Multi-Clinic Record Access: Test Isolation and Authorised Sharing
Evaluate record separation across clinics using synthetic identity, search, document and permission tests; distinguish branch sharing from independent organisations.
When a doctor works across clinics, a facility switch must establish the destination clinic, authorised role, patient and encounter before work continues. First resolve or deliberately hand over work in the previous clinic, including uncertain saves. After switching, verify the identity on the clinical record and any document or booking produced. A familiar patient name, a coloured header or access to both clinics does not by itself establish the correct context.
Visiting doctors, associates and clinic managers evaluating multi-location software or organising work across independent practices.
A doctor may see patients in two branches of one organisation, in two independent clinics, or in a host clinic while retaining a separate home practice. These arrangements can require different access and record-sharing rules. A dropdown containing both clinic names does not explain who controls each record or which information may be shared.
Before the first session, identify the organisation and location for each role. Establish which staff can view the doctor's records, where clinical correspondence arrives and who handles pending work between visits. If the same patient attends two practices, follow the agreed sharing and identity process rather than assuming one chart is universally available.
OWASP's multi-tenant guidance emphasises establishing organisation context from verified membership and enforcing it consistently. Applied here, the useful vendor question is whether permissions and record boundaries remain correct throughout a facility change, including background requests. A branch need not be a separate technical tenant.
Before moving to the next location, distinguish completed notes, saved drafts, unsaved edits and saves whose outcome is unknown. Those states require different actions. A draft can remain unfinished if it is stored in the correct encounter and its owner knows what remains; an uncertain save needs reconciliation before someone creates another copy.
Consider an invented sequence: at 09:58 a doctor saves a note at Clinic A; the connection stalls. At 10:00 the doctor selects Clinic B. At 10:00:02 a response to the earlier request arrives. That acknowledgement belongs to Clinic A's encounter. It must not make Clinic B appear to contain a saved note or overwrite the current patient view.
If the interface cannot make the outcome clear, stop the affected action, inspect the original encounter through the supported recovery path and seek support as needed. Do not repeatedly click save or sign in the hope that one attempt works. A screenshot of a spinner is evidence of an uncertain interaction, not evidence that no data reached the server.
Hand over time-sensitive results and patient questions to a named clinician or team under the local arrangement. A visiting doctor being physically elsewhere should not leave reception guessing whether a result can wait. Use the clinic's agreed communication route; clinical urgency remains a clinical decision.
Treat “switch requested” and “destination ready” as different moments. While the destination is loading, a previous clinic's patient list should not be mistaken for the new clinic's list. Check the full clinic name and location, especially when branches share a brand or abbreviated label.
Next verify the patient and current encounter. Shared family phones, repeated names and identical-looking local file numbers are common reasons to inspect more than the search result's first line. The clinic's identity process should work even when a person does not have a smartphone or cannot recall a file number.
WHO's May 2007 patient-identification solution recommends using at least two identifiers and checking that the intended patient receives the intended care. It does not prescribe a universal Indian UHID format or make a phone number a sufficient identity check.
Keep patient identity separate from encounter identity. The correct patient can still have an old visit open, an appointment at another location or a consultation assigned to a different clinician. Review the visible encounter date, status and location before documenting or issuing anything.
A doctor working at both sites may legitimately have different permissions at each. Recheck functions that matter for the session, such as editing, signing, exporting or reviewing another clinician's notes. A successful switch should not silently turn a restricted role into a broad owner role.
Source: WHO Collaborating Centre: Patient Identification, May 2007
A correct-looking workspace is only part of the journey. Printed prescriptions, appointment slips, invoices, referral letters and downloaded files can carry old clinic details from cached templates or a previous session. Use fictional records to inspect the actual output before relying on the setup.
Pay special attention to shared printers and downloaded files. The right document sent to the wrong printer can disclose information. A download named only “prescription.pdf” may be confused with an earlier file; opening and verifying it is more useful than assuming the latest filename is the correct version.
These checks are acceptance questions, not claims about Bractis or any vendor. Ask for the behaviour to be demonstrated in the configuration the clinic intends to use.
On a narrow screen, scroll the table horizontally to read all columns.
| Action | What to verify | A useful failure case |
|---|---|---|
| Search or queue view | Destination clinic and permitted population | Old results remain after the facility label changes |
| Document a visit | Patient, encounter, clinician and clinic association | A draft follows the doctor into a different encounter |
| Create a booking | Clinician availability and appointment location | A Clinic B booking uses Clinic A's address |
| Print or download | Clinic identity, author, patient and output version | A cached header or prior document is reused |
| Record a payment | Correct clinic ledger, patient and underlying charge | A fee or receipt belongs to the previous location |
| Review notifications | Which clinic and record the notification concerns | A late response opens or alters the wrong current view |
With the vendor's agreement, prepare synthetic records at Clinic A and Clinic B. Include a repeated patient name and deliberately similar local numbers. Give a trial user the exact roles intended for the visiting doctor. Do not use real patient records to discover access boundaries.
Open Clinic A in one tab and Clinic B in another. Check whether switching in one tab affects the other, and whether the behaviour is explicit. There is no single acceptable tab model for every product, but users need a predictable one. A hidden global switch that leaves the other tab looking unchanged is a risk to investigate.
Try refresh, browser back, a bookmarked encounter and a session that has expired. If membership is removed, test the next authorised action through an agreed trial procedure; a previously open page should not grant indefinite authority. Preserve synthetic test references and the observed outcome for vendor review.
Include a delayed request or interrupted connection only through a supported test setup. Check where the eventual success, error or retry instruction appears. Finish by reopening the source encounter and the destination encounter independently. A safe result should be explainable from their stored records, not just from a transient banner.
Review the doctor's actual working conditions: travel between sites, a laptop on unreliable mobile data, different reception teams and limited time at session end. Simplify the transition card around those conditions. The goal is a repeatable check at the points where the wrong clinic could change a decision, not an elaborate checklist that nobody can follow.
In the fictional 09:58–10:00:02 sequence, the save was initiated at Clinic A before navigation to Clinic B. The doctor needs evidence of the stored outcome at A and a clean, separately verified encounter at B. The time a response arrives does not change the clinic or patient to which the original action belongs.
Use these prompts in your clinic’s own notes. Record the result, responsible person and next action for each item.
No. Access depends on the organisations' authorised arrangements, the user's role and the relevant sharing process. Being the same clinician is not a complete access policy.
First demonstrate the product's tab and session behaviour with synthetic records. Each action must retain the correct clinic and encounter context, including after a switch or refresh.
Establish the stored outcome or use the documented recovery process before repeating the action. Keep the original clinic and encounter reference attached to any unresolved work.
No. Also verify patient, encounter and role, then inspect documents, bookings and other outputs where a stale clinic context could persist.
Treat each number within its issuing clinic or defined namespace. Follow the identity procedure and any authorised linking process; matching local numbers alone do not establish the same person.
Updated on . Workflow and acceptance guidance, not a statement that every facility needs a separate database or that a vendor already implements these behaviours. No independent clinical or legal review is claimed.
Technical separation guidance; clinic branches and independent organisations may have different authorised sharing models.
International patient-safety guidance; not an Indian identifier-format mandate.
Evaluate record separation across clinics using synthetic identity, search, document and permission tests; distinguish branch sharing from independent organisations.
Coordinate visiting consultants with realistic session capacity, records access, patient communication and named responsibility for work after the specialist leaves.
Design clinic patient numbers with stable identity, legacy aliases, duplicate resolution and readable formats; distinguish local UHIDs, visits and ABHA.
Rehearse Clinic A and Clinic B with synthetic patients, different roles and two tabs. Inspect a pending save, a refreshed encounter and downloaded outputs before concluding that switching preserves the right context.
Request a clinic demo