Concussions affect almost 50 million persons worldwide; with an estimated 3 million sports related concussions occurring in the US. The economic impact is estimated to be $800-$1500 per concussion; with the hidden expenses being far higher. Long-term healthcare costs, lost wages, and potential academic impacts for student-athletes contribute to an estimated total annual cost in the billions. Moreover, healthcare providers and insurers are increasingly focused on addressing these costs as part of broader sports safety initiatives. The lingering symptoms of post concussive syndrome contribute significantly to these costs. Care delivery for concussion patients is challenging; partially because of ineffective provider remuneration models. The GAIN-Lite protocol was recently announced providing a novel model for concussion followups using both traditional (enhanced usual care) and virtual care. The challenges patients face are fairly pragmatic - provider access, consistency and counseling. The challenges facing practices (esp in the United States) is delivering this high value care in a system with a poor reimbursement model. #concussion #tbi #digitalhealth
About us
Digital health tools designed to improve patient access and care coordination, built around a world-class network of orthopedic specialists. We believe the best orthopedic care is delivered by doctors, not digital apps.
- Website
-
https://theoshealth.com/
External link for Theos health
- Industry
- Hospitals and Health Care
- Company size
- 2-10 employees
- Type
- Privately Held
Updates
-
Google publishes their results w Med-Gemini.
Google recently published their experience w Med-Gemini which is (in plain speak) a highly trained, multi-modal large language model focused on healthcare application. Based on their evaluation criteria, the results are impressive; but most of their benchmarks are "standard" (specific task based) vs. "clinician aligned." Sergei Polevikov, ABD, MBA, MS, MA 🇮🇱🇺🇦 does a nice job summarizing what the Med-Gemini LLM does and doesn't do well. He also raises the point that their ultimate utility must center around clinical relevance for real world applications (such as the synthesis and reporting of multiple streams of complex data). He finishes by noting these tools are best refined in clinical settings by practicing physicians. (Nevermind that Mayo Clinic has never reported their experience using MedPALM-2). My team is preparing to do exactly this for MSK related conditions. We have a ways to go , especially training the LLM, superimposing the RAG and establishing a dedicated QA process. Trust will come from building a track record around real world clinical problems. For the lay-person, these AI tools are presented as quasi-omniscient with the potential to commandeer clinical oversight from physicians. With apologies to Vinod Khosla, their real value will be in helping seasoned clinicians manage large streams of data efficiently, deliver overarching context and hopefully improve diagnostic accuracy in decision making. Not having to scour an EHR to define a clinical narrative would be an added bonus. Worth a read, especially if you don't have the patience to sift through almost 60 pages of a white paper (which is in itself excellent.) Sergei Polevikov, ABD, MBA, MS, MA 🇮🇱🇺🇦 Vivek Natarajan Alan Karthikesalingam MD PhD
Med-Gemini by Google: A Boon for Researchers, A Bane for Doctors
sergeiai.substack.com