Medical Records Retention Rules for Hospitals in India

A common mistake is to set retention by where a record was created — radiology keeps things for X years, OPD for Y. Retention should follow the legal and clinical character of the record instead. An imaging study from a routine outpatient visit and an imaging study documenting an assault injury carry entirely different risk profiles despite being produced by the same machine.

Build the schedule around categories: routine outpatient records, inpatient case records, operative and anaesthesia records, medico-legal cases, obstetric and neonatal records, records of minors, imaging and other large binary objects, transfusion and implant traceability records, and consent documents. Assign each a period and a rationale, and record who approved it.

The rationale column is the part hospitals skip and later regret. When an assessor, a lawyer, or a new records manager asks why obstetric records are held for a particular period, a written rationale is the difference between a defensible policy and an arbitrary one.

MLC records should be treated as a separate class from the moment they are created, flagged in the system, and excluded from routine disposal by default. The trigger for release is the conclusion of proceedings, communicated in writing, not the passage of a fixed period. In practice many hospitals retain MLC files indefinitely because the cost of storage is trivial next to the cost of not having them.

A litigation hold is the mechanism that suspends normal disposal for any record connected to an actual or reasonably anticipated proceeding. It must be capable of being applied quickly, of covering an entire patient record rather than a single document, and of being audited afterwards. If your disposal process cannot demonstrate that held records were excluded, the hold has limited value.

Applying and releasing holds must itself be logged: who applied it, on what authority, which records were covered, and when it was lifted. A platform such as HealUDoc can carry a hold flag at the patient and encounter level so any automated retention job skips them and records that it did.

For patients treated as minors, the limitation period for bringing a claim generally does not begin to run until they reach majority. A retention period measured from the date of treatment is therefore the wrong measure entirely; the period should be measured from the date the patient attains majority, plus the applicable limitation window.

Obstetric and neonatal records deserve the same treatment for the same reason, and they carry a further complication: a single episode generates records for two patients whose retention clocks differ. Systems that file everything under the mother's record can make the neonate's documentation difficult to retrieve years later. Ensure the neonatal record is independently identifiable and independently retained.

This is one area where a general policy statement is genuinely insufficient. Ask your legal advisor to specify the calculation for your jurisdiction, encode it in the retention schedule as a formula rather than a fixed number of years, and make sure whoever administers disposal understands why some records with old treatment dates are not yet eligible.

Digitising records changes what you must be able to prove. A paper file speaks for itself; an electronic record has to be shown to have been produced by a system operating properly, and Indian evidence law requires an accompanying certificate from a person in a responsible position regarding the computer output. That requirement shapes your architecture more than most hospitals realise.

In practice it means you need to be able to demonstrate access controls, audit trails showing who created and amended each entry, integrity controls that would reveal tampering, backup and restoration evidence, and system logs covering the relevant period. Retaining the clinical content while discarding the audit trail is a common and serious error — the metadata is part of what makes the record credible.

Format obsolescence is the other digital risk. A record retained for twenty years must still be readable in twenty years, which argues for open, well-documented formats and for periodic migration with a documented chain of what was converted, when, and by what process. A proprietary viewer that ceases to exist takes the record with it.

Destruction of records is a governed act, not a housekeeping task. The process should generate a disposal certificate listing what was destroyed, the category and date range, the authority under which it was destroyed, the method, the date, and the signatures of the responsible officers. That certificate is retained permanently, because it is your evidence that a missing record was disposed of lawfully rather than lost.

Physical destruction should be witnessed and irreversible — shredding or incineration rather than disposal to a waste stream. Digital destruction must account for backups, archives, replicas, and any copies held by processors, which is precisely where most deletion policies fail. If a record is deleted from production and remains in a five-year backup, it has not been deleted.

The DPDP Act 2023 adds an obligation in the opposite direction: personal data should not be retained indefinitely once the purpose for which it was collected has been served and no legal requirement compels retention. Hospitals now have to be able to justify keeping data as well as deleting it, which makes the documented schedule the central artefact for both duties.

Start by inventorying what you hold and where — including the places records accumulate outside the records department, such as departmental drives, imaging archives, consultant devices, and off-site storage. Most hospitals discover holdings nobody owns, and unowned records are the ones that are neither retained properly nor disposed of properly.

Then write the schedule: category, retention period, the event the clock starts from, the authority or rationale, the owner, and the disposal method. Have it reviewed by legal counsel and approved at board or governing-body level, because retention decisions carry institutional risk and should not sit with a single manager.

Finally, make it operational. A schedule that requires manual identification of eligible records will not be followed. Encode the categories in the system, run the eligibility calculation automatically, route the resulting list for human approval before anything is destroyed, and log the whole cycle. HealUDoc can classify records by category at the point of creation so the eligibility calculation runs on data the system already holds rather than on a manual review of the archive. Review the schedule annually and whenever the regulatory picture changes.