Read the original article:
Becker's Payer Issues | Payer News
Our healthcare system badly lags other industrialized nations in terms of what we spend and the outcomes we achieve, and the culprit is fairly obvious: we spend five times our peer nations on administrative expenses. How bad is it? Our doctors still rely heavily on faxes — yes, faxes — for communications.
Killing fax machines won't solve the problem, of course. We need nothing short of a healthtech moonshot, one that's driven by the technology most capable of erasing administrative inefficiencies: artificial intelligence.
We're also acutely aware of the flaws in many of the industry's existing AI platforms, including those that perpetuate racial and ethnic biases. That's one reason why we're calling for a stepwise approach to this AI-driven healthtech moonshot.
And with AI as a focus of the next Trump administration — and one of his key advisors, Elon Musk — we propose the following blueprint for this technology-driven effort to reform our health system:
Step 1: Ethics
Unlike in other industries, ethics in healthcare cannot be an afterthought; it must serve as a north star by which we can navigate at every step.
And thanks to important research efforts and media coverage, the nation has grown more acutely aware of biases in prediction algorithms generally, and AI algorithms in particular. The work of addressing these biases must be more driven by the federal government, given its considerable control over research and care-delivery spending.
So to start, we must task the NIH to develop an AI ethics framework, in partnership with industry, by the end of 2025. Such a framework would address concerns about bias, accountability, and transparency. Beyond this, the NIH must, on an ongoing basis, gather ethics-related input from industry and conduct AI assessments across the market to ensure these technologies follow ethical guidelines.
Step 2: True Data Interoperability
In parallel with the ongoing work of safeguarding AI-related ethics, we must drive toward true data interoperability. The 2009 HITECH Act pushed health systems to digitize records, but they failed to make those records easily accessible to industry partners and patients. We must now require health systems and health plans to enable full data interoperability by 2028 or forfeit whatever federal funding they might enjoy. Here's how we propose tackling the work:
Phase 1: By the end of next year, we must establish interoperability standards, and develop a framework to guide data sharing across different health systems, thereby ensuring compatibility and efficiency.
Phase 2: By the end of 2026, we must offer meaningful financial incentives to spur health systems to achieve data interoperability benchmarks.
Phase 3: By the end of 2027 we must launch pilot programs in which select health systems can test their interoperability and refine their processes. The government must offer ongoing support for this process, continuously adjusting standards as the implementation continues.
Step 3: Ensure that all Data is Linked to an Encrypted Universal Patient Identifier (UPI)
As other developed countries have proven, UPIs are crucial to helping health systems support patients and their caregivers as patients move among cities, states and different provider groups.
Here's how we propose achieving this important goal:
Phase 1: By the end of 2026, develop encryption technologies to secure patient data and identifiers and ensure confidentiality.
Phase 2: By the end of 2028, integrate UPI technologies into existing electronic health record platforms, with government agencies working closely with health systems to ensure seamless access to data across all stakeholders — but most importantly, patients. This effort will require annual audits and updates to security measures and consent protocols.
Step 4: Build a Public Utility for NIH and Pharmaceutical Researchers
The nation's Cancer Moonshot, although well-intentioned, was doomed to underperform because it lacked the right foundational infrastructure. We can position an AI-driven HealthTech Moonshot for success by linking genetic data with clinical data by 2031 — all with government support and oversight.
Phase 1: This work should begin with the creation of a centralized data repository by the end of 2028 — one that houses both genomic and clinical data, and makes this data accessible for research.
Phase 2: Starting in 2029, the federal government should fund partnerships between NIH and private industry to help researchers mine the data for scientific breakthroughs.
Step 5: Build a National Institute of Artificial Intelligence
First, we must by 2030 identify key health challenges that have received disproportionately little attention from the private-sector, and could be well served by AI-driven research. Next, we must establish the National Institute of Artificial Intelligence by 2033 and seed research initiatives in those areas.
In sum, it will take a massive and concerted effort to deliver the U.S. to a place where its healthcare spending yields what it should for patients. Thanks to the recent and enormous advances in generative AI, however, we can do precisely that.
Yes, it will require a concerted and collective effort. But the benefits will be just as tremendous — not just for this generation of Americans, but for many to come.
Sachin Jain, MD, is president and CEO of SCAN Group. Amber Nigam is the co-founder and CEO of basys.ai.
コメント