How to Choose Clinic Software in India: 20 Questions for Your Demo
Use 20 demo questions, a clinic brief, supplier evidence and a worked scoring model to compare EMRs for Indian OPDs.
Evaluate an EMR export by whether the clinic can read, reconcile and reuse the records it needs. PDFs preserve a human-readable view; structured files support a different task and require documented field meanings, identifiers and relationships. Test attachments and a receiving-system import before relying on portability. A successful download or a folder full of JSON files is not an exit plan.
For Indian clinic owners comparing EMRs, planning migration, negotiating exit terms or checking access to records after a contract ends.
A patient’s authorized copy, a clinic continuity archive and a migration to another EMR are different projects. They may include different records, recipients and formats. Asking “Can we export our data?” without naming the purpose often produces a yes that does not answer the real requirement.
For a patient copy, readable documents and the appropriate scope may be central. For a migration, the receiving product may need patient identifiers, encounters, structured observations and attachment relationships. For recovery, the clinic needs a supported restore process and evidence that it works. Do not let one export button stand in for all three.
Start with a record inventory and a named destination. Include clinical documents, outside reports, prescriptions and relevant administrative information separately. Specify dates and any exclusions. This makes vendor responses comparable and reveals gaps before the clinic reaches the end of its contract.
A PDF may be easy to read but difficult to import as structured encounters. A CSV is useful for tabular data but needs definitions for columns, encoding and relationships. JSON can represent nested records, but its syntax does not tell another product what the fields mean. The extension is the beginning of the technical discussion, not its conclusion.
The NRCeS EHR standards reference distinguishes structural and semantic interoperability: exchanging a shape of data is different from preserving its meaning. Apply that distinction to the actual sample and receiving system rather than assuming a standards label guarantees a complete migration.
On a narrow screen, scroll the table horizontally to read all columns.
| Output | Useful for | Questions that remain |
|---|---|---|
| Readable PDF documents | A person reviewing the clinical record and issued documents. | Are all pages, dates, prescribers and corrections included? Can documents be connected to the correct patient and encounter? |
| CSV tables | Reviewing or importing defined rows such as patient indexes. | What do columns, blanks, statuses and dates mean? How are identifiers and links preserved? |
| JSON or another structured package | Representing related records for a documented import process. | Which schema/version is used? Which fields are proprietary? What can the receiving product actually consume? |
| Attachment archive | Preserving original reports and supporting documents. | Are the actual files included, or only links that expire when the old account closes? |
| Supported backup/restore package | Recovering the application or agreed record scope. | Which system can restore it, what is excluded and when was restoration tested? |
Source: NRCeS: EHR Standards for India—standards and interoperability
A useful data dictionary explains the meaning of each field, its format, allowed values and relationship to other records. “status: 2” is not enough without knowing whether 2 means draft, signed, inactive or something else. Dates also need clear event meanings and timezone handling.
Identify the stable keys. A family phone number should not become the migration’s patient identifier. Encounters, prescriptions and documents need relationships that survive a change of product. If the receiving system assigns new identifiers, the migration should retain a mapping back to the source rather than relying on similar names.
Ask how amended records, cancelled items and unresolved tasks are represented. A package that exports only the latest visible text may omit history needed to understand earlier decisions. Equally, a package containing all versions can create confusion unless the current and historical states are distinguishable.
For the fictional sample of 20 patients, 45 encounters and 12 attachments used below, ask for a package-level inventory alongside the records. The manifest should state when the snapshot was taken, the scope selected, the files delivered, their record counts where meaningful and any omissions. It gives the receiving team a way to distinguish a missing file from a file that contains no records by design.
A workable package can contain separate patient and encounter datasets, a document index, the actual attachment files and a data dictionary. Those are logical components, not a mandated folder layout or a promise about Bractis. If the vendor delivers them differently, require an equally clear route from an exported document to its patient, encounter, source and version.
On a narrow screen, scroll the table horizontally to read all columns.
| Package component | Expected sample scope | Acceptance question |
|---|---|---|
| Patient dataset | 20 distinct clinic patient references | Are identifiers stable and distinguishable even when names or phone numbers repeat? |
| Encounter dataset | 45 encounters linked to those patients | Does each encounter resolve to the intended patient and preserve its recorded date? |
| Document index and files | 12 attachment entries and their 12 delivered files | Does each entry resolve to a readable file associated with the correct context? |
| Data dictionary | Definitions for fields, code values, units and missing values | Can the receiver interpret the data without guessing at labels or numeric status codes? |
| Snapshot and exception record | Cut-off time, selected scope and known exclusions | What was deliberately omitted, and how will later changes be handled? |
For each important field, agree the source value, destination representation and handling when the value is absent or unsupported. For example, an empty allergy field is not automatically “no known allergies”; an absent measurement is not zero. Preserve uncertainty instead of filling gaps merely to satisfy a required destination field. Ask the receiving vendor how it represents an unknown value without creating a clinical assertion.
Keep original clinical time separate from import time. A document authored last year and imported today should not appear as a new consultation performed today. Preserve the supplied time zone or other date context; if it is missing, document the uncertainty rather than silently assuming it. Similarly, retain an original author as source attribution rather than making the importing staff member appear to have performed the consultation.
Where a destination cannot reproduce a source status or an amendment relationship directly, agree a readable representation and a clearly recorded limitation. The clinic should be able to identify the original and corrected versions and understand which is current. A mapping spreadsheet marked “done” is insufficient if the doctor cannot reconstruct that history in the destination.
FHIR provides a standard way to represent and exchange health information, but a format label alone does not establish the completeness of a clinic’s exit package. Ask for the FHIR release, implementation guide or profiles used, the supported record types and the validation results. The NRCeS ABDM implementation guide is based on FHIR R4 and describes health-record exchange artefacts; its scope should not be mistaken for a complete specification of every commercial EMR database.
HL7’s R4 validation guidance distinguishes checks such as structure, values, profile rules and business rules. It also explains that automated methods check computable aspects and have limits. A validator result therefore complements review of actual records; it does not establish by itself that every clinically relevant source item was exported, attributed to the right patient and interpreted correctly by the receiving product.
Use three separate acceptance observations: the package meets the agreed format rules; the expected source content and relationships are present; and the receiving workflow displays and uses that content correctly. If one passes and another fails, retain the distinction. Do not rewrite the conclusion as “FHIR passed, migration complete.”
A single patient with one uncomplicated visit is a weak migration sample. Include repeat encounters, two family members sharing a phone, a corrected prescription, an outside report, regional-script text and an unresolved follow-up task. These cases test identity, chronology, history, attachments and workflow state.
First compare the exported package with the source. Then inspect what the receiving system produces. Both checks matter: a complete export can still be imported incorrectly, and a polished destination screen can conceal omitted source records. Agree who investigates each kind of discrepancy.
Use fictional records for a sales demonstration. Any later real migration needs the clinic’s authorized access, handling and validation process. Do not send a full patient archive to a vendor merely to find out whether their import feature exists.
Count the record types agreed in scope and explain differences. A total of 1,000 files tells you little if some are duplicates or none contain the original laboratory attachments. Reconcile patients, encounters and documents separately, then inspect a sample of the links between them.
A hypothetical sample illustrates the problem: the source contains 20 fictional patients, 45 encounters and 12 report attachments. The export has 20 patient rows and 45 encounter rows, but the 12 “attachments” are authenticated URLs. Row counts match; portability is still incomplete because access depends on the old system.
The acceptance record should identify omissions and how they will be handled. “All essential data exported” is too vague unless essential has been defined. If a record type cannot migrate structurally, agree whether a readable archive is acceptable for that part of the history and how it will be found.
On a narrow screen, scroll the table horizontally to read all columns.
| Check | Evidence to keep | Failure example |
|---|---|---|
| Coverage | Counts by agreed record type and period, with explained exclusions. | Old encounters missing despite the patient count matching. |
| Identity and relationships | Sample links from patient to encounter to document. | Two relatives merged because a shared phone was treated as a unique key. |
| Content and chronology | Source-to-destination comparison of selected records. | Upload dates replacing test dates or draft notes becoming signed documents. |
| Attachments | Actual files opened outside the old authenticated session. | Export contains expiring links rather than portable documents. |
| Receiving-system result | Observed import output and a resolved discrepancy list. | The file parses successfully but medicine status or units lose their meaning. |
Continue the fictional 12-attachment sample. All 12 index entries resolve to delivered files, and every file matches the checksum supplied for it. During review, documents D04 and D09 are linked to the wrong patients and D12 cannot be read. These are three distinct affected documents, so nine of the 12 pass this sample’s combined linkage-and-readability check. File presence and byte matching alone would have missed all three problems.
A checksum compares bytes with a reference value. If an unreadable or incorrectly associated source file was exported unchanged, its checksum can still match. The check does not establish clinical accuracy, correct ownership or trustworthy provenance by itself. Keep the reference manifest controlled and inspect meaning and relationships independently.
On a narrow screen, scroll the table horizontally to read all columns.
| Check | Observed result | What remains unresolved |
|---|---|---|
| Entry and file presence | 12 of 12 present | Presence does not establish readability or the patient relationship. |
| Byte comparison against the supplied manifest | 12 of 12 match | The source bytes may already contain an unreadable file or accompany a wrong index link. |
| Patient/encounter association | D04 and D09 fail; 10 of 12 pass | Correct the two relationships and review the affected charts. |
| Readability | D12 fails; 11 of 12 pass | Investigate the source and delivered file; a replacement must be reviewed. |
| Combined acceptance in this sample | 9 of 12 pass; 3 distinct documents need action | Recheck corrected material and relevant relationships before closing the exception list. |
With fictional records and an agreed test setup, have an authorised reviewer open the delivered package in a clean session that is not signed into the old EMR. This exposes temporary links, embedded dependencies or viewers that still require the old service. It is a portability test, not a reason to cancel the live account or disable access before migration is accepted.
Use the access arrangement the clinic will actually retain. An encrypted archive is not useful if the only decryption route belongs to a departing vendor or employee. Document how authorised staff obtain the required credentials, how access is controlled and who can recover it under the clinic’s approved process. Do not place a password in the same unprotected handover message as the archive.
For an import, ask the receiving team to show a repeat visit, an amended document, a regional-language attachment and an unfinished task where these are included in scope. If a task cannot be migrated as an active workflow, define how it will be represented and reassigned. Preserve the source archive and exception evidence according to the clinic’s applicable arrangements; a successful import is not an automatic deletion instruction.
Define who can request, approve, receive and access the export. Use an appropriate protected delivery and storage process, and record custody of the package. A full-clinic archive should not become a routine email attachment or remain indefinitely in an untracked downloads folder.
Separate the clinic’s retention obligations from the vendor’s operational convenience. Agree what access remains during migration, what happens after the subscription ends and how deletion or retention is handled under the applicable contract and requirements. This article does not set a universal retention period.
If the vendor proposes a third-party transfer service, identify its role and the authorized data path. A migration may require technical support, but that does not justify sharing more information than the approved scope. Keep the specification and recipient clear before any real transfer.
Ask for export charges, delivery time, available formats, assistance with field mapping and the duration of read access after termination. Clarify whether attachments, older versions and more than one export attempt are included. A low subscription fee can be misleading if the clinic later discovers an unplanned exit charge.
State acceptance responsibilities: who supplies the package, who checks source completeness, who validates the destination and what support is available to resolve errors. The receiving vendor cannot guarantee information the source never exports; the source vendor cannot guarantee every importer understands it.
A well-defined exit plan also improves onboarding. It forces the clinic to understand what is being stored, which records are operationally important and how to verify continuity. That is a stronger foundation than relying on a general promise that the data belongs to the clinic.
In a fictional evaluation, the export opens and the patient/encounter totals match. The clinic then logs out of the old system and tries to open the exported report references. They fail because the package contains authenticated links. A second check finds that amended prescriptions have no current-version marker. The clinic records both gaps, requests actual attachments and documented version semantics, and repeats the destination import before accepting the proposed exit process. The lesson is to test independence from the old account, not merely the download itself.
Use these prompts in your clinic’s own notes. Record the result, responsible person and next action for each item.
No. Confirm the agreed release and profiles, source coverage, patient and encounter relationships, attachments and receiving-system behaviour. Automated format validation is one part of acceptance and does not prove that every required source item is present or clinically correct.
No. Ask for the clinic-wide scope, supported record types, identifiers, attachments, delivery method, limits and costs. Test a representative multi-patient package and its reconciliation separately from an individual patient-copy workflow.
It may preserve readable history, but the destination may need structured data for search, encounters and relationships. Define the intended use for each record type and test the receiving product with a representative sample.
No. The schema, field meanings, status values, identifiers and attachments must be understood by the receiving process. Parsing a file is not the same as reconstructing an accurate clinical record.
No. A backup needs an agreed restore process and evidence that the required state can be recovered. Ask separately what the export preserves and what the backup restores.
Only if the agreed access and portability requirements are actually met. Test access independently of the old account and clarify expiry. For a portable archive, the clinic may need the original files rather than temporary references.
Before contracting, alongside setup, migration and renewal costs. Record the scope and limits of the export service so the clinic can evaluate the whole commercial commitment.
Updated on . The examples describe evaluation and contractual questions, not a guaranteed Bractis export format or a universal Indian retention rule. Confirm the supported scope in the proposed configuration and agreement. No independent clinical or legal review is claimed.
Primary reference for structural and semantic interoperability, checked 25 October 2026. The package design and numerical examples are original Bractis analysis.
Primary reference for the R4 basis and health-record exchange scope, checked 25 October 2026. It is not cited as a specification for an entire vendor database export.
Primary explanation of validation dimensions and limits, checked 25 October 2026. The clinic acceptance exercises are original proposals, not an HL7 certification process.
Use 20 demo questions, a clinic brief, supplier evidence and a worked scoring model to compare EMRs for Indian OPDs.
Plan paper-to-EMR migration with a field map, record reconciliation ledger, transition log and clear cutover decisions for Indian clinics.
Compare EMR quotes with first-year and renewal calculations, monthly usage allowances, upfront commitments and invoice checks for Indian clinics.
Ask for a fictional export covering repeat encounters, amendments and attachments, then review the documented formats and the validation needed for your intended destination.
Request a clinic demo