Field Journal

India Case Study.

Logging progress, patient interactions, and real details from the deployment of Vivral's Patient Passport in rural India.

Visual Diagnostics & Severe Hospital Overcrowding

Today our patient went for another checkup. Before heading out, we noticed these black spots on his arm. We used Vivral's vision features, and it pointed out they might be blood clots. After physician review later, they indeed were.

When we went to the hospital, where his doctor is, the doctor didn't show up for hours, and there were 32 patients in front of him. Out of frustration and out of tiredness, our patient chose to go home and rest. He rescheduled for a day where the queue was certain to be smaller. This just highlights how poor the healthcare system is over here. This is the best hospital within a reasonable radius and they are extremely overworked.

We are really glad, however, that Vivral was able to identify the blood clot. We do certainly need more RL as the response also suggested other not-great, scary reasons. V2, which is in the pretraining stage, should solve most of these over-diagnosis issues.

Systemic Overload Top regional hospitals are operating vastly beyond capacity, forcing exhausted patients to abandon care.

Model Refinement Vivral correctly identified the clot, but V2 will address the tendency to over-diagnose unrelated conditions.

Camera-Based Vitals & Pacemaker Monitoring

A family member of a patient here needs a pacemaker, and it's reaching the end of its battery life. Back at home, their physician had already scheduled the replacement surgery, but the patient had to travel to see their ill relative. A major concern for the family was the condition of the pacemaker during the trip. All they knew was that if the pulse got stuck at 60bpm even with physical movement, they'd reached the danger zone and the pacemaker could stop at any moment.

To keep an eye on this, we used Vivral's on-device camera-based heart rate sensor to capture his heart rate before and after movement. Our first resting output read 60—exactly what the pacemaker forces the heart to beat at rest. After some uphill climbing, Vivral outputted a heart rate of 76bpm. We had a home nurse verify this through a manual pulse check just to be sure.

Honestly, one's phone is just as important as one's wallet now. With Vivral, patients can take general heart rate readings and actually ease their mind about their health without a clunky monitor. (Note: Vivral is not FDA-approved and is merely a tool for education. All tests were performed under medical supervision by a licensed physician.)

On-Device Monitoring Vivral's camera-based sensor effectively tracked heart rate changes during physical movement.

Medical Supervision A nurse verified the readings. Vivral gave the family peace of mind on the go.

Overcoming Literacy and Language Barriers

As we're rolling this out to more people, it's pretty clear that a lot of patients couldn't speak English well, and many couldn't write entirely. An AI assistant is basically useless if the patient can't even talk to it.

To fix this fast, we added a lightweight translation and transcription layer right into the app. Now, patients can just speak to Vivral in their local language without needing to type anything. Vivral takes the audio, translates it, figures out what they need, and speaks the clinical response back to them out loud. By making it voice-first, we completely bypassed the reading and writing barrier. It's a huge step in making sure the people who need it most can actually use it.

Multilingual Voice Layer Added a lightweight translation and speech-to-text pipeline for local dialects.

Accessibility Focus Voice feedback completely removes the reading and writing hurdle for our patients.

Rural Connectivity and the Triage Gap

Today we took a trip to a more northern region in Kerala. The internet out there was so terrible in so many areas that even Google Maps didn't work. We passed by thousands of villages, and the nearest functional hospital was over 100km away.

For these thousands of people, they all have cell phones, but they just don't have access to credible medical care on a daily basis. Most of these people don't even have usable cars. It's either they suffer and let NGOs and volunteer physicians occasionally take care of them, or they attempt a massive journey without knowing if they actually need to.

With an AI triage layer like Vivral, they can get a quick look at their symptoms right on their phone to see the severity of their case and figure out whether that 100km hospital trip is actually worth it. We clearly found a massive place for Vivral to be used out here.

Infrastructure Gap Hospitals are 100km+ away, internet is spotty, and getting to a doctor is a massive ordeal.

Triage Utility Vivral lets isolated patients figure out if a trip to the hospital is actually necessary.

Systemic Burden & The Need for Triage

I was talking with a patient's sibling today, and she told me a really heartbreaking story about losing her husband. He had a poorly done surgery, was in immense pain, and was rushed to the hospital showing stroke symptoms. But the hospital was so overcrowded that no doctor was available to see him. They waited for hours and then tried to transfer him somewhere else, but he passed away before anyone could help.

This story just proves why Vivral is so necessary. When we visit these hospitals, they are packed to the brim, a lot of times with people who don't even need acute emergency care. If we can use Vivral to handle the routine cases and do remote triage, we can clear out the waiting rooms so real doctors actually have time for the life-threatening emergencies.

Systemic Strain Overcrowded hospitals are leading to fatal delays for critical patients.

Triage Capability Vivral can filter out the routine stuff, keeping the waiting rooms clear for real emergencies.

Data Tracking & Supervised Care Adjustments

We noticed our patient's blood sugar was dangerously high for a diabetic today. Using Vivral to track it, we put together a plan to keep monitoring his levels and slowly bump up his insulin dosage from 10 to 15 units until things settled down.

It's really important to note that this whole process was strictly supervised and approved by a licensed doctor. Vivral was just the tool keeping everything organized and tracking the data so the doctor could make the right call. The patient's blood sugar is stable now, and Vivral is just running in the background, keeping a clean log of his vitals.

Vital Tracking Vivral tracked the crazy blood sugar swings so the doctor had clean data to work with.

Medical Supervision The doctor made all the calls on insulin; Vivral just organized the numbers.

Patient Passport & Automated Care Management

Today we really pushed the Patient Passport feature. We used Vivral's OCR to scan and upload our patient's entire medical history. It ate up all those complex files without a problem. Almost right away, Vivral started using that context to send automated, proactive notifications, like reminding him about upcoming follow-up visits.

It was amazing to see. We realized that translating confusing doctor's notes and discharge papers into plain English is huge—especially for older patients who might be forgetful after a big hospital stay. When they go home, they usually don't have the energy to manage their own care, so having an app automatically handle the schedule means important steps aren't forgotten.

OCR & Context We scanned in his whole medical history, and the AI instantly understood his health baseline.

Post-Discharge Support Explaining messy discharge papers in plain English is a lifesaver for older patients.

Hospital Logistics & Telemedicine Validation

Going to the hospital today really showed us the logistical nightmare these patients face. The place was insanely crowded, and we couldn't even find a wheelchair, which made getting around super tough and physically draining for the patient.

After waiting for hours just to get a few minutes with a stressed-out doctor, the final diagnosis was exactly what Vivral had already predicted before we even left the house. It really validated our telemedicine approach. We could have skipped all that physical and emotional exhaustion by just doing an AI-guided consult from the couch.

Infrastructure Gap No wheelchairs and insane crowds made the hospital trip a miserable experience for the patient.

Diagnostic Accuracy Vivral guessed the diagnosis perfectly before we even saw the doctor.

User Interaction & Model Preference

We're noticing the patient and their caretaker are using Vivral a lot—like 10 to 20 questions a day, asking everything from basic care stuff to deep medical questions. Vivral Swift is definitely the favorite model right now just because it's so fast and handles almost everything they throw at it.

Honestly, we haven't seen a real need to switch them to the Surreal model yet. Right now, we're shifting our focus to polish up the rehab features in the app to help out more with physical recovery.

Model Choice Vivral Swift is winning out because it's fast and gets the job done for daily chats.

Active Development Putting more dev time into the rehab features to help with recovery.

Wearable Integration & Diagnostic Refinement

The main goal today was getting Apple Watch support wired into the app. Now we can automatically pull in really accurate, continuous health data, like overnight vitals and sleep tracking. Having that kind of long-term data running in the background makes our preventative diagnostics way sharper.

We also pushed some patches to fix latency when the chat gets really long. While testing, we caught the AI hallucinating some weird stuff about PSA counts. Looked into it and found out it was just a typo error when it was reading the data. We patched it and we're hitting it with some RLHF to make sure it doesn't happen again.

Feature Deployed Hooked up the Apple Watch so we can pull overnight vitals automatically.

Patch Deployed Fixed a weird PSA hallucination caused by a typo, and smoothed out chat lag.

Large Context Imports & Model Optimization

Spent today importing massive patient histories into Vivral. The context sizes we were dealing with were huge. To make it work smoothly, we actually had to go in and rewrite the core Patient Passport backend so it could handle and search through that much data without choking.

We also realized that when people ask simple, casual questions (like "how do I bathe a patient who can't move?"), the AI was giving pretty weak answers compared to when we asked long, formal medical questions. So, we've started some Reinforcement Learning (RL) to train the model to be better at handling just regular, everyday questions.

Model Performance Vivral Swift is handling the daily load better and faster than the other models.

Architecture Update Had to rewrite the backend so it wouldn't crash when loading huge medical files.

Model Refinement While visual detection succeeded, the model showed a tendency to over-diagnose. Upcoming RL and the V2 model will focus on narrowing differential diagnoses.

Camera-Based Vitals & Pacemaker Monitoring

A family member of a patient here requires a pacemaker, which is currently nearing the end of its battery life. Before traveling to visit an ill relative, their physician had already scheduled a replacement surgery. Naturally, the family was deeply concerned about the pacemaker's condition during the trip. They were advised that if the patient's heart rate remained stuck at 60 BPM despite physical exertion, the device had entered a danger zone and could fail at any moment.

To monitor this, we utilized Vivral's on-device, camera-based heart rate sensor to capture readings before and after movement. Our initial resting output read exactly 60 BPM—the baseline rate enforced by the pacemaker. After the patient engaged in some uphill climbing, Vivral successfully recorded an increased heart rate of 76 BPM. We then had a home nurse verify this reading via a manual pulse check.

Today, a smartphone is as essential as a wallet. With Vivral, patients can take general heart rate readings to gain immediate peace of mind regarding their health. Note: Vivral is not FDA-approved and is intended merely as an educational tool. All tests described were performed under the direct medical supervision of a licensed physician.

On-Device Monitoring Vivral's camera-based sensor accurately tracked heart rate changes during physical exertion.

Medical Supervision Readings were corroborated by a nurse; Vivral functions as an educational tool, not a certified medical device.

Overcoming Literacy and Language Barriers

As our deployment expands, it has become starkly apparent that language proficiency and general literacy present significant hurdles to healthcare access. Many of the patients we interact with have limited English proficiency and struggle with written communication entirely. An intelligent medical assistant is only as effective as its accessibility; if patients cannot interact with the tool naturally, its diagnostic potential is severely bottlenecked.

To immediately address this, our team engineered and deployed a lightweight translation and transcription layer natively into the Vivral application. This localized update enables patients to speak directly to the assistant in their regional dialects without the friction of typing. Vivral instantly processes the audio, translates the intent, and reads its clinical responses back aloud. By prioritizing a voice-first, multilingual interface, we have effectively removed both literacy and language requirements, ensuring the most vulnerable demographics can reliably seek care.

Multilingual Voice Layer Deployed a lightweight translation and speech-to-text pipeline to support regional spoken dialects.

Accessibility Focus Text-to-speech audio feedback successfully eliminates reading and writing barriers for elderly or low-literacy patients.

Rural Connectivity and the Triage Gap

Today, our team traveled to a more northern region of Kerala. The infrastructure shift was immediate and drastic; internet connectivity became so unreliable that even basic GPS navigation applications like Google Maps ceased to function. As we passed through thousands of rural villages, a harsh reality set in: the nearest functional hospital for these communities is often over 100 kilometers away.

Despite this extreme isolation and a widespread lack of personal vehicles, nearly everyone we observed possessed a cellular device. For these populations, credible medical care is virtually inaccessible on a daily basis. They are left with grim choices: endure their conditions while waiting for sporadic visits from NGOs and volunteer physicians, or attempt a massive, arduous journey without knowing if it is truly necessary.

This environment crystallized a critical use-case for Vivral. By providing an accessible AI triage layer, patients can receive immediate, preliminary symptom assessment directly on their phones. Vivral can help determine case severity and intelligently advise whether a 100km hospital trip is medically warranted, bridging the catastrophic gap between having a smartphone and having access to actionable medical guidance.

Infrastructure Gap Observed extreme rural isolation where hospitals are 100km+ away and mobile internet is severely degraded.

Triage Utility Vivral provides vital symptom assessment, helping patients confidently decide if an arduous hospital journey is required.

Systemic Burden & The Need for Triage

Today, a conversation with the patient's sibling revealed a heartbreaking story regarding the loss of her husband. Following a poorly executed surgery, he experienced immense pain and was rushed to the hospital exhibiting symptoms of a stroke. Tragically, due to severe hospital overcrowding, a doctor was unavailable to attend to him. After an agonizing wait, they attempted a transfer to another facility, but he passed away before receiving care.

This tragic account powerfully underscores the absolute necessity for Vivral. As we observed during our own hospital visits, facilities are filled to the brim, often with patients who may not require acute emergency care. Resources like Vivral can act as a critical triage layer, managing routine cases and utilizing telemedicine to bridge the gap when in-person doctors are overwhelmed. As populations increase and the ratio of available physicians declines, a reliable, scalable system must step in to handle everyday healthcare, preserving the time of real doctors for the most complex, life-threatening cases. We are finding profound, real-world evidence validating this critical gap in the market.

Systemic Strain Overcrowded hospitals lead to fatal delays; many patients occupying space could be triaged remotely.

Triage Capability Vivral provides a necessary buffer, managing routine inquiries to free up specialized doctors for critical emergencies.

Data Tracking & Supervised Care Adjustments

We noticed our patient's blood sugar was dangerously high for a diabetic. Using Vivral's tracking features, we devised a plan to continuously monitor their levels and incrementally adjust their insulin dosage (from 10 to 15 units) until a safe baseline was achieved.

Crucially, this entire process was strictly supervised and approved by a licensed medical professional. Vivral acted purely as an organizational and tracking tool; all care and dosage decisions were provided directly by the attending physician. The patient now maintains stable blood sugar, and Vivral continues to automatically track and organize these vitals for ongoing review.

Vital Tracking Vivral effectively monitored and organized fluctuating blood sugar levels to provide clear data for the physician.

Medical Supervision All insulin adjustments were planned and executed exclusively by a medical professional, emphasizing Vivral's role as a supportive assistant.

Patient Passport & Automated Care Management

Today we tested the full capacity of Vivral's Patient Passport. Using Vivral OCR, we uploaded our patient's entire medical history, successfully storing and indexing all complex medical files. Almost immediately, Vivral processed this context and began providing automated, proactive notifications, including reminders for upcoming follow-up visits.

This capability proved transformative. We found that parsing and explaining complex doctor's notes and discharge papers is essential—especially for elderly patients experiencing memory issues following a major medical event. Often, patients returning home lack the strength or clarity to manage their own post-discharge care; having an intelligent assistant automatically synthesize this information ensures critical care steps aren't missed.

OCR & Context Successfully digitized comprehensive medical histories, allowing the AI to instantly grasp the patient's full health narrative.

Post-Discharge Support Automated notifications and plain-language explanations of discharge papers provided a vital safety net for elderly care.

Hospital Logistics & Telemedicine Validation

Today's hospital visit highlighted significant infrastructural challenges; the facility was severely overcrowded, and a lack of available wheelchairs made navigation difficult and physically taxing for the patient. Despite a multi-hour wait for a brief consultation with a busy physician, the final diagnosis was identical to what Vivral had already predicted.

This experience underscores the value of Vivral's telemedicine integration. Much of the physical and emotional strain of the visit could have been avoided by performing an AI-assisted consultation from the comfort of home, leveraging the patient's full medical history for a more efficient and comfortable experience.

Infrastructure Gap Observed severe overcrowding and lack of mobility aids, causing patient nausea and unnecessary physical strain.

Diagnostic Accuracy Vivral accurately predicted the physician's final diagnosis in advance, validating its predictive capabilities in a real-world setting.

User Interaction & Model Preference

We observed that the patient and caretaker engage with Vivral frequently, asking between 10-20 questions daily, ranging from basic care inquiries to complex medical questions. Vivral Swift has emerged as the preferred model due to its exceptional speed and comprehensive coverage of most queries.

At this stage, we haven't identified a compelling need for the Surreal model. We are currently focusing on refining the rehabilitation features within the application to further support recovery and will monitor its impact as it evolves.

Model Choice Vivral Swift remains the favorite for its responsiveness and utility in day-to-day interactions.

Active Development Refining the rehab feature set to better address specific patient recovery needs.

Wearable Integration & Diagnostic Refinement

Today's primary objective was the integration of Apple Watch support into the Vivral ecosystem. This enhancement allows us to automatically ingest continuous, high-fidelity health data, including overnight vitals and sleep analysis. This continuous telemetry provides crucial longitudinal data that vastly improves the accuracy of our preventative diagnostics.

Concurrently, we deployed a suite of general performance patches aimed at stabilizing latency during long-form conversational diagnostic sessions. During QA testing, we identified a hallucination involving Prostate-Specific Antigen (PSA) counts. Root-cause analysis revealed it stemmed from a typographical processing error during data ingestion. We have isolated the fault and are currently employing targeted Reinforcement Learning from Human Feedback (RLHF) to systematically correct the behavior and prevent similar misinterpretations.

Feature Deployed Apple Watch telemetry integration for continuous overnight vital monitoring and advanced diagnostic baselines.

Patch Deployed Latency optimization for long-form diagnostic sessions and RLHF adjustments to correct typographical parsing errors regarding PSA counts.

Large Context Imports & Model Optimization

We spent today importing massive patient histories into Vivral. The context sizes were enormous. In order to accommodate this effectively, we had to update and rewrite the core Patient Passport architecture to efficiently manage and retrieve from these large contexts.

Additionally, we discovered that simple, informal queries (e.g., asking for the technique to bathe a patient who has difficulty moving) yielded surprisingly poor responses compared to long, formal questions. To combat this, we have initiated Reinforcement Learning (RL) targeted specifically at improving the system's ability to handle basic questions.

Model Performance We found that Vivral Swift was the optimal model for the majority of queries that arose today. It was significantly faster and answered relevant questions better than alternatives.

Architecture Update Patient Passport storage rewritten to support massive historical context windows.

AI was used to generate some of this text. Nonetheless all text is based on real scenarios. Our paper covers the case-study in detail.