vendors that do much more than deliver a high-quality, technologically capable product. The best practices below were reported by vendors’ most successful AI customers.

Deliver comprehensive services for AI platform

The need for these services is critical when you consider that the adoption cycle of AI in healthcare and the maturity model for clients are still in early stages. Customers cannot simply purchase the tools or technology and be expected to achieve success without proper guidance and support from their vendor. That would be like handing future homeowners blueprints and materials and wishing them luck on building their new houses.

A director of data sciences outlined a vendor’s approach: “We have both a dedicated account manager and a dedicated client-facing data scientist, so it’s basically like we bought the AI tool. We got the software, but then we also got a free data scientist along with it. The vendor came in heavy with the training. They sent us to an all-day course to make sure we were familiar with the tool and all the settings in the tool. We’ve built a model, but it’s not exactly performing the way we want. We just call up the vendor, and they’re right there.”

Vendors that offer comprehensive services are really showing their customers that client success is their ultimate goal. I cannot emphasize enough how important it is to put in place these comprehensive services during the early-adoption phase. For those vendors who haven’t yet implemented comprehensive services, that is definitely something to consider.

Be a humble, active partner

Healthcare organizations frequently face information and tool overload. New tools are frequently perceived as extra hoops to jump through. Done right, AI tools should fit within their users’ regular routines. 

One CIO recommends spending time observing clinician workflows to make sure AI tools fit: “Any CTO, CIO, or CDO should be sensitive to clinicians’ workloads. I watched what different types of clinicians do on a daily basis with each patient and found places in the workflow where we could insert predictive models in the reports. That way, the tool is part of a normal daily routine and is not going to disrupt that at all. The clinician doesn't have to run another report or print anything out; it’s going to show up inside the EMR with this extra information they didn’t have before.” 

Being transparent going into the project will go a long way in terms of creating the needed partnership between vendors and their customers. All AI healthcare vendors are learning as they go in this very young market. In this report KLAS sees that vendors who have good partnerships are those who are responsive to constructive feedback.

The partnership between a vendor and a healthcare organization was described by a CIO: “Our vendor is always ready to work with us. They are listening, they are acting, and they are timely. They are always willing to jump on the next plane to meet with us. It takes a lot to create something. More importantly, when I brought my clinicians to the table, it’s an exact science for them to make them believers. It is kudos to them and the physician that they brought on board, and my team that worked with them of course, but we still have variable results.”

To learn more about the other best practices we found, please read the full report. It has been an incredible journey to watch this market evolve and progress. We at KLAS are excited to see what more AI technology will bring to the healthcare space in the near future.




     Photo cred: Adobe Stock, Gorodenkoff