Back to T2 Flex

Reducing Patient Leakage: How Better Access Turns Referrals Into Scheduled Visits

ShareLinkedInEmail

Patient leakage drops when referrals are worked quickly by trained agents who can schedule correctly in the EHR. That is contact center work. T2 Flex runs that function for health systems and handles more than 50 million patient calls a year. The team is organized around the health system's value streams, so the measures reported are referral conversion and scheduled visits rather than call counts. Geisinger brought T2 Flex agents into its internal team to handle a call surge and kept those agents afterward.

What patient leakage is and how to measure it

Patient leakage is the share of patients who are referred for care inside your system and then receive that care somewhere else or not at all. Referral leakage is the same idea applied to a single referral order.

The simplest way to measure it is a four-step funnel.

  1. Referral received. The order is placed in the EHR or arrives from an outside practice.
  2. Patient contacted. Someone reaches the patient and confirms the next step.
  3. Visit scheduled. An appointment is booked with the right provider and the right visit type.
  4. Patient arrived. The visit takes place.

Leakage happens at every step between the first and the last. A referral that is never worked is a loss. A referral that is contacted but never scheduled is a loss. A scheduled visit that becomes a no-show is a loss. Most health systems only report the first and last numbers. The steps in between are where the fix lives.

To calculate a leakage rate, divide referrals that did not result in an arrived visit by total referrals in the period. Track it by service line and by referral source. The number will be different for cardiology than for dermatology, and different for internal referrals than for outside practices.

Where referrals are actually lost

Most leakage is operational. It comes from a few predictable points.

Unanswered calls. A patient calls to book a referred visit and hangs up before an agent answers. If the abandonment rate on a scheduling queue is high, a share of those callers book elsewhere or do not book at all. The same missed call prevents rescheduling after a cancellation or a no-show.

Slow outbound follow-up. Many referrals require the health system to call the patient. If the first contact attempt happens a week after the order, the patient has often moved on. Speed to first contact is the single most important access metric for referral conversion.

Wrong scheduling. The patient is booked with the wrong provider, the wrong visit type, or at a location that does not perform the service. The visit is cancelled or rescheduled and some patients do not come back.

No-shows. A booked visit does not happen. Without a follow-up call to rebook, the referral is lost.

Portal messages not worked. Patients use the portal for messages, history, payment, and test results. Scheduling requests arrive as messages. If those messages sit in a work queue for days, the request goes stale.

Long hold times and poor call handling. Patients who cannot reach the office or wait too long describe the experience to others. That costs the health system future referrals as well as the one on the call. Urgent calls that are not routed correctly are a safety issue as well as an access issue.

Each of these is a question of whether a trained person reached the patient and completed the booking. The sections below cover the levers that fix them.

The access levers that move conversion

The table below lists the operating levers that change referral conversion, the metric that tracks each one, and a working target. Targets should be set against your own baseline.

LeverMetricWorking target
Speed to first contactHours from referral order to first outbound attemptWithin 24 to 48 hours
Answered callsAbandonment rate on scheduling queuesLow single digits
Right the first timeFirst call resolutionResolved on the first contact for most scheduling calls
Specialty scheduling accuracyReschedule and cancellation rate caused by booking errorsTrending down month over month
Outbound persistenceContact attempts per unworked referralThree attempts across phone, text, and portal message
No-show recoveryShare of no-shows rebooked within seven daysMajority rebooked
Portal responseTime to respond to a portal scheduling messageSame business day

Each lever has an owner in the contact center. Speed to first contact depends on staffing for outbound work. Abandonment depends on staffing for inbound peaks. First call resolution and scheduling accuracy depend on agent training in specialty rules and in the EHR. No-show recovery depends on having a team assigned to work that list every day.

What an outsourced access team does differently

An outsourced patient access team is a staffing and operating model. The differences from a stretched in-house team are practical.

Dedicated agents. Agents are assigned to your queues and your service lines. They learn your providers, your locations, and your scheduling rules. The Dedicated vs. Shared Call Center Agents article covers the difference in detail.

Cross-training. Agents are trained across departments so that call surges in one area can be absorbed by agents from another. At Geisinger, T2 Flex provided cross-departmental training that went beyond COVID-related calls. The result was a team that could respond to surges across departments.

Call transcription and quality scoring. Every call is transcribed. Quality and sentiment are scored. Supervisors can see which calls ended without a booking and why. T2 Flex used this voice system at Geisinger for real-time performance monitoring and quality assurance.

Workforce management for peaks. Staffing is planned against call volume by hour and by day. Geisinger's workforce management software from T2 Flex improved resource allocation during regular and peak periods.

Work inside the client's EHR. T2 Flex agents schedule directly in the health system's EHR with credentialed access. That removes the hand-off between a third-party system and your scheduling templates. The Patient Access Outsourcing page describes the integration setup.

Measured on business drivers. T2 Flex runs each engagement in two-week sprints with shared dashboards. The metrics are the ones in the table above, tied to the health system's referral and scheduling value streams rather than to call counts alone.

Case example

The Geisinger engagement shows the operating model at scale. Geisinger serves three million patients across 36 practice sites and three hospitals in northern and central Pennsylvania. During the COVID-19 pandemic, patient calls for information and appointments surged. Geisinger's Patient Contact Center agents were specialized by department and the center needed to scale without lowering its care standards.

T2 Flex recruited and trained remote agents and integrated them into Geisinger's internal team. Agents were cross-trained to work across departments. The voice system provided transcription, sentiment and quality scoring, and call metrics. Geisinger handled the surge while maintaining its standards. After the surge subsided, Geisinger kept the T2 Flex agents as permanent team members. Read the Geisinger case study for the full account.

How to run a 90 day leakage reduction pilot with a partner

A pilot limits risk and produces a number you can act on. The structure below has worked for health systems starting with an outsourced access partner.

Days 1 to 15: baseline. Pick one or two service lines with known leakage. Pull the four-step funnel for the prior 90 days. Record speed to first contact, abandonment, first call resolution, and no-show rate. Use the T2 Flex Agent Calculator to size the team against call volume.

Days 16 to 30: staff and train. The partner recruits or assigns agents to the pilot queues. Agents are trained in your EHR, your scheduling rules for those service lines, and your scripts. Phone routing, callback options, and work queues are set up.

Days 31 to 75: operate. The partner works inbound calls, outbound referral follow-up, portal scheduling messages, and no-show rebooking for the pilot service lines. Every call is monitored and scored. Weekly reporting shows each metric against baseline.

Days 76 to 90: measure and decide. Compare the funnel to baseline. Decide whether to expand to more service lines, adjust the model, or stop.

Write the exit criteria before the pilot starts. A target such as first contact within 48 hours on most referrals is easy to measure and hard to argue with.

Next step

Schedule a Consultation to review your referral funnel with T2 Flex. You can also read more on the Patient Access Outsourcing page and the Patient Appointment Scheduling page.

Page reviewed by Kevin Torf, Managing Partner, T2 Group, Updated November 2026

Frequently Asked Questions

How do you calculate patient leakage?

Divide the number of referrals that did not result in an arrived visit inside your system by the total number of referrals in the period. Report it by service line and by referral source, since the rate differs for cardiology and dermatology and for internal and outside referrals. Track the steps in between as well: referral received, patient contacted, visit scheduled, and patient arrived. The gaps between those steps show where the loss happens and which team owns the fix.

What is a good referral conversion rate?

It depends on the service line and the referral source, so a single benchmark is not useful. Set a target against your own baseline from the prior 90 days and report the trend by service line. Most health systems find that speed to first contact is the metric that moves conversion the most. A referral contacted within 24 to 48 hours of the order converts at a higher rate than one contacted a week later.

Can a call center reduce leakage?

Yes. Most leakage comes from referrals that are not contacted quickly, not scheduled correctly, or not rebooked after a no-show. All three are contact center functions. A team that is staffed for outbound follow-up, trained on specialty scheduling rules, and measured on first call resolution and no-show recovery will convert more referrals. T2 Flex staffs and reports on those measures for health systems and works inside the client's EHR.

How fast should a referral be contacted?

Within 24 to 48 hours of the order. The chance of conversion falls with every day of delay, because the patient's attention moves on or the patient books elsewhere. Outbound follow-up needs its own staffing so it is not displaced by inbound volume. A working rule is three contact attempts per unworked referral across phone, text, and portal message, with the first attempt inside the 48 hour window.

Ready to explore a partnership?

Let's Connect

We use cookies. View our Privacy Policy