Referral Tracking in Healthcare: A State Dictionary and Worklist Template
Build a referral tracking record with clear states, timestamps, evidence, owners and next actions, including cancellations, reopenings and handoffs.
Define the tracking unit
Begin with one unique referral episode. That episode can have several incoming documents, multiple outreach attempts and more than one appointment occurrence. Keeping those related records separate prevents repeated activity from inflating referral volume.
Retain the source order or referral identifier where available, the receiving or sending organization, the earliest documented receipt and links to supporting records. A patient identifier alone is insufficient because the same person can have more than one unrelated referral.
Do not force every task into a single linear status. A referral can have a booked appointment while an administrative document request remains open. Track the episode's principal state and any parallel tasks so completion of one does not hide unfinished work elsewhere.
HL7's FHIR R4 Task resource separates task status, business status and ownership. The dictionary below is an original operational design; it is not a replacement for implementation-specific terminology or a clinical protocol.
For software-category differences, use the separate referral management versus tracking comparison. This guide focuses on the record and worklist a team needs to maintain.
Use a state dictionary with evidence requirements
Adapt the following to your local workflow. Treat state names as promises about what the record proves. Staff should not need to interpret “in progress” differently depending on who last touched the referral.
| State | Entry evidence | Required next action, then exit evidence |
|---|---|---|
| Received, not reconciled | Transmission or source request received | Match to an episode and validate administrative intake. Exit: identity and episode relationship resolved. |
| Information needed | A specific required administrative item is missing | Request the item from its appropriate owner. Exit: item received or authorized disposition recorded. |
| Ready for assigned action | Required inputs for this administrative step are available | Perform the approved action. Exit: action result recorded. |
| Outreach in progress | Approved attempt made and outcome captured | Continue or escalate under the approved contact workflow. Exit: response, booking or documented disposition. |
| Booking confirmed | Linked appointment confirmation exists | Monitor changes and the next assigned task. Exit: attended, cancelled, rescheduled or other recorded outcome. |
| Administrative follow-up due | Expected outcome evidence or document remains outstanding | Request, reconcile or route it. Exit: evidence received and appropriately routed. |
| Administratively closed | Closure reason and authorized evidence recorded | None unless reopened or another task remains. Exit: reopening event if later work is required. |
Clinical priority, coverage decisions and medical appropriateness are outside this dictionary. Where the workflow depends on one, record the responsible authorized role and the administrative follow-up. Do not let staff infer a clinical decision from the age of a record.
Avoid a catch-all “lost” state unless its meaning is precisely documented. “No confirmed outcome after approved follow-up” communicates uncertainty more accurately than claiming that a patient received no care anywhere.
Make the worklist actionable
A useful row should tell a covering colleague what to do without reconstructing the entire history. Keep enough context to act, with links to the full authorized record rather than copying sensitive detail into uncontrolled notes.
| Field | Why it belongs | Example using fictional data |
|---|---|---|
| Episode ID | Stable reference across systems | REF-1042 |
| Principal state | Current administrative position | Information needed |
| Blocker | Specific reason work cannot advance | Missing receiving-office acknowledgment |
| Next action | Concrete administrative step | Confirm receipt through approved channel |
| Owner and backup | Accountability during absences | Intake team A, coverage team B |
| Next-action due time | Operating follow-up commitment | Local date and time under internal policy |
| Last meaningful event | Evidence of actual progress | Request sent, response pending |
| Source link | Inspectable underlying record | Authorized record reference |
Do not use “follow up” as the entire next-action description. Specify what information or response is needed, from whom and what happens if it does not arrive. A clear next action reduces dependence on the original coordinator's memory.
The inbound and outbound ownership guide helps decide which side of an interoffice handoff owns each step. A message sent to another organization does not automatically mean that organization accepted the work.
The referral intake guide covers reconciliation before the worklist stage. Preserve those intake identifiers so staff can recognize a resend without creating a second episode.
Use more than one clock
Episode age begins with the documented receipt or creation event selected for your workflow. State age begins when the referral enters its current state. Next-action age measures whether the assigned follow-up time has passed. These clocks answer different questions.
If a referral moves from one owner to another, its episode age should not reset. If it enters a new state, the state clock can change while the history remains. A cosmetic edit to a note should not erase time spent waiting for an external response.
Store event time separately from data-entry time when they differ. A coordinator may record an event after it occurred. An integration may deliver a message late. The report extraction timestamp is neither the referral receipt time nor evidence that the referral advanced.
Use approved operational escalation rules to decide what appears first in a worklist. Administrative age can identify overdue follow-up, but it does not determine clinical urgency. Keep the clinical priority supplied by the authorized process visible as a separate field where relevant.
The referral dashboard dictionary explains how these event definitions feed aggregate measures. A dashboard built on unstable clocks can show improving speed merely because records were reassigned or edited.
Walk through a reopening without losing history
Hypothetical example: referral REF-1042 is received on day 1. A missing administrative item is requested on day 2 and received on day 4. A linked appointment is booked on day 5 for day 20. On day 10, the patient requests a different date, and staff cancel the original occurrence while arranging a replacement.
The episode still began on day 1. Its first-booking event remains day 5. Its current state should show that the scheduling path needs an action, rather than continuing to display “booking confirmed” for a cancelled occurrence.
On day 12, a new appointment is confirmed for day 25. This is a second booking transaction associated with the same referral episode. The referral counts once in a unique-episode booking measure. Both appointment occurrences remain in the history for cancellation and scheduling analysis.
If the episode had been administratively closed before the patient responded, record a reopening with its reason and owner. Do not create a new episode simply to make the open-work count convenient. The history should explain why work resumed and which prior evidence still applies.
Distinguish document receipt from clinical review
A returned consult note is an administrative evidence event. Record the document, originating organization, associated visit, received time and routing destination. If the visit link is uncertain, keep that uncertainty in a reconciliation task rather than marking the episode complete by assumption.
The consult-note return guide covers requests and receipt. Your tracking record should not imply that a clinician read or acted on the document merely because a file arrived.
HL7's Appointment resource represents a booking and its statuses. It does not make a local label called “complete” evidence that every downstream documentation or review activity occurred. Keep the specific event visible to avoid collapsing several responsibilities into one word.
If your organization has an approved clinical closure process, link to its recorded decision and keep the appropriate authority. This administrative template should neither replace that process nor create a new clinical completion rule.
Check for hidden tracking failures
Run a short exception review using synthetic cases and a permitted sample of actual records. Look for unowned open tasks, missing next-action times, negative durations, duplicate episodes, booked states without appointment evidence and closed records with outstanding assigned work.
Test a handoff during an absence. Can the backup identify the next step from the record? If the answer requires calling the absent employee, improve the field or reference that failed. Do not respond by adding every possible detail to the worklist; put the right information in the right maintained source.
Reconcile the total open count to the rows staff can see. Check permissions, filters and status mappings before concluding that a record has disappeared. A view that hides another team's tasks should still make the handoff and responsibility understandable.
For a transition to referral coordination automation, use these same checks as acceptance cases. Automation should preserve inspectable episode history and next-action ownership rather than create a second contradictory tracking record.
FAQ
What statuses should a referral tracker include?
Should a referral have one owner or several?
When should an aging clock reset?
Can a spreadsheet serve as a referral tracker?
What does closing a referral mean?
Sources
- HL7 FHIR R4: Task, task states and accountable ownership.
- HL7 FHIR R4: Appointment, appointment representation and distinct booking-related statuses.
