Prepare
Collect the required information and expose gaps before review.
Digital health engineering · Controlled workflows · External systems
Build digital health software that stays clear, traceable and controlled.
Health and pharmacy workflows depend on clear authority. Each action needs a permitted user, a visible state and a defined response when information is incomplete or an external service rejects, delays or cannot process the next step.
Twin Global turns confirmed operational and regulatory requirements into product behaviour that can be reviewed, tested, released and changed without losing control.
Roles and permissionsWorkflow statesTraceable eventsRelease readiness
Responsibility before interface
Health software has little tolerance for unclear responsibility.
Health-related work can pass through preparation, review, authorised decision, external exchange and follow-up. The product must show who owns the current step, what that person is allowed to do and what must happen before the workflow can move forward.
Review
Show source context, prior changes and reasons for correction.
Authorise
Reserve the final action for the permitted role and record the decision.
Exchange or recover
Make external service responses, delays and recovery actions visible to the right team.
Requirements remain a shared responsibility
The client and its advisers confirm the applicable legal, clinical, privacy and regulatory obligations; Twin Global translates them into product rules, access controls, acceptance criteria and test scenarios.
Interactive role explorer
Model roles, states and permissions.
The same record should not create the same experience for every user. Tasks, states, permissions and next actions must reflect each user’s responsibility.
Preparation
Stop incomplete work before it reaches review.
The preparer sees the current task, available information and unresolved gaps without being mistaken for the final decision-maker.
- Task
- Assemble required information
- State
- Draft or incomplete
- Permission
- Create and correct within scope
- Next action
- Submit a complete record for review
Review
Give the reviewer enough context to judge readiness.
The reviewer sees source information, earlier changes and unresolved issues, then returns the item or advances it with a clear reason.
- Task
- Check completeness and context
- State
- Awaiting review
- Permission
- Review, annotate and return
- Next action
- Correct or advance with a reason
Authorised decision
Keep final authority explicit and traceable.
Only the authorised user can approve the final step. The platform records the decision, state change and traceable history linked to the action.
- Task
- Make the authorised decision
- State
- Ready for authority
- Permission
- Approve, reject or return
- Next action
- Trigger the permitted exchange
Exception and recovery
Turn a technical failure into an operational next action.
The support role sees the failed step, provider response and recovery history without receiving broader access than the task requires.
- Task
- Investigate and recover the workflow
- State
- Rejected, delayed or unresolved
- Permission
- Inspect and retry approved actions
- Next action
- Restore or escalate with enough context
External-service integration and exceptions
Design for success, rejection and recovery.
A dependable health platform does not treat the successful path as the whole product. It must prevent avoidable errors, translate external responses into clear workflow states and guide authorised users through correction, re-review or escalation.
Validation before exchange
Stop incomplete or conflicting work before it moves forward.
Required information, role authority and workflow context should be checked before review, approval or external submission. A clear validation state reduces avoidable rejection and keeps the user focused on what must be corrected.
- Identify missing, invalid or conflicting information.
- Prevent an incomplete record from appearing ready.
External response
Turn rejection, delay or unavailability into a visible product state.
The platform should keep the external response linked to the originating task, explain the operational effect and identify who is responsible for the next action. A failed exchange must never look complete.
- Translate external service responses into understandable workflow language.
- Show the reason, owner, status and permitted next step.
Controlled recovery
Correct and resubmit without losing the original context.
Recovery should preserve earlier requests, responses, corrections and decisions. Where the change affects an authorised action, the workflow should return through the required review or approval step before another exchange is attempted.
- Record correction, retry and reconciliation history.
- Escalate unresolved or superseded items with enough context to act.
Controlled information
Keep every important action traceable.
A useful event history should explain who acted, what changed, when it changed and why the workflow moved.
The product should retain who acted, when the action occurred, what changed, which state followed and what an external service returned. That history helps authorised teams investigate a rejected item, understand an amendment and verify which decision or version moved forward.
- Actor and authorityIdentify the responsible user and permitted action.
- State and timeRecord when the workflow moved and why.
- Information changePreserve relevant changes and version context.
- External responseKeep external service outcomes linked to the originating task.
Quality and release preparation
Prepare the whole workflow for controlled use.
Release readiness begins with confirmed requirements, acceptance criteria and representative journeys—not a final round of screen checks.
Testing should cover role boundaries, permitted state transitions, information changes, external services, negative scenarios and recovery paths. Environment configuration, deployment preparation and post-release checks make the release easier to verify and support.
- Requirement traceabilityConnect confirmed requirements to acceptance criteria and test outcomes.
- Role and state testingExercise permitted actions, blocked actions and complete workflow journeys.
- Exception and integration testingTest rejected, delayed, unavailable and corrected external exchanges.
- Release readinessPrepare configuration, migration, deployment checks, monitoring and recovery.
Controlled change after release
Let the product evolve without weakening the controls around it.
Each material change should be assessed against roles, states, integrations, stored records and earlier behaviour. Versioned requirements, regression testing, release approval and a practical recovery plan help teams introduce change deliberately.
- 01Assess impact
- 02Version the requirement
- 03Test the affected journeys
- 04Release, verify and monitor
Australian digital-health delivery in practice
ScriptStream puts controlled workflow engineering into practice.
See the customer story for the full project scope, conformance milestones and current production status.
Australian digital health · Early production
ScriptStream
Twin Global Solutions developed ScriptStream as a specialist-prescribing platform with preparation, review, authorised approval, external exchange and exception handling. The programme involved Healthcare Identifiers, eRx, Australian Digital Health Agency electronic prescribing, PBS authority-related workflows and PRODA-related access arrangements.
View the ScriptStream storyPlan the engineering brief
Turn health-platform requirements into a clear engineering brief.
Discuss the workflow, external services and release requirements for your health platform.
