Evaluating Clinic Software Features: Separate Demonstrated, Configured and Planned
Turn vendor feature claims into scoped evidence, exception tests, dependency checks and purchasing decisions using a practical capability record.
Ask a clinic-software vendor to show how access is restricted, how one clinic's information is separated from another's, how usable records are restored and how support staff obtain access. Request scoped evidence, dates, limitations and named owners. Encryption, row-level security or a compliance badge can describe part of the system; none alone proves that the complete clinic workflow is secure.
Indian clinic owners and doctors evaluating cloud software, renewal terms or a move from an existing records system.
Start with the information and work the clinic cannot afford to expose or lose: patient records, attachments, account access, clinical attribution, receipts and unresolved tasks. Ask who can read, change, export or administer each category. A general promise that data is 'safe in the cloud' does not answer those questions.
Separate the vendor's responsibilities from the clinic's. The vendor may operate the service while the clinic controls staff accounts, shared devices, local exports and access decisions. A secure hosting arrangement cannot compensate for a doctor sharing an unrestricted password with everyone at reception.
Request an evidence summary that is useful to a non-specialist: what was assessed, the version or environment, when it was assessed, major limitations and who handles unresolved issues. Sensitive security reports may need controlled review rather than public disclosure. A refusal to publish full technical details is not itself proof of a flaw, but a buyer still needs adequate evidence for its decision.
OWASP distinguishes authentication from authorisation and recommends checking permissions on each request. A successful login proves identity only within the system's mechanism; it does not establish permission to every patient record or administrative action. Ask for a demonstration using the actual intended role rather than the vendor's unrestricted account.
Use an agreed, authorised test environment with synthetic data. A receptionist should be able to complete permitted desk work while restricted clinical or administrative actions remain controlled. A visiting doctor should have the intended branch and patient access. Inspect the result of a prohibited action, not merely whether the interface hides its button.
Include record attachments, exports and old sessions in the discussion. A correctly restricted patient screen is not enough if a previously generated download remains available beyond its intended access. Ask how role changes, staff departures and lost devices affect active sessions and retained local files.
Keep the review within the agreed scope. Do not probe another real clinic's records or run intrusive tests on a production service without authorisation. A qualified security assessment can cover risks that a purchasing demonstration cannot; the clinic walkthrough is one layer of evidence, not a penetration-test substitute.
A vendor may use separate databases, schemas or shared tables with row-level policies. OWASP describes trade-offs and the need to test isolation through the application's actual access path. The architecture name alone does not establish the deployed boundary. In particular, a database control can be ineffective for requests using a role that bypasses it.
For the buyer, ask for a simple two-clinic synthetic example and the assessed scope. Are searches, attachments, exports and background operations restricted to the intended clinic? Does switching a facility change only the permitted context, or could a stale screen expose the wrong information? Ask which parts were covered by independent testing and which remain outside its scope.
Distinguish tenants from branches. Two locations of one organisation may intentionally share some information, while two unrelated clinics must not gain access through the same mechanism. The vendor needs to understand the clinic's organisational arrangement; a generic demonstration of two different logos does not settle the data boundary.
Ask how privileged support and maintenance paths are controlled. A system can enforce ordinary staff access while allowing broader vendor operations for legitimate maintenance. The relevant questions are authorisation, purpose, scope, time limits where applicable and evidence of what was done.
NIST's small-business guidance recommends protecting and testing backups. Ask what is backed up, how recovery is tested and whether the restored result includes the records and attachments the clinic needs. A nightly job marked successful is not the same as evidence that a clinic can resume work from it.
Separate the recoverable data point from the time taken to resume service. In a fictional incident, the usable copy represents data at 08:00 and the service fails at 10:30. There may be a 2.5-hour data gap to reconcile. A technical restore finishes at 12:00, but the clinic cannot resume until validation and access checks finish at 13:00. Technical restoration took 90 minutes after failure; operational interruption lasted 150 minutes.
Ask how work created after the recoverable point is identified and reconciled, including paper notes or other authorised fallback records. Do not assume a restore automatically reconstructs missing transactions or that reopening the website establishes clinical completeness. The clinic and vendor need a defined acceptance and recovery owner.
Retention, deletion and legal preservation need a separate, applicable policy. Do not accept a generic retention number as a complete answer for every record type and circumstance. Establish what the service stores, what the clinic must retain and how deletion requests, contractual exit and backups are handled under the relevant requirements.
On a narrow screen, scroll the table horizontally to read all columns.
| Concern | Useful evidence | Limit to keep visible |
|---|---|---|
| Staff access | Allowed and blocked synthetic tasks under intended roles. | A role name or hidden button is not proof of enforcement. |
| Clinic separation | Scoped test coverage across records, files, exports and operational paths. | One database setting does not prove the entire boundary. |
| Backup recovery | Dated restore exercise, included data and business-readiness result. | A successful backup job is not a demonstrated restore. |
| Support access | Authorisation process, attributable activity and review arrangements. | Necessary maintenance access still needs clear control. |
| Incident response | Named contacts, detection and communication responsibilities. | A generic response promise may omit the clinic's own actions. |
| Data exit | Opened sample export with documents, identifiers and relationships checked. | An export button does not prove a complete usable handover. |
Identify the clinic's contact for an incident and the vendor's route for urgent security concerns. Agree what information each side needs and how updates will be provided. Do not invent a legal notification deadline from a sales brochure; applicable requirements and the actual circumstances need appropriate review.
Ask who operates important dependencies and how responsibility is divided. A hosting provider, messaging service or payment integration may affect part of the workflow without taking responsibility for the entire clinic record. The contract and support model should make the boundaries understandable, including relevant service changes.
Test a sample data exit using synthetic records. Open the files, inspect attachments and determine whether the receiving system could understand dates, authorship and relationships. Separate a patient's readable document from a practice's structured migration needs. An export that looks complete in a folder may still omit linked information.
NIST's SSDF can support a discussion about secure development and maintenance practices with the supplier. It is not a certificate that the product meets Indian legal requirements. Ask how vulnerabilities are handled and what evidence is available, while keeping functional export and recovery acceptance separately testable.
Source: NIST SP 800-218: Secure Software Development Framework 1.1
Distinguish missing evidence from a demonstrated failure. If the vendor has not shown a restore, mark recovery unassessed rather than claiming the backups do not work. If an authorised synthetic test exposes the wrong clinic's data, record the failure and require an appropriate response before relying on that boundary.
Prioritise by consequence. A reassuring logo or a long list of controls should not offset an unresolved essential access or recovery issue. Decide what evidence is necessary before adopting the service, what can be resolved during a controlled implementation and what requires specialist advice.
Review responsibilities after staff changes, additional branches, new integrations and major service changes. Keep clinic devices and accounts maintained, and test the relevant recovery arrangements at a frequency appropriate to the risk and agreed plan. Security is ongoing work shared across people and systems, not a permanent property obtained when a subscription is purchased.
A fictional vendor restores the application by 12:00 after a 10:30 failure. The recoverable copy represents 08:00 data. The clinic still needs to confirm record and attachment availability, restore appropriate staff access and identify work from the intervening 2.5 hours. Those checks finish at 13:00. Reporting only a 90-minute technical restore hides a 150-minute operational interruption and a separate data-reconciliation problem. The buyer therefore asks for both the recoverable point and the business-readiness criteria, with named owners. These times illustrate the distinction and are not a service-level promise or a recommended recovery target.
Use these prompts in your clinic’s own notes. Record the result, responsible person and next action for each item.
No. It can be an important control, but effectiveness depends on configuration, actual request roles and the wider application paths. Ask for scoped evidence covering the deployed boundary, including files and exports, rather than relying on the feature name.
Encryption addresses part of the problem. Appropriate access, account protection, recovery, support controls, device handling and usable exit arrangements also matter. Ask what each claim covers and what remains untested.
Ask what data is included, the recoverable point, protected storage, restore-test evidence and the steps needed to resume work. A successful backup schedule does not establish a complete or timely recovery.
Use only an agreed, authorised environment with synthetic clinics and records. Do not probe real third-party data. A purchasing walkthrough can support the decision but does not replace a properly scoped professional security assessment.
Review the exact certification or assessment scope, validity and product coverage, then assess the requirements relevant to your clinic. A generic badge, hosting location or technical framework is not a complete legal-compliance determination.
Updated on . This is a purchasing evidence framework, not a security audit, penetration test, compliance certification or statement about any vendor's deployed controls. Recovery figures are fictional. No independent clinical or legal review is claimed.
Technical guidance on least privilege, default denial and request-level checks; not an Indian legal compliance certificate.
Isolation includes database, storage, cache and operational paths; architectural labels alone are not proof.
US technical risk guidance, updated August 2026, including protected and tested backups; not Indian law.
February 2022 final framework and supplier discussion vocabulary; it does not certify a vendor or define the editorial maturity labels below.
Turn vendor feature claims into scoped evidence, exception tests, dependency checks and purchasing decisions using a practical capability record.
Evaluate a second clinic with incremental economics, branch-specific schedules and billing, controlled record access and a reversible opening trial.
Evaluate EMR recovery using a worked encounter ledger, attachment checks, usable-service milestones and reconciliation of work after a migration cutoff.
Ask Bractis for a scoped security discussion and authorised synthetic demonstrations of role boundaries, clinic separation, restore evidence and usable export. Record what was actually shown and what needs specialist review.
Request a clinic demo