Without policies that regulate Artificial Intelligence (AI) and machine learning (ML) there could be dire consequences in every sector of the healthcare industry.
This is what Brian Scorpelli and Sebastian Holst did during the HIMSS22 Global Health Conference in Orlando entitled “A Modest Proposal for AI Control in Healthcare”. Scarpelli is the Senior Global Policy Counsel for the Connected Health Initiative and Qi-fense, a consulting group based in Holst AI and ML.
“ML features do more than challenge domain-specific applications of technology,” Scarpelli and Holst wrote. “Many of these features force the evaluation and retouling of core manufacturing, quality and risk frameworks that serve as the basis for today’s industry-specific regulations and policies.”
Here are some key points from their presentation on AI / ML growth and the need for regulation.
1. AI is very promising.
AI can make revolutionary changes in health care from all angles. It can reduce administrative burdens for providers and payers and allow them to expand resources in the healthcare system to serve the vulnerable patient population. It can handle public health emergencies such as the COVID-19 pandemic and help improve both preventive care and diagnostic efficiency.
2. Machine learning is a fast growing industry.
According to Scarpelli and Holst, there has been an increase in machine learning products since 2015, starting with processing applications including products that process radiological images and then advancing into diagnostic applications, especially with the help of triage and priority in the radiological space. .
The number of patents coded for machine learning and health informatics increased from 165 in 2017 to 1,100 in 2021.
3. Why AI needs control.
While AI is promising, legal and ethical challenges must be addressed. For example, one of the main themes of the HIMSS22 conference was the achievement of health equity and the elimination of implicit bias. This is one of the major challenges of AI, because AI solutions are biased. Tania M. Martin-Mercado, MS, MPH, According to a clinician, several sessions focused on how different teams are needed when designing AI solutions to ensure that programs do not have the same bias as society. Presented by the researcher on “How latent bias affects AI in health care”.
During her presentation, she exemplified that “an online tool for assessing breast cancer risk is lower for black or Latin women than white, even though each other’s risk factor is the same.”
4. Regulators need guidance.
A wide variety of health agencies, including the FDA, HHS, CMS, FTC and the World Health Organization, are seeking guidelines and developing regulations from a variety of stakeholders, including AI developers, physicians and other providers, patients, medical associations and academics. Companies.
5. What should AI strive for.
Scarpelli said the focus for successful AI follows four key elements. It must:
- Improve access to health care.
- Empower patients to manage their own health.
- Strengthen the relationship that patients have with their health care teams.
- Reduce administrative and cognitive burdens on physicians and patients.
This article actually appeared on the MedicalEconomics.com website