A referral queue can show plenty of activity and still leave an important question unanswered: what happened next? Calls made, messages sent, and appointments requested describe work performed. They do not, by themselves, show whether a referral was received, whether the patient was seen, or whether information returned to the referring practice.
A useful referral scorecard separates workload, process movement, patient access, and information return. This framework is an operational starting point, not a universal benchmark or a substitute for clinical quality measures. Practices should define each measure around their own workflow, systems, and responsibilities.
Start by defining what “closed” means
Referral workflows often contain several distinct milestones:
- Sent: the referral and required information were transmitted through the approved channel.
- Received: the receiving organization confirmed receipt, when that confirmation is available.
- Scheduling outcome: the appointment is scheduled, declined, deferred, or still pending.
- Visit outcome: the appointment occurred, was missed, or was canceled. “Scheduled” is not the same as “seen.”
- Information returned: the referring practice received the specialist’s report or other agreed documentation.
- Clinical follow-up: an authorized clinician reviews and acts on information that requires clinical judgment.
These milestones answer different questions. A practice may choose one endpoint for an administrative queue and another for a clinical quality measure. State the endpoint plainly so that a percentage cannot be mistaken for more than it represents.
Use a small scorecard with clear denominators
Four views are usually more useful than one blended “completion rate.” The examples below are proposed operating measures; choose only the ones that support a decision your team can make.
1. Queue coverage
Formula: open referrals with a current owner, status, next action, and review date ÷ all open referrals in the defined queue.
This checks whether work is visible and actionable. It does not measure whether patients obtained care. Agree on what counts as “current” and exclude records that are outside the queue’s scope.
2. Aging by current status
Report the number of open referrals in each status and how long they have remained there. For example, separate “awaiting receiving-office response” from “patient has not scheduled” and “report not yet received.” Use a locally chosen review interval and flag exceptions that need escalation under the practice’s policy.
A single average age can hide a small number of very old items. A count by age band or a short list of the oldest unresolved items can make the queue easier to review.
3. Access and appointment outcomes
For a defined referral cohort, distinguish referrals with an appointment scheduled, appointments documented as completed, and outcomes that remain unknown. Keep the denominator consistent: specify whether it includes all referrals sent, only those accepted by the receiving organization, or another clearly defined group.
Do not label an unknown outcome as a failure or a success. Record it as unknown and decide who will verify it, through which approved channel, and when.
4. Information returned to the referring practice
Track whether the agreed report or consultation information has arrived for the referral cohort being reviewed. The Agency for Healthcare Research and Quality’s referral guidance recommends tracking referrals, confirming and documenting completion, and obtaining information about the referral result. Tool 21 also offers a sample progress check using records from referrals made about a month earlier; that example is a practice-improvement exercise, not a required benchmark.
The Centers for Medicare & Medicaid Services also maintains a defined quality measure for receipt of a specialist report. Its specifications apply to the eligible clinicians and reporting context described by CMS. A local operations dashboard should not be presented as that measure unless it follows the applicable specifications.
Make comparisons fair enough to act on
Before comparing teams, locations, or months, confirm that they use the same inclusion rules, status definitions, time window, and evidence of completion. Referral mix and receiving-office processes can differ. A percentage without its denominator and observation window is difficult to interpret.
Separate operational exceptions from patient-level detail in leadership reports. Use the organization’s approved systems and access controls for any record-level follow-up; do not place identifiable patient information in general dashboards or marketing tools.
Turn the numbers into a review routine
- Choose one referral type and define when it enters and leaves the queue.
- Assign an owner for each open item and a backup or escalation route.
- Review status, next action, and aging on a set cadence chosen by the practice.
- Record outcomes using agreed categories, including “unknown” and “awaiting information.”
- Sample completed records to check whether the status is supported by the documentation.
- Adjust the workflow when recurring delays or missing handoffs appear; do not assume a metric alone identifies the cause.
The goal is not to maximize calls or push every referral through the same path. It is to make the next step visible, distinguish administrative follow-up from clinical decisions, and show what evidence supports each reported outcome.
Build a scorecard around your actual workflow
A focused workflow review can identify what enters the referral queue, where items wait, who owns exceptions, and what evidence counts as completion. Start with one queue and a small set of measures your team can maintain consistently.
Contact MEDICOSMOS to discuss a defined administrative workflow.
Sources
- Agency for Healthcare Research and Quality. Make Referrals Easy: Tool 21.
- Centers for Medicare & Medicaid Services. Closing the Referral Loop: Receipt of Specialist Report (CMS50v14).
Written and edited by MEDICOSMOS.
Editorial note: Prepared by MEDICOSMOS’ operations team using internal playbooks and assisted drafting tools, and reviewed for accuracy and clarity before publication.
