🚨 Pulse of Innovation 🚨

Breaking news in the healthcare AI

Improving Accessibility with META AI Smart Glasses

Meta introduced its Ray-Ban smart glasses in 2021, blending fashion with functionality. Initially designed for photo and video capture, the glasses have since evolved into a powerful AI-enhanced wearable. Recent updates enable real-time features like live transcription, contextual reminders, and object recognition, significantly expanding their utility.

Some key applications in healthcare include:

1.  Accessibility for Low-Vision Users: Partnering with “Be My Eyes,” the glasses enable live video assistance, connecting users to volunteers who guide them in tasks like identifying surroundings or reading mail. This innovation enables users to identify objects, navigate spaces, and manage tasks, empowering independence in daily life.

2.  Language Translation and Live Transcription: Real-time speech translation and transcription capabilities make these glasses invaluable in clinical settings with diverse patient populations, breaking down language barriers during consultations or emergencies.

3. Contextual Assistance for Caregivers: Healthcare professionals and caregivers can leverage AI-driven reminders and object recognition to manage patient care efficiently, such as tracking medications or recording observations hands-free.

4. Telemedicine Support: With real-time video and AI assistance, the glasses could be used to guide remote diagnostics, enabling specialists to assist onsite clinicians or patients in underserved areas.

5. Augmented reality (AR) overlays could assist surgeons during procedures by displaying real-time data or anatomical guides.

• Looking ahead, Meta’s Orion project is set to take AR glasses to the next level with always-on AI, seamless interconnectivity, and enhanced medical applications, including virtual simulations for training.

• As wearable AI technology advances, it holds the promise of reshaping diagnostics, patient monitoring, and global accessibility, marking a significant step forward in the convergence of medicine and technology.

Nature Article

🧑🏼‍🔬 Bench to Bedside👨🏽‍🔬

Developments in healthcare AI research and innovations

Is AI Delivering on Improving Physician Efficiency and Reducing Burn Out?

AI tools have variety of impact in healthcare field from improving diagnostics to analytics to note taking which shall ultimately be reflected by improving physician efficiency and reducing burnout. A recent study from China explores how AI in radiology is impact physician efficiency.

• The study included over 6500 radiologist across China and divided in two groups of those using AI and those who did not. The prevalence of burnout was significantly higher in the AI group compared with the non-AI group (40.9% vs 38.6%; P < .001). After adjusting for covariates, AI use was significantly associated with increased odds of burnout (odds ratio [OR], 1.20; 95% CI, 1.10-1.30), primarily driven by its association with EE (OR, 1.21; 95% CI, 1.10-1.34). A dose-response association was observed between the frequency of AI use and burnout (P for trend < .001). The associations were more pronounced among radiologists with high workload and lower AI acceptance. A significant negative interaction was noted between high AI acceptance and AI use.

• Contrary to general belief this study highlights that use of AI use in radiology is increasing physician burnout. One of the most common use of AI in healthcare is in field of radiology as it involves automated image processing AI tools. We have covered several articles describing effectiveness of AI in radiology particularly reducing diagnosis times and improving physician efficiency. These findings covering over 6500 radiologist are insightful and prompts rethinking.

• The study does not go in details of radiologist burnout reasons however highlights low acceptance/adaptability of AI as one of the factors associated higher burnout. More specific reasons of burnout could help with better implementation of these AI tools really improve radiologist efficiency.

   

  

JAMA Article

Are Ambient AI Clinical Documentation Tools Efficient: NEJM DAX Copilot Study

• AI-driven tools are being used for clinical documentation by leveraging advanced machine learning (ML) algorithms and natural language processing (NLP). These technologies automate the summarization of clinician-patient interactions, reducing administrative burdens and enhancing efficiency in healthcare. One notable innovation in this space is Nuance’s Dragon Ambient eXperience (DAX) Copilot, an AI-enabled scribe integrated with electronic health record (EHR) systems.

Study Insights from Atrium Health :

• A study published in NEJM AI evaluated the impact of DAX on clinician workflows at Atrium Health. The study involved 238 clinicians. Participants used the Dragon Medical One (DMO) version of DAX, capturing conversations via the PowerMic Mobile DMO/DAX app on their smartphones.

• The study focused on two primary outcomes:

  1. EHR Use Metrics: Time spent on EHR tasks outside work hours, time dedicated to notes, same-day note closure rates, and note length.

  2. Financial Metrics: Gross revenue per visit and work relative value units (wRVUs) per visit.

  Findings:

Although no statistically significant differences were observed between intervention and control groups after adjusting for variables , some trends were noteworthy:

1. Time Savings: 18% of participants reported over an hour of reduced daily EHR time, repurposing it for other tasks or enabling earlier work completion, the free time was directed towards other activities or leisure which was probably not captured by measuring objective metrics like EHR or financial metrics.

2. Real-Time Efficiency: DAX demonstrated its capability to accurately transcribe multi-speaker conversations and accommodate linguistic variations, enabling rapid documentation - within 30 seconds.

Limitations and Implications:

• The study was not a randomised controlled trial and faced potential unmeasured confounders. Additionally, comprehensive editing times were not evaluated, and initial inefficiencies due to the learning curve may have influenced outcomes.

• Despite these limitations, DAX showcases the potential of AI-powered tools to alleviate the documentation burden, reduce clinician burnout, and streamline workflows.

• Proper training and implementation are critical to fully realising these benefits. Future research should explore subgroup-specific impacts, cost-benefit analyses, and the influence of such tools on the patient-clinician relationship to ensure equitable and effective adoption.

  NEJM AI Article

AI: The Great Health Equalizer?

• The promise of AI in healthcare lies in its potential to break down barriers, making diagnosis, treatment, and research accessible to all, regardless of geography or resources. AI-powered diagnostic tools can deliver accurate results in remote, underserved areas, bridging gaps in healthcare access. For example, portable AI-driven imaging devices are already transforming care in rural clinics.

• In research, tools like AlphaFold enabled protein structure predictions once confined to elite labs to be available to anyone with internet access, empowering scientists in resource-limited settings to innovate for their communities.

• True equity in healthcare hinges on democratising AI itself which is achievable by decentralising power. An article published in the Future healthcare journal elaborates on community driven approaches for achieving this.

  Key concepts:

• The democratisation of AI involves co-creation and participatory design, where diverse communities collaborate, bringing unique perspectives to shape equitable solutions. Evaluating and addressing biases inherent in these collaborations ensures inclusivity and better outcomes. For instance, reflective co-design processes challenge AI solutionism, focusing instead on creating tools that genuinely serve underserved populations.

• Technological democratisation includes using open-source, reproducible systems and decentralised training models like federated learning. These allow institutions across borders to collaboratively train and validate AI without compromising sensitive patient data. Tools like AI “passports” that document model development, testing, and validation are key to ensuring transparency and accountability.

• Moreover, policy and governance play pivotal roles in incentivizing diverse perspectives and supporting community-driven projects. Investments in collaborative frameworks not only safeguard data but also ensure that AI systems are rigorously tested before integration into clinical practice.

 Conclusion:

• From portable diagnostic tools in remote areas to global access to breakthroughs like AlphaFold, AI is breaking barriers. However, its true allure lies in fostering inclusive development and regulation, making healthcare a shared global resource rather than a privilege for a few.

  

  Source: Science Direct