India Has Built the Digital Roads. Cancer Patients Still Need a Guide
How AI can make patient navigation work at Indian scale, and why the navigator must stay human
Every week, some version of the same scene plays out at the big cancer hospitals in Kolkata, Mumbai, Delhi, and Chennai.
A family arrives from a district town after an overnight journey. The father is the patient. The son is holding a plastic carry bag, the kind you get from a cloth shop, and inside it is the entire medical history: a biopsy report from a private lab, two CT scans on CDs nobody can open anymore, a discharge summary written partly by hand, prescriptions from three different doctors, and a photocopy of a ration card in case it is needed for a scheme.
They don’t know which counter to go to first. They don’t know whether the biopsy needs to be repeated here. They don’t know if the scan is too old. They are not even sure which of the papers matters. With OncoVOIS, we plan to try to solve this issue for the patients.
Nobody in that family lacks courage. What they lack is a guide.
That is what patient navigation is. And after years of doing this work at CancerFax, I am convinced that India is closer than almost any country to making navigation available to every cancer patient, not just the lucky few. The reason is a combination of two things: the digital infrastructure India has quietly built and the AI tools that can finally sit on top of it.
But only if we build it the right way.
The real problem is not always treatment. It is the path to treatment.
India adds roughly 15.6 lakh new cancer cases a year, according to the WHO’s Global Status Report on Cancer 2026, with nearly 8.7 lakh deaths. ICMR’s National Cancer Registry Programme estimated close to 15.7 lakh cases for 2025, with Uttar Pradesh, Maharashtra, West Bengal, Bihar, and Tamil Nadu accounting for the largest numbers.
When people talk about these figures, the conversation usually turns to hospitals, machines, and oncologists. All of those shortages are real. But there is a quieter problem that causes just as much harm, and it is the one I see most often.
Patients get lost between steps.
A suspicious scan in a district hospital. A referral to a city. A biopsy that takes weeks to schedule. A pathology report that reaches the wrong desk. A family that runs out of money for the next trip and simply stops coming. Each step makes sense on its own. But the patient is not really “owned” by any one place, and so when they drop out, often nobody notices until they come back months later with a disease that has moved beyond the point where it could be cured.
An oncologist from India, Dr. Uddiptya Goswami, described this well in a recent piece for CancerWorld: in low- and middle-income countries, the greatest barrier is often not that treatment is unavailable, but that the path to it is fragmented and deceptively simple to fall out of. I could not agree more.
India has already proven that navigation works
This is not a theory. India has its own evidence.
Tata Memorial Centre, together with the Tata Institute of Social Sciences and Tata Trusts, started KEVAT, India’s first formal patient navigation program, including a one-year postgraduate diploma in oncology patient navigation. According to the CancerWorld report, around 130 KEVAT navigators working across nine Tata Memorial Centre hospitals have supported roughly six lakh patients over five years. In advanced lung cancer, navigation was associated with a median nine-day reduction in the time to start cancer-directed treatment.
Nine days may sound small. In advanced lung cancer, it is not.
During COVID, these navigators did much more than book appointments. They helped families find financial support, accommodation, and transport. That is what navigation actually looks like in India: part clinical coordinator, part translator, part social worker, and part the person who remembers your name.
And then the math breaks
Here is the uncomfortable part. One hundred and thirty navigators, however skilled, cannot follow fifteen lakh new patients a year. Even if we trained ten times as many, the numbers would still not close.
A navigator’s day is also not spent the way people imagine. A large share of it goes into reading documents, building timelines, chasing reports, calling patients who missed appointments, and explaining the same terms again and again in different languages. All of that is important. Very little of it needs a human being to do the first draft.
This is exactly where AI belongs. Not in place of the navigator. Underneath the navigator, doing the heavy lifting so the human can spend their time where only a human can help.
India’s unusual advantage: the roads are already being laid
What makes India different from many countries at a similar income level is that the digital groundwork for these services already exists, at a national scale.
Health records that can follow the patient. Under the Ayushman Bharat Digital Mission, more than 93.95 crore ABHA health IDs had been created and over 105 crore health records linked as of July 2026. ABHA is built around consent-based sharing, which means, in principle, a patient’s reports from a district lab and a city hospital can be pulled together with their permission instead of travelling in a plastic bag.
Telemedicine that already reaches villages. eSanjeevani, the national telemedicine service, has crossed nearly 49 crore consultations through over 19,000 hubs and more than 1.4 lakh spokes, many of them at Ayushman Arogya Mandirs in rural areas. The habit of seeing a specialist through a screen is no longer new for millions of Indians.
Language technology as public infrastructure. BHASHINI, run under MeitY, is India’s national language digital public infrastructure, using AI for translation and speech across Indian languages. Government health platforms are already beginning to integrate it to serve people in all 22 scheduled languages.
Think of these as roads. They are wide and they are getting longer every month. But a road does not tell a frightened family which turn to take. That is the gap AI-supported navigation can fill.
Six places where AI can genuinely help Indian cancer patients
1. Turning the plastic bag into a case file
Indian medical records are messy in a very specific way. Scanned PDFs, phone photos of reports taken under tube lights, handwritten notes, and a mix of English, Hindi, Bengali, Tamil, and Marathi all appear on the same page. Modern AI models can now read these documents and pull out the facts that matter: diagnosis, histology, stage, biomarkers, treatments given, and dates.
The goal is not for AI to interpret the case. The goal is for a human coordinator to receive a clean, organized summary in minutes instead of hours, with every fact linked back to the page it came from so it can be checked.
2. Noticing what is missing
Often the most important thing in a cancer file is what is not there. No PET-CT. No molecular testing. No PD-L1, HER2, EGFR, or ALK results were available, which is what one would normally expect. Imaging that is five months old.
An AI system that flags these gaps on day one, instead of a specialist discovering them on day fifteen, can save precious weeks. In India, where every extra trip to a city costs a family money and wages, getting the right tests requested the first time matters even more.
3. Speaking the patient’s language, literally
A Santhali-speaking grandmother in Jharkhand and a Malayalam-speaking fisherman in Kerala should not have to understand English medical terms to understand their own illness. AI translation and voice tools mean a navigator can explain a treatment plan in the patient’s language, and a patient with limited reading ability can ask questions by voice instead of typing.
Get CancerFax - Advanced Cancer Treatment’s stories in your inbox
Join Medium for free to get updates from this writer.
This needs care. A single mistranslated word in a pathology report can change a recommendation. Anything clinically important should still be reviewed by a person. But the default no longer has to be English.
4. Spotting who is about to fall off the path
This is perhaps the most powerful use, and there is good evidence for it from other countries.
In one US program, natural language processing flagged imaging reports that suggested possible pancreatic cancer, and a human navigation team then coordinated the next steps. The average time from imaging to biopsy fell from 22 days to 7, and the time to treatment fell from 56 days to 34. At Massachusetts General Hospital, a model that predicted which patients were likely to miss appointments, followed by a call from a navigator, brought the no-show rate down to 10.2 percent compared with 17.5 percent under usual care.
Imagine that logic applied across Indian cancer centers. A patient whose biopsy was advised three weeks ago but never done. A patient who has missed two chemotherapy cycles. A family that stopped answering calls. AI can surface these names every morning. A human navigator then picks up the phone.
5. Helping families with the money and the paperwork
For many Indian families, the hardest part of cancer is not the medicine. It is the cost and the documents needed to get help with it. Government schemes, insurance pre-authorizations, hospital estimates, and reimbursement forms. Families lose time and sometimes give up on treatment because this part becomes too much.
AI assistants can help a family understand which documents are usually needed, prepare checklists, and explain the process in simple language. Final decisions on eligibility always rest with the scheme or insurer, and families should be told that clearly. But nobody should abandon treatment because a form was confusing.
6. Looking wider when local options run out
Sometimes a patient has been through every standard option available in their city. This is where navigation extends beyond borders.
At CancerFax, a large part of our work is helping patients explore what may be possible elsewhere: a second opinion from a major Indian center, a clinical trial, or advanced options in China such as CAR-T therapy, TIL therapy, or other cell and gene therapies that are being developed and tested there at remarkable speed. AI can compare a structured case summary against long, technical trial criteria and produce a shortlist worth a closer look.
It does not decide eligibility. Trial investigators and treating oncologists do. And many of these options are investigational, meaning they may not work for everyone. But AI helps ensure a promising possibility is identified even if no one has time to read a forty-page protocol.
Where AI navigation in India can go wrong
I want to be honest about the risks, because in healthcare the harm from careless technology falls on the most vulnerable first.
English-only and smartphone-only tools exclude the people who need help most. If an AI navigator only works in English, on a new smartphone, with steady internet, it will serve urban, educated patients who already have options. It will miss the family with the plastic bag.
Models trained abroad may not understand Indian patients. Disease patterns, cancer types, treatment pathways, and costs in India differ from those in the US or Europe. A tool that has not been tested on Indian data should not be trusted to triage Indian patients.
An alert with nobody to act on it is useless. AI can flag a patient at risk of dropping out. If no navigator has the time or authority to follow up, the flag means nothing. Technology cannot compensate for a missing team.
Privacy is not optional. Cancer records are among the most sensitive data a person has. India’s Digital Personal Data Protection Rules, notified in November 2025, are coming into force in phases, with full compliance expected by May 2027. Any company building AI for patients should treat consent, security, and data minimization as foundations, not afterthoughts.
AI must never sound like a doctor’s verdict. It should not tell a family whether a treatment will work. It should not deliver bad news. It should not quote survival percentages from studies of different patients. A frightened family hears anything that sounds like a prediction as a promise or as a sentence.
What we are building at CancerFax, and the rules we hold ourselves to
At CancerFax we are building an oncology AI app into our patient navigation work, starting with the most time-consuming part: reading and organizing medical records so our team can review cases faster and more completely. We are doing it with a few rules that are not negotiable for us.
A human reviews every case summary before it reaches a doctor or a family. Every fact the AI extracts is linked to its source document. Gaps and missing tests are flagged, never guessed. Patients communicate in the language they are comfortable with. Consent comes first, always. And AI never tells a patient what their outcome will be.
The point of all of these guidelines is simple. Our coordinators should spend less time on paperwork and more time on the conversations that actually help a family: explaining options honestly, answering the late-night questions, and staying with them through decisions that are frightening and difficult.
The guide matters more than the road
India has done something many richer countries have not. It has built national digital rails for health records, telemedicine, and language. That is a genuine achievement.
But a cancer patient does not experience a health system as infrastructure. They experience it as a series of doors, queues, reports and decisions, usually while scared and often while far from home. What they need is someone who knows the way and walks with them.
AI can help India give that guide to many more people than our current workforce ever could. It can read the reports, catch the gaps, speak the language, and notice who is slipping away. What it cannot do is hold a family’s trust. That still has to be earned by people.
If we get this balance right, the family with the plastic bag will one day arrive at the hospital with their records already organized, their missing tests already done, and a navigator already expecting them.
That is the future worth building.
If you or someone you love is facing a complex cancer diagnosis and you are unsure about the next step, our team at CancerFax can help review medical records, identify missing investigations, arrange expert second opinions, and explore advanced treatment options in India, China, and other international centers where appropriate.
This article is for general information only and is not medical advice. A qualified oncology team should always make treatment decisions after fully reviewing the patient’s diagnosis, stage, test results, and treatment history. Investigational therapies and clinical trials carry uncertainties, and eligibility can only be confirmed after expert evaluation.
Sources
- World Health Organization, Global Status Report on Cancer 2026
- ICMR National Cancer Registry Programme estimates for 2025, as reported in Lok Sabha
- Uddiptya Goswami, “The Missing Link in Cancer Care: Can AI-supported patient navigation close the gap in LMICs?" CancerWorld, 1 September 2026
- Tata Memorial Centre, KEVAT Patient Navigation Program
- Ministry of Health and Family Welfare / PIB updates on the Ayushman Bharat Digital Mission, July 2026
- Union Health Ministry data on eSanjeevani, as reported in August 2026
- Digital India BHASHINI Division, MeitY, and Ministry of Ayush MoU, May 2026
- Digital Personal Data Protection Rules, 2025, notified by MeitY in November 2025
