An ABDM Compliance Roadmap for Indian Hospitals

Two registries sit underneath everything else. The Health Facility Registry holds a verified record of the hospital, and the Healthcare Professional Registry holds verified identities for the clinicians who practise in it. Neither is administrative decoration, because shared records are attributed to a registered facility and an enrolled professional.

Facility registration asks for consistent detail on ownership, address, services offered, and infrastructure, and a group with several branches should expect to register each site rather than the organisation. Professional enrolment depends on individual clinicians completing their own verification, which is the step that runs slowest because the IT team cannot do it on their behalf. Begin it early, track it as a named workstream, and assume visiting consultants will need reminding more than once.

Both registries need maintenance after go-live. A department that closes, a branch that relocates, or a consultant who leaves creates a gap between what the registry asserts and what the hospital actually does. Give registry upkeep an owner in the same way you would a licence renewal.

An Ayushman Bharat Health Account gives the patient a portable identity that records can be linked to, and the natural moment to capture or establish it is at registration. Hospitals that create a separate ABHA counter usually find it staffed on quiet days and abandoned on busy ones. The linkage has to sit inside the flow the front desk already follows, between identity capture and encounter creation.

Three paths need designing: the patient who already holds an ABHA and can present it, the patient who wants one created, and the patient who declines. Each path must end with a registered encounter, because a patient without an ABHA still needs care and a bill. Designing only the first path is what produces queues and improvised workarounds.

This is the highest-leverage workflow decision in the whole programme. Everything downstream depends on records being attached to the right identity at the point of registration, and no later reconciliation recovers a linkage that was never made. A companion article on this blog covers the desk-level design in detail.

ABDM standardises on HL7 FHIR R4, which means shared documents are structured resources rather than scanned pages or free text. A hospital that stores discharge summaries as PDFs can produce a document reference, but it cannot produce the coded problems, medications, and results that make a record useful to the next clinician. The distance between those two states is where most of the real work sits.

Audit what your system captures as discrete data today: diagnoses, medications with dose and route, laboratory results with units and reference ranges, vitals, allergies, and procedures. Anything held as free text will need a structured field, a terminology binding, and usually a change in how clinicians document. That is a clinical change programme, not a mapping exercise.

Doing this properly pays for itself well beyond compliance. The same structured record improves internal reporting, decision support, and the patient-facing view in a portal. HealUDoc's EHR (/ehr) and patient portal (/patient-portal) draw on the same structured capture, so the investment is not spent solely on satisfying an external requirement.

Sharing under ABDM is not authorised by a login or by an agreement between two institutions. A request carries a consent artefact granted by the patient through a consent manager, stating what may be shared, for what purpose, over which period, and with whom. Your system's job is to honour exactly that scope and nothing wider.

There is a practical consequence here that surprises teams. If the record store can only export a whole chart, it cannot serve a consent limited to laboratory results from a defined window, so filtering by information type and date range has to work at request time. Retrofitting that capability later is considerably harder than designing for it.

Log every consent-based disclosure alongside the artefact it relied on. When a patient or a regulator asks what was shared and on what basis, the answer must come from a record rather than from recollection. That log is also the evidence an accreditation assessor will ask to see.

Integration is proved in a sandbox environment through a staged sequence of milestones, each building on the last rather than arriving as a single event. Plan for iterations. Teams routinely underestimate how many cycles it takes to satisfy record-format and consent-handling checks, particularly when clinical documentation is changing in parallel.

Be sceptical of any claim that a product makes a hospital ABDM ready on its own. Software can support registration, identity capture, structured records, and consent handling, but the registrations, the workflow, the documentation quality, and the ongoing upkeep belong to the hospital. Ask a vendor which milestones their integration has passed and what explicitly remains yours to do.

Give the programme a realistic horizon and review it quarterly. Registry accuracy, ABHA linkage rates, structured-capture completeness, and consent-request handling all drift without attention. A hospital that treats ABDM as a live obligation rather than a completed project is the one still compliant a year later.