?? Neuralink Implants Device In Human Brain

Good morning. It's official: spring is right around the corner.

At least, according to Punxsutawney Phil, the groundhog who did not see his shadow at the Groundhog Day celebration this morning.

That means, according to the legend, we're in for an early spring!

Until then, let’s dive into the movers and shakers of HealthTech this week!

Enter your email here to receive this newsletter in your inbox directly from HealthTech Zen.

PULSE POINTS

Latest Breakthroughs This Week

Here’s the most important breakthroughs and advancements we spotted in the healthcare technology space this week.

• Neuralink: Elon Musk’s brain science startup, implanted its device into a human for the first time, it teased more details on its product: called Telepathy, it’ll enable people to control “almost any device, just by thinking.”
• Research: Researchers at Georgia Tech just developed a new diagnostic blood test for ovarian cancer using AI and metabolic data that achieved 93% accuracy in detecting the disease. The model calculates a personalized probability score for each patient based on their unique blood compound profile.
• HealthTech AI: More and more wellness companies have been advertising using artificial intelligence in their apps and products. But is AI-powered health tracking making us healthier yet?
• Engineering: SliceGPT introduces a novel post-training sparsification scheme to reduce the resource demands of large language models. By replacing weight matrices with smaller ones and reducing embedding dimensions, it can remove up to 25% of model parameters in major models like LLAMA2-70B and OPT 66B, while retaining up to 99% task performance.
• EHR: EHR interoperability is the capability of different EHR systems to share, interpret, and collaboratively use data across various healthcare environments.
• Arkenea: hire this 14+ years experience custom healthcare software development company to develop or maintain your healthcare applications. Request for a free quote.** from our sponsor.

SNIPPETS

The Latest In HealthTech

• AI Developers: The World Health Organization released new guidelines on the ethics and governance of LLMs in healthcare. Reactions from the leaders of healthcare AI companies have been mainly positive. However, one leader pointed out that fear or the risks associated with LLMs shouldn't hinder innovation, and another noted that the guidance may have failed to mention a major topic.
• Policy: 50-State telehealth report shows slowing reform progress. The report found that telehealth bills filed in 2023 did very little to drive innovation, but instead stuck to the status quo of narrow coverage expansions and payment mandates. However, three states, Idaho, Louisiana, and Utah, stood out for their progressive work in increasing access to care for their state via cross-state legislation.
• Pharma: The New York State Senate passed a package of bills aimed at addressing prescription drug costs. New York pays the third-highest percentage of prescription drug spending out of all 50 states. If signed into law, New York would be the second state after California to manufacture its generic drugs to reduce costs and resolve drug shortages.
• M&A: Cardinal Health, a healthcare product distributor and data analytical services company announced a $1.2 billion cash deal to acquire Specialty Networks, a platform focused on independent physician practices. The former said that the goal is to expand across specialty therapeutic areas and enhance provider offerings.
• Survey: Over 70% of executives are highly focused on data considerations like data availability, quality, compliance, security, and privacy during implementation. Organizations could benefit from filling blind spots like focusing on consumers, governance, and the workforce.
• Stats: Within workforce challenges, 87% of hospital CEOs ranked shortages of technicians as the most pressing issue, followed by shortages of registered nurses (86%), and burnout among non-physician staff (79%). Get the whole story here.

DECODER

What is Digital Therapeutics?

Digital therapeutics are devices and software used to treat a wide array of diseases. They can be used independently or with medications to treat behavioral and physical disorders.

Read more about it here.

DEEP DIVE

AI Outperformed Clinicians in Identifying Social Determinants of Health from EHR Data

What is happening: Large language models outperformed healthcare providers in mining patients’ medical records to find their social needs. Factors for social determinants of health (SDoH) include housing, transportation, financial stability, and community support. Because screening for such factors in clinical notes is so challenging, researchers tested AI tools to see how they compared.

Among 770 cancer patients who received radiation therapy, 48 had an adverse social determinant hidden in their clinical notes. The best-performing model identified 45 of those patients.

Why it matters: Social determinants of health play a crucial role in a patient's health outcomes once they leave the doctor’s office. The screening for these factors is tough and even when it occurs, the findings are scattered across clinical notes from the patient’s visits searching for it and its subsequent utilization for clinical follow-ups is a challenge.

While a subset of billing codes exists to capture SDoH data for quality improvement and health equity efforts, they are sorely underutilized. Healthcare LLMs for identifying SDoH could augment efforts to identify patients who may benefit from social resource support, especially in cases of severe medical conditions such as cancer care where the distance a patient lives from a major medical center or the support they have from a partner can substantially influence outcomes.

The way forward - specialized learning models: Compared to larger models like OpenAI’s GPT-4, the specialized models showed less bias in determining SDoH based on race, ethnicity, or gender. Smaller, specialized models are deemed less prone to bias and more cost-effective compared to larger models like GPT-4.

Data extraction using AI can enhance physician experience and patient outcomes however operational challenges, including clinician alert fatigue, bias mitigation, and obtaining patient consent remain.