krishnakumar

Founder @evidence_md

[krishnakumar @krishna93677]

Calcium score would be a best option

[ANI @ANI]

#WATCH | Bengaluru, Karnataka: Dr Devi Shetty, Founder and Chairman, Naryana Health says, "Every Indian, at the age of 17 must get the blood test done. This is the guideline now issued by the Cardiology Society of India because if the cholesterol level is high, they can manipulate the diet and all the men at the age of 35 to 40 should undergo a routine test along with CT scan of the heart. Anyone with the family history of heart disease must go for the checkup at the age of 30 itself, not wait for long time. All the diabetics should go for the checkup even earlier than 30 years. It is important that everyone should know their numbers. They should know what their blood pressure is, cholesterol number is, their heart numbers are. Today we find large number of young people getting into extreme sport. Even if you are very young, 17 or 18, if you are going for extreme sport, you must go for a checkup... The best way to know what the cardiac problems are to the test done. As I said, simple tests which can be done within one hour, once a year. If the CT scan is normal, for the next seven years or ten years you don't need to bother..."

[EvidenceMD @evidence_md]

Introducing the EvidenceMD API: Clinical Reasoning for Modern Healthcare

As healthcare becomes increasingly complex, access to transparent, evidence-based clinical reasoning is more important than ever. The EvidenceMD API enables healthcare organizations, digital health platforms, and developers to incorporate clinical reasoning capabilities directly into their products and workflows.

Designed for modern healthcare applications, the API helps transform medical knowledge and evidence into structured, context-aware reasoning that can support a wide range of clinical and operational use cases.

https://t.co/Lhh5YrEgR4

Key capabilities include:

• Evidence-based clinical reasoning
• Seamless integration with healthcare applications and workflows
• Scalable infrastructure for enterprise environments
• Secure and reliable access to clinical intelligence
• Structured outputs built for interoperability and innovation
Whether you’re building provider tools, AI-powered healthcare solutions, care management platforms, or next-generation health technologies, the EvidenceMD API provides a foundation for bringing clinical reasoning into modern healthcare experiences.

Built for healthcare teams. Powered by evidence. Designed for modern clinical reasoning.
#physicians #doctors #healthtech #medtwitter #healthcaredevelopers #developers #aihealth #ClinicalReasoning #ModernHealthcare #HealthcareAPI #DigitalHealth #HealthTech #EvidenceBasedMedicine #HealthcareInnovation #MedicalTechnology #Interoperability #AIinHealthcare

[EvidenceMD @evidence_md]

A recent study published in Nature Medicine offers valuable insights into the capabilities of large language models in clinical contexts. While frontier general-purpose models like GPT, Claude, and Gemini demonstrate strong baseline performance, the research highlights a critical distinction: answering complex medical questions requires more than standard literature retrieval.

This is the exact challenge EvidenceMD was built to solve.

Rather than functioning primarily as a journal retrieval tool, EvidenceMD is a deeply medically fine-tuned model designed for complex clinical reasoning. Our focus is on:

🔹 Transparent Reasoning: Providing clear, step-by-step clinical logic rather than a "black box" output.

🔹 Complex Question Answering: Synthesizing nuanced medical information to answer intricate clinical queries.

🔹 Deep Medical Fine-Tuning: Moving beyond surface-level retrieval to deliver true clinical understanding.
As the landscape of medical AI evolves, the focus must shift from simple information retrieval to transparent, medically grounded reasoning.

Read the full study here:

https://t.co/2H5KbvaRcF

#ClinicalAI #physicans #NatureMedicine #HealthTech #MedicalInformatics #doctors #medtwitter #medED #medLLM #LLM #clinicalreasoning

[krishnakumar @krishna93677]

A new Nature npj Digital Medicine paper on an autonomous AI agent for Emergency Department decision support highlights where healthcare AI is heading.

https://t.co/uPaHGSS111

For years, the focus has been on documentation and note generation.

But the bigger opportunity is helping clinicians navigate complexity—reviewing evidence, synthesizing patient data, and supporting clinical reasoning.

At EvidenceMD, we believe AI should do more than document care.

It should help clinicians make sense of information and arrive at better-informed decisions.

Documentation saves time.

Clinical intelligence creates value.

#physicians #MedicalAI #ClinicalReasoning #HealthcareAI #DigitalHealth #doctors #clinicalAI

[krishnakumar @krishna93677]

NEJM AI made an important observation:

https://t.co/yLcbl5nt7u

AI scribes reduce documentation burden, but they are not true productivity tools—yet.

Most solutions today focus on:
✅ Transcription
✅ Note generation
But physicians spend significant time on:
• Clinical reasoning
• Chart review
• Evidence retrieval
• Differential diagnosis
• Documentation refinement

At @evidence_md , we believe AI should do more than write notes.

Our vision is to build a clinical reasoning assistant that helps physicians analyze cases, review evidence, generate differentials, and create high-quality documentation.
Documentation is the starting point. Clinical intelligence is the future.

#HealthcareAI #ClinicalReasoning #MedicalAI #DigitalHealth #EvidenceBasedMedicine #LLM #medical #physicians #doctors

[krishnakumar @krishna93677]

One of the biggest problems with medical AI today is that the models are trying to know everything. When a model is trained across thousands of topics, medicine becomes just another subject. The result is confident answers that can sometimes be wrong. In healthcare, a hallucination is not a small mistake. It can misinform patients, influence clinical decisions, and create real world consequences. Medical AI should be built on evidence, clinical reasoning, and transparency. Every answer should be traceable, verifiable, and held to a higher standard.

The future of healthcare AI is not about bigger models. It’s about building systems that clinicians and patients can trust.

#medical #clinicalai #aihealth #ai #medtwitter #meded #doctors #physicians

[د. عبدالله البيروتي @aalbeirutty]

لكل طبيب وطالب طب..

لا تستخدم ChatGPT
لا تستخدم Claude
لا تستخدم Gemini

أخطاؤهم كبيرة ، وهلوستهم كثيرة ..

هنا جبت لك 6 منصات ذكاء اصطناعي طبية تقدر تستخدمها وانت واثق 🩺

[krishnakumar @krishna93677]

@TataMotors_Cars Thank you and DMed you

[krishnakumar @krishna93677]

Hi @TataMotors_Cars,

Facing a serious issue with my Tata Nexon AMT at only ~26k km. During driving, the car suddenly became sluggish and gear shifts became very slow/delayed. Restarting the car temporarily fixed the issue, but it returned again later.

Service center is now saying it is a clutch issue and asking for customer payment. Considering the low kilometers and intermittent AMT behavior, I am concerned this may involve AMT actuator/calibration/transmission control issues rather than normal wear and tear alone.

Requesting escalation and proper diagnostic review please. This also created a safety concern during driving.

Can someone from Tata Motors please assist?

[krishnakumar @krishna93677]

@AzureSupport DMed you

[krishnakumar @krishna93677]

@AzureSupport - I have my ticket opened for last 10 days and no proper support. Could you please help here?

[krishnakumar @krishna93677]

@pankaj Such a scam run. You raise 33 M just to integrate 10 api’s and give model for free - Is this a business model? Very sorry for investors.

[krishnakumar @krishna93677]

Healthcare does not suffer from a lack of information. It suffers from a lack of structured, reliable reasoning over that information. Simply connecting a system to PubMed or large volumes of medical literature does not create evidence based care. Retrieval alone cannot distinguish weak studies from strong ones, resolve conflicting results, or apply findings correctly to an individual patient.

What the market truly needs are strong foundational medical models built specifically for clinical reasoning. These models must understand study quality, hierarchy of evidence, bias, and context. They should be able to synthesize across trials, guidelines, and real world data, and translate that into patient specific insights rather than generic summaries.

Evidence in medicine is not just about access to papers. It is about interpreting that evidence correctly, weighing it appropriately, and applying it safely. Without this layer of reasoning, systems risk amplifying noise instead of reducing it.

Patients also increasingly expect transparency and justification for decisions that affect their health. Evidence based reasoning is not only a clinical necessity but also a trust requirement.

Foundational medical models that embed this kind of structured reasoning can reduce variability, minimize errors, and bring consistency to care. This is the shift from information retrieval to true clinical intelligence.

#clinicalai #medical #doctors #doctor #physicians #healthai #medllm #med #medtwitter #medical #health #LLM

[krishnakumar @krishna93677]

Foundational models are setting a new standard for AI in medicine.

General models are trained broadly. They perform well on language, but medical decision-making requires structured understanding, consistency, and domain-specific reasoning.

Foundational medical models are built with that context at the core.

They are trained on large-scale, domain-relevant data and learn rich representations of medical knowledge, allowing them to handle complexity beyond surface-level text processing

Here’s where the difference becomes clear:

Domain-aligned understanding:

They capture how clinical variables interact, not just how they are described

Stronger generalization:

They perform more consistently across real-world scenarios because they are trained on medical distributions, not generic internet data

Multi-task capability:

A single model can handle multiple workflows with minimal adaptation, reducing the need for fragmented systems (Nature)

Deeper reasoning:

They can identify patterns, weigh competing possibilities, and support complex decision pathways rather than simple retrieval

Scalable improvement:

Performance improves predictably with more data and scale, making them long-term infrastructure rather than short-term solutions

The shift is simple:

Foundational medical models reason within context.

That difference is what makes them far more reliable for real-world medical workflows.

#medicalai #doctors #medicalLLM #LLM #healthcare #ai #reasoningmodel #physicians #medtwitter

[krishnakumar @krishna93677]

Most people are underestimating this.

Medical AI is not just another LLM use case. It needs deeply specialized fine tuning trained on real clinical reasoning, patient variability, guidelines, and evidence synthesis. Without that, you are not building something reliable, you are just building a system that sounds confident while being wrong.

At the same time, the market is getting flooded with wrappers and agent based workflows. They look impressive in demos and feel powerful on the surface, but when you look closely there is very little depth. No real clinical understanding, no strong reasoning, and no defensibility.

Agent workflows are not the moat. Orchestrating APIs is becoming easy. Building something that can actually reason like a clinician is the hard part.

What will really matter is depth. Models that are tuned for the domain, systems that can search and ground answers in real evidence, and the ability to understand a patient over time instead of giving one off responses.

Wrappers might win attention for now. But in healthcare, depth is what will win trust.

#clinicalai #healthcare #medicalai #clinical #medLLM #LLM #health #doctors #medtwitter #med

[EvidenceMD @evidence_md]

Introducing a new standard for Evidence based AI in healthcare:

Transparent First Reasoning.

We built EvidenceMD because complex clinical questions require more than just a quick output. They require verifiable logic. EvidenceMD is an advanced clinical reasoning model that doesn't just give you the answer; it shows you its work.

Unprecedented Transparency: Expand the "Thought Process" window to see exactly how the AI evaluates symptoms, labs, and medications.

Dynamic Visualizations: Complex differentials and biomarker comparisons are instantly transformed into clear, readable tables and visual aids.

Deep Context Memory: Retains highly complex patient data to provide accurate, multilayered, and evidence based clinical insights.

#ClinicalAI #HealthcareInnovation #ClinicalReasoning #MedicalDocumentation #HealthcareSupport #FutureOfHealthcare #healthtech #healthai #medicalLLM #Clinicalreasoning #physicians #doctors #nurses #peerreviewed #biomed #medicalresearchers #medicalintelligence #clinicalcopolit #ClinicalAI #MedicalAI #ChainOfThought #ClinicalDecisionSupport #HealthcareInnovation #aihealth #medicalLLM #medLLM #physicians #medical #medicalstudents #CME #MedicalAI #clinicalreasoning #biomedical #AIinhealthcare #medicaleducation #clinicalAI #cmecredits #CE #CPD #clinicalleadership #biomedicalresearchers #medicalresearchers #healthcareworkers #nurse #healthcareworkers #medicalscribe #aiscribe #medtwitter

[krishnakumar @krishna93677]

Healthcare needs better SOTA model

[EvidenceMD @evidence_md]

Most AI models in healthcare focus on documentation, summarization, or answering medical questions. However, real-world clinical workflows require something more complex: clinical reasoning combined with operational healthcare intelligence.

At EvidenceMD, we developed a state-of-the-art reasoning model designed specifically for clinical environments by integrating two critical capabilities into a single system.

1️⃣ Advanced Clinical Reasoning

The model processes structured and unstructured medical data including:

• Laboratory results
• Imaging reports
• Patient history and symptoms
• Medications and comorbidities
• Clinical notes

Using evidence-based medical knowledge, the system performs multi-step reasoning similar to physician diagnostic workflows, generating structured clinical interpretations, differential considerations, and guideline-aware insights.

This enables hospitals and healthcare professionals to analyze complex patient data faster while maintaining strong evidence alignment.

2️⃣ Insurance & Documentation Intelligence

Healthcare delivery also depends heavily on documentation quality, medical necessity validation, and payer requirements.

EvidenceMD incorporates an insurance intelligence layer that understands:

• Medical necessity criteria
• Documentation completeness
• Coding context and clinical justification
• Payer-driven decision pathways

This allows hospitals and healthcare teams to align clinical documentation with reimbursement requirements, reducing friction between care delivery and administrative workflows.

Why this matters

By combining clinical reasoning + insurance intelligence, EvidenceMD bridges two critical layers of healthcare operations:

• Clinical decision support
• Documentation quality
• Revenue cycle alignment
• Operational efficiency in hospital systems

The result is a reasoning system built not just to understand medicine, but to operate within real clinical and healthcare infrastructure.

#ClinicalAI #HealthcareInnovation #ClinicalReasoning #MedicalDocumentation #HealthcareSupport #FutureOfHealthcare #healthtech #healthai #medicalLLM #Clinicalreasoning #physicians #doctors #nurses #peerreviewed #biomed #medicalresearchers #medicalintelligence #clinicalcopolit #ClinicalAI #MedicalAI #ChainOfThought #ClinicalDecisionSupport #HealthcareInnovation #aihealth #medicalLLM #medLLM #physicians #medical #medicalstudents #CME #MedicalAI #clinicalreasoning #biomedical #AIinhealthcare #medicaleducation #clinicalAI #cmecredits #CE #CPD #clinicalleadership #biomedicalresearchers #medicalresearchers #healthcareworkers #nurse #healthcareworkers #medicalscribe #aiscribe #medtwitter

[EvidenceMD @evidence_md]

Most AI models today are general-purpose LLMs trained on massive internet-scale text corpora. They perform well across many domains, but medicine is fundamentally different. Healthcare requires deep medical knowledge, clinical reasoning, and interpretation of structured medical data. A model trained primarily on general web data is not optimized for this level of complexity.

Recent medical AI benchmarks such as MedQA, PubMedQA, MedMCQA, and MMLU-Medicine highlight this gap. Models adapted with medical literature and clinical datasets consistently outperform general LLMs on clinical reasoning tasks and safety-critical medical evaluations.

The next evolution goes even further.

Medicine is inherently multimodal. Clinical reasoning involves interpreting medical knowledge together with structured clinical data. Models designed to process these signals together are significantly better aligned with real clinical workflows.

This is where specialized medical AI becomes critical.

EvidenceMD is a clinically fine-tuned multimodal medical AI designed for physicians and researchers.
Instead of acting as a general chatbot, EvidenceMD is designed to understand medical knowledge, clinical context, and complex medical questions, while synthesizing evidence from medical literature.

The goal is not just answering questions.

The goal is clinical intelligence aligned with evidence-based medicine.

The future of healthcare AI will not rely on general LLMs alone.
It will rely on clinically specialized, multimodal medical intelligence.

#ClinicalAI #HealthcareInnovation #ClinicalReasoning #MedicalDocumentation #HealthcareSupport #FutureOfHealthcare #healthtech #healthai #medicalLLM #Clinicalreasoning #physicians #doctors #nurses #peerreviewed #biomed #medicalresearchers #medicalintelligence #clinicalcopolit #ClinicalAI #MedicalAI #ChainOfThought #ClinicalDecisionSupport #HealthcareInnovation #aihealth #medicalLLM #medLLM #physicians #medical #medicalstudents #CME #MedicalAI #clinicalreasoning #biomedical #AIinhealthcare #medicaleducation #clinicalAI #cmecredits #CE #CPD #clinicalleadership #biomedicalresearchers #medicalresearchers #healthcareworkers #nurse #healthcareworkers #medicalscribe #aiscribe #medtwitter

[krishnakumar @krishna93677]

@Airtel_Presence @airtelindia Here is my service request number # 10853496529 - Wi-Fi not working for 4 days for no reason. Is this how you service your customers?

[krishnakumar @krishna93677]

@airtelindia here is my request ID: 654439745 -My DTH has not been working for the past 3 days, and my earlier complaint was closed without proper resolution. A technician visited and changed something, but now the service is not working at all. Is this how you treat customers?