Written by Zain Qayyum,
Continuity of care is usually taught as a clinical idea. The same clinician, or at least the same care team, follows a patient over time so that nothing is forgotten and no decision is made without context.
In day-to-day practice, continuity tends to fail for nonclinical reasons. It fails because a document did not move, a phone call was not made, or a decision sat in a queue that nobody owned. The clinical judgment was sound. The administrative follow-through was not.
Nowhere is this clearer than at the referral. A referral is the moment a patient passes from one organization to another, and it is the point where continuity is most exposed. For case managers, care coordinators, nurse navigators and anyone responsible for a patient across settings, understanding exactly where that hand-off breaks is the difference between a referral that becomes an appointment and one that quietly disappears.
Where Continuity of Care Actually Breaks
A referral looks like a transfer. It behaves like a hand-off.
In a transfer, responsibility moves cleanly from one party to another, and the receiving party confirms they have it. In a hand-off between separate organizations, with separate systems, separate staff, and separate priorities, there is a window where nobody is certain who holds the patient. That window is where continuity is lost.
The scale of this is well documented. A retrospective cohort study of 38,334 specialty referrals to a children’s hospital, published in The Journal of Pediatrics in 2023, found that 62% of referrals resulted in an appointment being scheduled and 54% resulted in a completed visit. Nearly half of referred patients never arrived. The study also found completion was lower for Black patients, Native Hawaiian and Pacific Islander patients, Spanish-speaking families and publicly insured patients, which tells us these gaps do not fall evenly.
The communication failure runs in both directions. The landmark Milbank Quarterly review of specialty referrals in the United States described breakdowns at every stage of the referral process, and noted that primary care physicians frequently do not know whether a patient actually attended the specialist, or what the specialist recommended. A referral can be clinically appropriate, accepted, attended, and still fail as an episode of coordinated care, simply because the information never returned to the person who initiated it.
The Six Stages a Referral Passes Through
Most discussion of referral management treats it as a single event. It is more useful to treat it as a sequence, because the failures cluster in specific places.
A referral typically moves through six stages:
- The referral arrives, usually as a document.
- Someone identifies what it is and extracts the patient and referring provider.
- Someone checks insurance and any authorization requirements.
- Someone decides which provider and which appointment type it needs.
- Someone contacts the patient and books, or records that booking is not possible.
- A report goes back to the referring clinician.
Stages That Fail Loudly
Stages one through three fail in visible ways. A document arrives in the wrong place. A name is misread. An authorization requirement is missed, and the claim is denied later.
These failures are frustrating, but they are recoverable, because something or someone eventually notices. A denied claim generates a task. A misfiled document gets found when somebody goes looking for it. The system produces a signal.
Stages That Fail Silently
Stages four and five are different. These are the stages that require a judgment: which provider is right for this patient, what appointment type this warrants, what to do when the patient cannot be reached or cannot be accommodated.
When a judgment has no clearly assigned owner, the referral simply waits. It has been received correctly. It has been read correctly. Nothing about it is wrong. It is sitting in a queue, complete and accurate, waiting for a decision that nobody has been made responsible for making.
Why Silent Failures Are the Hardest to Catch
This is the central problem in referral coordination, and it is worth stating plainly: a referral waiting on a judgment produces no error state.
Every quality process we build is designed to catch wrong things. Exception reports, denial queues, incomplete-record flags. A referral stuck at the provider-matching stage triggers none of them, because by every available measure it is in good order.
The practical test for whether this is happening in your organization takes about a minute. Ask how many referrals arrived last month, and you will get a precise number. Ask how many of those patients were never contacted, and in most organizations you will get an estimate or a pause. The first question counts documents, which systems do well. The second requires knowing what should have happened next, which is a coordination question, not a record-keeping one.
If the second number is not readily available, referrals are almost certainly being lost, and nobody will be able to say how many.
What Care Coordinators Can Measure
Three measures surface silent failure better than referral volume does.
Time From Arrival to First Patient Contact
Not time to filing, and not time to processing. Time from the referral arriving to the moment someone actually speaks to or reaches the patient.
Time to filing is the measure most organizations can produce easily, and it is the one that flatters them. A referral can be filed within an hour and sit unactioned for three weeks. Time to first contact is what the patient experiences, and it is the only one of the two that correlates with whether they eventually attend.
Referrals Received Against Referrals Contacted
A simple ratio, calculated monthly, from two numbers that most organizations hold separately and rarely compare. The gap between them is your leakage, and seeing it as a single figure tends to change the conversation faster than any amount of process mapping.
Look at referrals from two months ago rather than last month, so that anything genuinely still in progress has resolved.
Unbookable Referrals, Counted With Reasons
Some referrals cannot be booked. The patient declines, moves away, is not appropriate for the service, or cannot be reached after reasonable attempts. That is legitimate.
What is not legitimate is for those referrals to leave no record. If a patient cannot be booked, that should be an explicit status with a reason attached, and it should generate a message back to the referring office. Counting these monthly tells you something useful about demand, access and referral quality. Leaving them uncounted means a genuine inability to serve a patient is indistinguishable from an administrative failure to try.
Closing the Loop Back to the Referring Clinician
An appointment attended is not a closed loop. The loop closes when information returns to the clinician who initiated the referral.
This is formalized in CMS Quality ID #374, “Closing the Referral Loop: Receipt of Specialist Report,” which sits in the Effective Communication and Care Coordination domain. The measure holds the referring clinician accountable and asks a single question: did the referring clinician receive a report back from the clinician they referred to?
It is worth noticing what that measure does not ask. It does not ask whether the referral was sent, whether it was received, or whether the patient attended. It asks whether information came back. That reflects something coordinators understand instinctively: a referral where the patient was seen but the referring clinician learned nothing has produced a visit without producing continuity.
For referring offices, the practical consequence is that the single most valuable thing a specialist can send is not the consult note weeks later. It is a short acknowledgment on the day of receipt, confirming the referral arrived and will be actioned, followed by confirmation of what was booked or why it could not be. Those two messages cost almost nothing and resolve the uncertainty that otherwise drives repeat calls, duplicate referrals and patients falling between two organizations each assuming the other holds them.
Both are also the easiest part of the loop to automate, since neither message needs a person to compose it. Sending them reliably is standard in referral workflow tools such as Medsender.
Practical Steps for Care Coordination Teams
Separate the Decision Queue From the Processing Queue
Referrals awaiting a judgment behave completely differently from referrals awaiting processing, and mixing them in one list hides the first behind the second.
Maintain them separately, and make sure the decision list sits in front of the person authorized to make that decision. The volume is usually small. The delay is usually large. It is one of the few changes in this area that costs nothing and shows a difference within a month.
Make the Unbookable Message Standard
Agree on a short, standard message that goes back to the referring office whenever a patient cannot be booked, with the reason. Most organizations have no such message, and the default is silence.
Silence is read by the referring office as the referral having been received and handled. It is the single most common reason a patient is lost while two organizations each believe the other is responsible.
Agree Who Owns the Patient Between Organizations
The window between referral sent and appointment booked is where ownership is genuinely ambiguous. Most referral agreements specify clinical criteria and say nothing about this window.
Write it down. State how long the receiving organization has to make contact, what happens after a set number of failed contact attempts, and at what point responsibility returns to the referring clinician. This is not a technology question, and organizations that answer it clearly lose fewer patients than those with better systems and no agreement.
Conclusion: Continuity Is an Administrative Discipline
Continuity of care is often described as a quality of the clinical relationship. The evidence suggests it is at least as much a property of the administrative process surrounding that relationship.
Nearly half of referred patients in a large cohort study never completed the visit. Referring clinicians routinely do not find out what happened to the patients they sent. Neither of these is a failure of clinical judgment, and neither will be solved by clinical training.
For case managers, care coordinators and anyone holding responsibility for a patient across settings, the useful work is unglamorous and specific. Find the stages where a decision has no owner. Measure contact rather than filing. Say something back when a patient cannot be booked. Agree in writing who holds the patient in the gap.
The referrals that get lost are rarely the complicated ones. They are the ordinary ones, correctly received and accurately recorded, waiting for a decision nobody was asked to make.
Author Bio: Zain Qayyum is co-founder and co-CEO of Medsender, a healthcare workflow automation company. He was studying pre-med and interning in a hospital emergency department when he first saw how much of patient care depends on paperwork moving correctly, and he left pre-med for computer science to work on the problem. He led development of Medsender’s first AI models and the prototype of its voice agent, and writes about referral workflow, document intake and front office operations in specialty practices.
References
Centers for Medicare & Medicaid Services. (n.d.). Quality ID #374: Closing the referral loop: Receipt of specialist report. Quality Payment Program. https://qpp.cms.gov
McDonald, K. M., Schultz, E., Albin, L., Pineda, N., Lonhart, J., Sundaram, V., Smith-Spangler, C., Brustrom, J., Malcolm, E., Rohn, L., & Davies, S. (2014). Care coordination measures atlas update (AHRQ Publication No. 14-0037-EF). Agency for Healthcare Research and Quality. https://www.ahrq.gov/ncepcr/care/coordination/atlas.html
Mehrotra, A., Forrest, C. B., & Lin, C. Y. (2011). Dropping the baton: Specialty referrals in the United States. The Milbank Quarterly, 89(1), 39–68. https://doi.org/10.1111/j.1468-0009.2011.00619.x
Srinivasan, A. P., Sullivan, E., Desai, A., Adiele, A., Bell, S., & Lion, K. C. (2023). Disparities in pediatric specialty referral scheduling and completion. The Journal of Pediatrics, 260, 113485. https://doi.org/10.1016/j.jpeds.2023.113485
