Running an Open-Source EMR in India: What I Learned the Hard Way

I have installed, broken, rebuilt and lived with more open-source EMR software than I would like to admit. Before we settled on OpenEMR for our telemedicine practice, I tried several others, including OpenMRS, the Java-based system that many hospital projects in Africa and Asia run on. Every one of them was free to download. None of them was free to run.

This is not a feature comparison. Vendors and review sites already publish plenty of those. This is what I learned from actually running an open-source EMR for real patients in India: where the hard work is, where the risk is, and how a small clinic should think about it before spending a single rupee or a single weekend.

Why open source at all?

Commercial EMRs charge per doctor, per month, forever. More importantly, they decide what your workflow looks like. If their prescription format does not match how you practise, you adapt to the software, not the other way round.

Open-source EMRs such as OpenEMR give you the source code. You own your data, you choose where it lives, and you can change anything. That last part is both the biggest advantage and the biggest trap.

Installation is easy. Customisation is the real work.

Getting OpenEMR running took an afternoon. Getting it to work the way an Indian practice works took far longer.

Out of the box, most open-source EMRs are built around American or international workflows: insurance billing codes, US-style forms, drug databases that do not know Indian brand names. An Indian outpatient practice needs different things. Prescriptions must meet the requirements of the Telemedicine Practice Guidelines. Patients expect the prescription on their phone, often as a PDF on WhatsApp or email. Doctors want to type a brand name they actually prescribe and have it appear instantly, with the dose and duration they usually use.

Every one of those changes meant reading someone else’s code, understanding how the system was put together, and making a change that would not break when the next version came out. Upgrades are where many customised installations quietly die. If you change core files without keeping careful notes, the next upgrade either overwrites your work or refuses to run.

Two things made this manageable for me.

The first was the community. The OpenEMR forums and the wider open-source community have answered almost every question I have had, often years before I thought to ask it. Someone, somewhere, has already hit your error message. Search the forums before you search anywhere else.

The second was AI. Modern AI assistants have changed what a doctor with limited programming experience can do with open-source software. They can explain a confusing piece of code, suggest how a change should be made, and help you test it. They are not perfect, and you must check their work, but they turned months of work into weeks. One rule matters more than any other here: never paste real patient data into an AI tool. Use made-up test patients. The moment patient data leaves your system, you are responsible for where it went.

Where should your EMR live?

This is the decision most doctors skip, and it is the one that matters most for patient safety.

Option 1: On a laptop or your clinic’s own network

For a small clinic with one or two doctors working from one location, running the EMR on a single computer or on the clinic’s internal network is completely feasible, and in many ways the safest choice. OpenEMR runs well on an ordinary laptop or a small desktop machine. Nobody on the internet can reach it, because it is not on the internet.

The risks here are physical rather than digital. Laptops get stolen, hard disks fail, and clinics flood. If you go this way, three things are non-negotiable:

  • Turn on full-disk encryption (BitLocker on Windows, FileVault on a Mac), so a stolen laptop does not mean stolen patient records.
  • Take automatic, encrypted backups every day, and keep at least one copy outside the clinic.
  • Give every staff member their own login. Shared passwords make it impossible to know who looked at what.

Option 2: Online, on a server you control

An online EMR is far more convenient. Doctors can see patient history from home, from a second clinic or during a video consultation. At Dofody, our doctors consult patients across Kerala and beyond online, so an EMR that only worked inside one building was never an option for us.

But convenience has a price. The moment your EMR is reachable from the internet, it is reachable by everyone on the internet, including people scanning for exactly this kind of software around the clock. You are no longer just a doctor keeping records. You are the person responsible for guarding a database of other people’s health information.

Why not shared hosting?

Many doctors start with cheap shared web hosting because that is where their clinic website already lives. For an EMR, I would strongly advise against it.

On shared hosting, your EMR sits on the same machine as hundreds of other customers’ websites. You do not control the server’s software versions, its firewall, its security updates or who else has access to the machine. If another site on the same server is hacked, yours may be next. You often cannot even see proper access logs.

We moved our EMR to a dedicated virtual server: a machine where only our software runs and where we control every layer. That gives you the ability to close every door that does not need to be open, to restrict who can even reach the login page, to apply security updates the day they are released, and to keep detailed logs of who accessed what. It also means that every one of those jobs is now yours. A dedicated server is safer only if someone actually does the work of looking after it.

The work that never ends: security and the DPDP Act

Setting up the EMR was a project. Keeping it secure is a permanent job.

India’s Digital Personal Data Protection Act, 2023, and the DPDP Rules, 2025, now set out what every organisation that handles personal data must do, and a clinic’s patient records are exactly that. The main obligations apply from May 2027. For a clinic running its own EMR, that means in practice:

  • Clear notice and consent. Patients should know what data you collect, why, and how they can ask for it to be corrected or erased.
  • Reasonable security safeguards. Encryption, access controls, so that each person only sees what they need, and logs of who accessed which records, kept for at least a year.
  • Breach reporting. If patient data is exposed, you must inform the Data Protection Board and the affected patients. The penalty for failing to protect personal data can run up to ₹250 crore.
  • Children’s data. Records of patients under 18 need verifiable consent from a parent or guardian.
  • Keeping only what you need. Old data you no longer have a reason to hold is a liability, not an asset, although medical records rules still require you to keep records for a minimum period.

For us, this means regular software updates, reviewing who has access, checking logs, testing that backups can actually be restored, and treating every new feature as a question: does this create a new way for patient data to leak? This is not a task you finish. It is a habit you keep.

If that list feels heavy for a two-doctor clinic, that is a good reason to run the EMR offline on the clinic’s own network until you genuinely need it online.

Is it worth it? Every doctor I know says yes

After all of that, you might wonder whether an EMR is worth the effort. I have not met a single doctor who started using one properly and then went back to paper.

The reasons are simple and they show up in the first month:

  • Legible prescriptions, every time, with no pharmacist phoning to ask what you wrote.
  • The whole history in seconds. Previous visits, blood test trends, current medicines and allergies on one screen, instead of a patient searching their bag for last year’s file.
  • Safer prescribing. Allergy and interaction warnings catch the mistakes tired doctors make at the end of a long OPD.
  • Continuity of care. A colleague seeing your patient on your day off knows exactly what you did and why.
  • Records that cannot be lost to a leaking roof, provided you take your backups seriously.

Paper records feel simple only because their risks are invisible. The file that went missing, the prescription the pharmacist misread, the allergy nobody wrote on the cover: those costs are real, they just never appear on a bill.

Where to start

If you are a doctor thinking about an open-source EMR, my advice is this. Start small, offline, with one doctor and fake patients. Learn the system before you change it. Keep a written note of every customisation. Lean on the community and on AI, but never share patient data with either. And move online only when the convenience is worth the responsibility that comes with it.

If you are curious why so many health-tech products struggle in India, including EMRs that look perfect in a demo, I have written about that in Why Health-Tech Products Fail in India.

Build the systems, not just the practice

Choosing and running an EMR is one chapter of a much bigger shift: moving from being the doctor who does all the work to being the doctor who builds the systems that do it reliably. That shift is what my upcoming book, The Digital Doctor’s Blueprint, is about. It covers digital infrastructure, patient data, practice workflows and building a practice that does not depend on you being in the room for every decision.

Pre-order The Digital Doctor’s Blueprint and get early access to Chapter 1.

more insights