What Is Referral Leakage? Separate Incomplete Referrals From Network Movement
Referral leakage can describe different events. Define whether you mean incomplete care or care outside a network before choosing a measure, interpreting the result or assigning follow-up.
A patient who completed the intended visit at another organization is not the same as a patient whose referral remains unresolved. A missing result in your EHR does not, on its own, distinguish those situations. The first useful step is to describe the evidence you have and the outcome you do not yet know.
Start with the question behind the word
When someone says "our referrals are leaking," ask what event they observed. Did a receiving office fail to find the referral? Did the patient remain unscheduled? Was a visit completed outside the organization? Or is the problem that no specialist information has returned?
Each question points to different evidence and a different owner. A coordinator can investigate a missing handoff. An analyst may need more complete data to examine network movement. A clinician may need to review returned information. One leakage label should not collapse these responsibilities into a single queue action.
Use plain labels in the working report. "Outcome unknown at the cutoff" is more informative than "lost" when the only evidence is an unresolved record. "Visit documented outside the defined network" describes network movement without deciding that the movement was inappropriate or avoidable.
Separate four observable situations
The table below is a terminology aid. It proposes how to label evidence, rather than a clinical judgment or a universal reporting standard.
| What is known | Clear working label | What it does not establish |
|---|---|---|
| Referral has no confirmed completed visit within the chosen observation period | No documented completion in the measured window | That care never happened anywhere |
| A completed visit is documented outside the defined network | Confirmed outside-network completion | That the patient received unsuitable care or that movement was preventable |
| An appointment is booked but not yet completed | Scheduled, outcome pending | A completed referral |
| The visit is documented but the referring team lacks the expected information | Completed visit, information return unresolved | That the patient did not attend |
This last distinction matters when discussing a closed loop. The IHI/NPSF referral guide considers communication across referral steps, not merely a booking event. A scheduling result and the return of information should therefore remain distinguishable in your local reporting.
Our closed-loop referral management guide covers the broader follow-up process. Use it when the task is establishing ownership and communication, rather than estimating where visits occurred.
Choose an event your data can observe
Before calculating a rate, list the data available. An internal appointment system may show completed visits at your own locations. It may not reveal every visit elsewhere. A referral queue may record an administrative closure without proving that a visit occurred. A claims dataset has its own coverage and observation limits.
Describe those limits beside the result. If your system cannot observe outside visits, do not label every missing internal completion as confirmed outside-network leakage. The record supports an unknown outcome until another source establishes more.
Published research illustrates the importance of the wording. Patel and colleagues' 2018 study examined documented specialist appointment completion in one large health system using an earlier period of EMR data. Its setting and documented endpoint matter. It is not a current national rate for all referral leakage or a measurement of your organization's performance.
Use outside evidence to understand a method or a limitation, then build a denominator that matches your actual data. Resist using a memorable benchmark merely because it makes a business case look clearer.
If your team uses the same label for several different outcomes, book a Linear Health demo and bring the definitions you want the workflow to preserve.
Keep patient choice and unknown information visible
An outside visit may reflect patient preference, an available appointment, location, coverage or another circumstance. Record a reason only when it is documented and appropriate to retain. Do not manufacture a reason from the destination alone.
Likewise, a patient who has not responded to an outreach attempt should not automatically be labeled as having declined care. An attempt, successful contact, stated preference and confirmed appointment are different events. Keeping them distinct improves the accuracy of the administrative account.
Consider a hypothetical referral whose internal appointment record remains empty. Later, the receiving office confirms that the patient attended elsewhere. The correct update is to record the new evidence and its source. The earlier unknown should not be rewritten as though the team had known the outcome all along.
In another hypothetical example, a visit occurred at the intended office, but the expected information has not returned. The relevant unfinished task is information follow-up. Counting that case as both a missed visit and a communication gap would describe the same record inaccurately.
Send each question to its measurement owner
Once the event is defined, use the guide built for that task:
- Referral-to-appointment conversion addresses movement from received referrals to the defined appointment outcome.
- Referral tracking addresses the states and evidence staff need to follow individual records.
- The referral leakage benchmark report addresses comparisons and their limitations.
- The referral leakage cost calculator addresses a separate financial scenario, with assumptions that should not be treated as observed losses.
Keep the definition attached as the number moves between these tasks. A finance worksheet based on incomplete referrals should not later be described as a measured outside-network revenue loss. A conversion rate based on bookings should not become a completion rate in a presentation.
Explain what the number cannot prove
A leakage percentage alone does not identify the cause, the appropriate clinical response, how much work was preventable or what a particular technology will change. It also does not prove that every unknown outcome represents a patient who went without care.
An honest report pairs the number with its definition and unresolved evidence. That may make the result less dramatic, but it gives the team a clearer action: confirm an outcome, repair a handoff, clarify a status or examine a data gap.
Linear Health's referral coordination software belongs in a discussion about the administrative work behind those records. Evaluate it against the specific visibility and follow-up problem you have identified, rather than a leakage label that changes meaning during the conversation.
Name the problem before choosing the fix
A shared definition gives the team a better starting point for action. Book a demo to discuss the referral records and administrative handoffs you want Linear Health to help make visible.
Frequently asked questions
What does referral leakage mean?
Is outside-network care always a failed referral?
How is leakage different from noncompletion?
What data is needed before calculating leakage?
Why can two leakage percentages describe different problems?
Sources
Sources accessed October 20, 2026.