The assessment of AI technology as part of AI management is crucial in the healthcare environment, says many experts. The tools provided by the supplier must be unthreatening, effective and ethical to ensure the best results for all patients.
As part of the management structure, there should be an effective system of understanding and risk assessment related to AI in clinical exploit. This approach defines four risk categories and an assessment system to assess risk in each category.
These categories are:
- Correctness and transparency
- Honesty and justice
- Integrated work flow
- Security and privacy
Emphasizing the risk areas, the healthcare system can make more conscious decisions that tools for exploit, and technology implementation teams can develop a risk relief suitable for tools or work flows.
Establishing a solid management framework in interaction with suppliers will allow healthcare organizations to exploit the benefits of commercial AI systems while minimizing the risk and maintaining public trust, said Glenn Wasson, administrator of analytics at UVa Health. Has a doctorate in computer science.
Wasson will deal with this topic at the Himss25 educational session entitled “Dear suppliers AI: we need it.”
At UVA Health, Wasson oversees the way in which the care system and analyzes the data on patient care and tests. His duties include data operations, analyst, data learning and data visualization.
In this role he is working on everything, from predictive care among residents, to modeling the risk of population to the spilling of the hospital ranking algorithms. He is particularly interested in creating problems solving problems.
We conducted interviews with Wasson to preview his Himss25 session.
Q: What do hospitals and healthcare systems have to understand about AI tools compared to previous software?
AND. AI goes to every aspect of today’s healthcare undertaking, opening the door to the potential improvement of everything, from diagnoses and treatment planning to settlement and resource management to research and discovering medicines.
But AI systems also contain a recent risk that is not present in previous generations of software. Organizations must understand this risk and their potential effects and alleviating to effectively rule artificial intelligence in their company.
Understanding the risk and benefits of AI systems is a sophisticated question. Supplier organizations rarely have access, capacity and/or talent for analyzing the supplier’s code.
Instead, in this session he discusses the dialogue between suppliers and suppliers who can emphasize the sources of risk. This dialog box will require increased transparency of data, algorithms and work flows, but it can also support suppliers develop the supplier’s trust in their tools.
Q: How will you focus on artificial intelligence in the Himss25 session?
AND. This session concerns artificial intelligence, which has many varieties, which is the latest artificial intelligence. As part of healthcare, AI’s ability to analyze evidence based on past human experience allows them to be used in many different scenarios.
They include forecasting diagnosis and selection of treatment, personalized medicine, staff planning, beds of beds, supply chain, remote monitoring, improvement in invoicing and coding, reduction of costs through efficiency and many others.
In this session, we want to discuss various cases of exploit for artificial intelligence and work flows that allow AI and people to cooperate in making decisions. We will discuss how understanding the actions arising from these decisions and the risks related to these activities can build the certainty that the AI system will be unthreatening and correctly provided to the necessary support.
Q: What is one of the different letters with which you hope that the participants of Himss25 will leave the session and will be able to submit an application when they return home to their organizations?
Questions will include various risk sources – data sets, work flows, etc. – and will require qualitative and quantitative answers.
That is why this dialogue is not to be simply a discussion between AI professionals – for example, data scientists or statistics – but they must include leaders and operators who understand the intended environment and work flows.
Finally, we will discuss how the vigorous nature of learning today’s advanced AI means that this dialogue must be pending.