What Is Care Leakage? Definition, Causes, and How to Stop It (2026)
Care leakage is the patients and ordered care that fall out of the loop between visits. A surveillance interval opens and no one calls. An echo, MRI, lab panel, or colonoscopy is ordered and never scheduled. A referral is accepted and never booked. A post-op follow-up lapses. A care program the patient qualified for is never enrolled. Each is a decision the physician already made that the practice never carried through. Care leakage is different from patient leakage, which is patients choosing another practice, and from referral leakage, which is inbound referrals that never convert. In specialty medicine the gap is wider than in primary care because the work between visits is longer, interval-based, and split across the endoscopy scheduler, the imaging center, the surgery coordinator, and the front desk.
This guide covers a working definition of care leakage, where it happens in orthopedics, cardiology, gastroenterology, urology, ophthalmology, and dental practices, how to measure it in your own records, and how patient activation agents close the loop.
Key Takeaways
- Care leakage is care the practice already decided on that never happened, not new demand the practice failed to win.
- In a 2023 JAMA Network Open analysis, only 58.4% of high-risk tests and referrals ordered at in-person visits were completed on time, and 42.6% after telehealth visits.
- Interval-based care leaks worst: 45.8% of post-polypectomy patients in one cohort had late or missing surveillance colonoscopy, and about half of glaucoma patients lapse from care for more than a year.
- The fix is an owner, not a reminder: something must scan the records daily for open loops, reach the patient, book the slot, and report that the visit happened.
- 82% of patients try to book care outside a practice's regular office hours, so the outreach has to answer back at night and on weekends or the loop stays open.
The Data Behind the Decision
The clearest evidence comes from loop-closure research. A Harvard-affiliated primary care network tracked 4,133 orders for colonoscopy, dermatology referral for suspicious lesions, and cardiac stress testing. Only 58.4% of orders placed at in-person visits and 42.6% placed at telehealth visits were completed within the recommended window. Colonoscopy referrals placed in person closed 56.9% of the time and cardiac stress tests 63.2% (JAMA Network Open, 2023).
Interval care leaks even more. When a Singapore center audited 419 patients after colonic polypectomy, 13.8% had their surveillance colonoscopy on schedule, while 45.8% were late or never returned (Journal of Gastrointestinal Oncology, 2019). In ophthalmology, a 2026 study of a glaucoma clinic cited IRIS Registry data showing that half of glaucoma patients experience a lapse in care longer than a year, and found that a single patient-portal message re-engaged only 12.5% of lapsed patients within 30 days (Ophthalmology Glaucoma, 2026).
Referred care leaks at the handoff. Across 103,737 referral scheduling attempts to 20 specialties in one large health system, 38.9% never got an appointment date and only 34.8% ended in a documented completed visit (Journal of General Internal Medicine, 2018). 82% of patients try to book care outside a practice's regular office hours, which is exactly when the scheduler who owns the open order has gone home.
What counts as care leakage, and what does not?
Care leakage has a narrow definition on purpose: care your practice already ordered, recommended, or scheduled that did not happen on time. The five common forms are a missed surveillance recall, an ordered diagnostic that was never scheduled, an accepted referral that was never booked, a follow-up visit that lapsed, and a care program (chronic care management, remote monitoring, a post-surgical outcomes program) the patient qualified for but never joined.
Two neighbors get confused with it. Patient leakage is a patient who leaves your practice for a competitor or a health system, often after a referral out for imaging or therapy. Referral leakage is inbound: a primary care physician sends you a patient and that patient never becomes a booked visit. Care leakage is different because the clinical decision has already been made and documented. It shows up in quality measures and bundle-era engagement scores, and it is invisible on the schedule, because an appointment that was never created never becomes a no-show.
How is care leakage different across specialties?
Orthopedics. The loop is post-operative. A total joint patient is supposed to be seen at two weeks, six weeks, three months, one year, and then at implant intervals. Add the ordered MRI never scheduled and the injection series that stopped at two of three. Most groups run Epic, athenahealth, or ModMed, and the surgery scheduler is a different person from the clinic scheduler, so the open order has no single owner.
Cardiology. The loop is diagnostic. Echocardiograms, Holter and event monitors, stress tests, and lipid panels are ordered at the visit and scheduled later, often by a separate imaging desk. Heart failure patients are supposed to be seen within 7 to 14 days of discharge, and post-discharge visits are the single most leaked appointment in the specialty. Practices run Epic, Oracle Health, or athenahealth, and the volume is heavily Medicare, where a missed echo becomes a missed quality measure.
Gastroenterology and urology. The loop is interval surveillance: colonoscopy at one, three, five, or ten years depending on pathology, PSA and confirmatory biopsy on active surveillance, cystoscopy after bladder tumor resection. The order lives in a pathology letter or a recall field in gGastro or ModMed that no one queries. Ophthalmology and dental. The loop is chronic monitoring: pressure checks and visual fields for glaucoma, diabetic retinal exams, six-month hygiene recall and periodontal maintenance. Published dental cohorts show most patients miss their individualized recall interval entirely.
How do you measure care leakage in your own records?
Start with four queries your EHR can already run: orders older than 30 days with no linked appointment (echo, MRI, labs, colonoscopy, PT); patients whose recall date passed 30 days ago with no future visit; accepted inbound referrals with no appointment after 14 days; and procedure patients whose next planned follow-up was never created. Count patients, not tasks, and put a name next to each list. If no one owns a list, that list is your leakage rate.
Then measure recovery, not just exposure. For every open loop, track four states: found, reached, booked, seen. Most practices can report found. Almost none can report seen, and that reporting gap is why care leakage persists in well-run groups.
How to Stop Care Leakage in a Specialty Practice: 6 Steps
- Define the loops you will own. Write down the five to ten order types and recall intervals that matter clinically and financially for your specialty. In cardiology that is echo, stress, monitor, and post-discharge visit. In GI it is surveillance colonoscopy.
- Pull the open-loop lists weekly, then daily. Query the EHR for each loop type with no linked future appointment. The first pull is the backlog, and it is recoverable revenue and quality credit.
- Assign one owner per list. A person or an agent must be accountable for every open loop until it is booked or documented as declined.
- Reach patients in a two-way thread, not a blast. Reminders go out. Outreach comes back. Use phone, text, and email in one conversation so the patient can answer at 9 p.m. and get booked.
- Book into real slots during the outreach. The conversion happens when the patient is holding the phone. Outreach that ends with instructions instead of an appointment leaks again.
- Report found, reached, booked, seen. Close the loop on your own process. Reconcile booked against completed visits monthly and treat the seen rate as the number that matters.
What Should Still Go to a Human?
Whether an overdue surveillance interval can be extended for a stable patient, whether a symptom reported during outreach changes the urgency of an ordered test, and whether a patient who declines a colonoscopy needs a different conversation with the gastroenterologist are clinical judgments. Billing disputes, prior authorization denials, and any patient who is upset about a missed result should go straight to a named staff member with the full record open. Automation should find and route these moments, not decide them.
Where Clinekt Fits
Clinekt is the patient activation platform for specialty practices: four AI agents that share one memory of every patient and cover the journey from first click to the care between visits. Care leakage is the job of the Recall Agent. It continuously scans the practice's records for overdue follow-ups, lapsed care plans, surveillance intervals that opened, ordered-but-never-scheduled visits, and dormant charts. It reaches each patient by phone, text, and email in one two-way thread, in the practice's voice, books them into open slots, and reports found, reached, booked, and seen.
The other agents close the rest of the loop. The Inbound Agent answers the patient who texts back at midnight. The Care Management Agent runs the monthly check-ins for programs the patient was never enrolled in. Clinekt OS is the one memory behind all four, so the patient who arrived through the website in spring is recognized when they text in winter. Baldwin Bone & Joint, an orthopedic group, produced 263 qualified surgical leads and 159 booked appointments, a 60% booking rate, in a single quarter; the same find, reach, book mechanism closes an open echo order or a lapsed PSA. Deployment is live same day, syncs to the EHR, and is HIPAA compliant and SOC 2 Type II. See the platform and orthopedics pages.
Run your own numbers in the leakage calculator or book a demo to see the agents work against your records.
Frequently Asked Questions
What is the difference between care leakage and patient leakage?
Patient leakage is a patient who leaves your practice for another provider, often after being referred out. Care leakage is care your practice already ordered or recommended that never happened, whether or not the patient stayed. A patient can be fully loyal and still leak an ordered echo.
Is care leakage a revenue problem or a quality problem?
Both. Every unscheduled diagnostic, missed surveillance study, and lapsed follow-up is a visit or procedure that did not bill. It is also a quality measure not met, a bundle engagement score not earned, and in the worst case an abnormal finding that sat unworked.
Why do appointment reminders not fix care leakage?
Reminders only fire for appointments that exist. Most care leakage is an order or interval that never became an appointment in the first place, so there is nothing to remind about. You need something that scans the records for open loops and creates the appointment.
Which specialties have the most care leakage?
Any specialty with long intervals and ordered diagnostics. Surveillance colonoscopy in GI, glaucoma monitoring in ophthalmology, active surveillance in urology, post-discharge follow-up in cardiology, and implant follow-up in orthopedics all show published loss-to-follow-up rates near or above half.
How fast can a practice start recovering leaked care?
The backlog exists today in your EHR. Once the open-loop queries have an owner, the first recovered bookings come within days. Clinekt's Recall Agent goes live the same day, with no IT project and no new staff workflow.
Care leakage is already in your records, and it is recoverable. Estimate what it is costing your practice with the leakage calculator, then book a demo and watch the Recall Agent find the open loops in your own data.