All articles tagged with #health data
WIRED tests Muse Spark, Meta's health-focused AI, which encourages users to paste lab results and biometric data for analysis, but privacy risks loom since data can be stored to train models and used for ads, HIPAA protections don’t apply to consumer AI, and experts warn the tool is not a doctor and can give risky guidance (including an extreme fasting plan).
Garmin is developing a third‑party Garmin Chat Connector to link Garmin Connect data with AI assistants such as ChatGPT and Claude, enabling natural‑language questions about sleep, workouts, and recovery. The system, a cloud‑hosted MCP server, securely connects data via a token‑based URL and exposes 16 data tools across five categories; it is not an official Garmin product and is nearing completion as part of a broader move toward conversational health analytics.
Ultrahuman unveils Ring Pro, its third-gen smart ring, with up to 15 days of battery life and 250 days of on-device health history, plus a Pro Charging Case that adds ~45 more days and stores a year of data. The ring features a redesigned heart-rate sensor, faster processor, and improved safety features. Preorders are $479, US availability is blocked due to Oura patent issues, and shipments start in March. The Jade real-time biointelligence AI platform will roll out globally, offering health insights and, in future updates, smart-home controls and services.
The article outlines five practical settings you can adjust on your Apple Watch to improve heart-rate measurements, emphasizing a snug wrist fit and proper calibration so readings are more reliable during workouts and everyday wear.
The author tested OpenAI's ChatGPT Health and Anthropic's Claude by uploading a decade of Apple Health data (and later medical records) to obtain a simple cardiac-health grade. Outputs swung widely—ChatGPT gave an F, then a D after adding records; Claude varied between C and B—driven by noisy VO2 max estimates and changing resting heart-rate data. Experts say the analyses are not medical advice, and privacy promises to protect data aren’t HIPAA-covered. While AI health tools could unlock insights, the article argues they’re not ready for personal clinical decisions and should mainly support prepping for a doctor visit.
OpenAI has agreed to acquire Torch, a one-year-old AI healthcare app, for about $100 million in equity (roughly $60 million upfront with the rest in stock to retain employees), a move that extends OpenAI’s reach into health-data analysis from providers like Kaiser Permanente to apps like Apple Health and firms such as Prenuvo.
Fitbit's app is being redesigned with a new AI-powered Personal Health Coach, built with Google's Gemini, offering personalized health and fitness guidance, better data visualization, and customization features, available in an open beta for Fitbit Premium subscribers, with a focus on scientific research and partnerships to ensure reliable health advice.
The UK Biobank completed the world's largest whole-body imaging project, scanning 100,000 volunteers over 15 years, creating an extensive dataset that is revolutionizing disease research, diagnostics, and AI development for early detection and understanding of health and aging.
RFK Jr. suggests all Americans use health-tracking wearables to monitor their health, but experts raise concerns about data security, privacy, and government access, emphasizing the risks of data breaches and misuse.
Washington and 19 other states sued the federal government over the decision to share Medicaid enrollees' personal health data with ICE, raising concerns about privacy and deportation efforts.
The article criticizes the current state of AI-generated fitness summaries from wearables, highlighting their superficiality, lack of context, and limited usefulness, especially in complex scenarios like injuries. Despite widespread adoption and positive user feedback, the summaries often fail to provide meaningful insights or actionable advice, revealing the limitations of AI in understanding nuanced health data. The author suggests that these features are more about marketing than delivering real value, and hopes for more personalized, insightful AI in the future.
The article reflects on the collapse of 23andMe, emphasizing lessons for the future of consumer genomics, including the importance of clinician partnerships, the limitations of HIPAA compliance, and the need for better governance and transparency of health data. It highlights the company's strategic missteps and the broader implications for health data ownership, trust, and regulation in the evolving digital health landscape.
The increasing availability of fitness trackers and wearables has provided users with an overwhelming amount of health data, but experts caution that more data does not necessarily lead to better health outcomes. While these devices offer a range of measurements from heart rate to sleep time, not all data is considered helpful, and the real impact depends on the individual wearer. Additionally, many of these devices are not regulated by the FDA, and some claims about their capabilities have been cautioned against.
Researchers in Denmark have developed an AI "death calculator" called life2vec, which analyzes personal and medical data to predict a person's mortality timeline with 79% accuracy. Using a Danish public health registry dataset, the algorithm outperformed other tools at predicting early mortality, suggesting it could be a valuable tool for identifying key factors that affect life outcomes and enabling earlier interventions.
A study published in The Lancet suggests that over 7,000 Covid-related hospital admissions and deaths in the UK during the summer of 2022 could have been prevented if the population had received the recommended number of vaccine doses. The research, which analyzed anonymized electronic health data for the entire UK population over the age of five, found that 44% of the population was under-vaccinated, with younger people being among the most likely to skip doses. The study highlights the importance of vaccine uptake in preventing severe Covid illness and suggests that similar health data analysis could be used to understand and address other diseases.