Key factors that need to be considered when deploying AI products at your facility
In recent blogs, we have looked at how AI and other image processing can be deployed in healthcare environments and explored some of the challenges facing the companies that develop these technologies.
But healthcare departments like radiology also need to carefully consider who they partner with when it comes to rolling out new technologies. Today, we look at five key questions that you should be asking when considering the suitability of an AI provider’s products for your facility.
- How does your product fit into my clinical workflow?
It’s important that a vendor understands the imaging ecosystem. In productivity tools and critical care findings, for example, new applications ideally need to be applied to the relevant data before the radiologist reads the exam. Whereas for condition-specific information, such as reports on MS progression, the referring physician may not need a report for several hours. Does the vendor understand these needs and does their technology deliver against these requirements automatically?
Equally, do they require you to use a different viewer? They should understand that outputs should be integrated into a standard reading workflow, most likely in PACS, to avoid breaking workflow by requiring a change to separate viewer or even workstation.
- What else do you do?
Any project to bring on a new vendor will absorb significant hospital resources – the workload placed on hospital IT and PACS administrators alone to deploy a single application at a single facility is considerable. Configuring an application to fit into your standard workflow takes even more time, and that is not even considering the additional resources required by legal, contracting, purchasing, training and project support and monitoring.
Doing all of the above for one niche application just isn’t feasible. But one way to mitigate this investment is to deploy platforms that provide quick access to multiple technologies, which can be easily integrated into existing workflows.
- What will you do with my data?
It’s vital to understand how a vendor uses the data that its AI technology is designed to work with. Personal health information protection is utterly key, and ideally PHI should never leave the hospital IT infrastructure. Of course, it’s pretty clear that you have the right to use a patient’s data for their benefit, but what’s not so clear is who the hospital has the right to share it with to create additional products that might bring benefit to the wider population.
- What is the resource loading?
We’ve already touched on the impact of implementing new technologies on a hospital’s human resources, you also need to explore how a new application will impact on your technical infrastructure. The most important consideration is the performance of the core systems that drive departmental workflow – it’s no use adding a new tool that brings down the PACS or RIS. Considerations around new products should cover internal factors such as processing time, hardware requirements and cloud connectivity – are you happy with the provider’s cloud host? Also consider external factors such as loading on PACS, network connectivity, triggering, caching and so on.
Find out these answers early to prevent a lot of pain later.
- How many sites in the US have purchased your solution?
The only real way to know that all the previous questions above have been answered to satisfaction is to find out who else is using it. And this should not be in terms of trials or tests – but about real customers who have paid for the solution and enjoyed a successful deployment.
Contact us to find out more on how the Blackford Platform can be used to deploy multiple AI, Machine Learning and intelligent image process technologies.