New patient volume is short of plan — MSOs

Referrals are being placed. Nobody can say whether the patient was seen.

PRESENTATION — WHAT THE OPERATOR SEES

Primary care is referring. The specialists say their schedules are not full. When someone asks what happened to last month's referrals, the answer takes two weeks to assemble and comes back partial. The instinct in the room is to generate more demand — before anyone has established what happened to the demand already generated.

DIFFERENTIAL — LIKELY CAUSES, MOST LIKELY FIRST

01
Referrals placed that never reach a documented scheduled appointment

The referral order is written, and the scheduling attempt either never happens or is never recorded. The demand exists, was physician-generated, was directed at a named specialist, and evaporates before the calendar. Read as the internal referral funnel — placed, contacted, scheduled, seen — the loss has a location; read as a total, it does not.

Across 103,737 referral scheduling attempts, the documented appointment scheduling rate was 61.1% — meaning roughly four in ten never reached a documented scheduled appointment at all. Inside a single integrated health system.

Patel MP et al., J Gen Intern Med. 2018;33(5):715-721. One academic health system, 20 high-volume specialties, July 2015 – June 2016. Population limit: do not present as a national rate.

NOT THIS IF — Not this if the group can produce, for a closed month, a reconciled count of orders placed against appointments scheduled. If that reconciliation exists and shows a high scheduling rate, the loss is downstream.
02
Scheduled and then lost — cancellation and no-show

The appointment exists and the encounter does not. Cancellation is the larger of the two losses by a wide margin, and it is the one most often mistaken for a no-show problem.

Of the same 103,737 referral scheduling attempts: 18,531 (17.9%) cancelled and 4,117 (4.0%) no-shows — cancellation is more than four times the no-show loss. Separately, 73% of medical practices report no-show rates stayed the same (60%) or decreased (13%) in 2025 versus the prior year, while 27% report increases.

Patel MP et al., J Gen Intern Med. 2018; MGMA Stat poll, August 12, 2025 (265 applicable responses).

NOT THIS IF — Not primarily a no-show problem if cancellations dwarf no-shows in the group's own data, which is the pattern in the only peer-reviewed referral series with a stated denominator. Investing in no-show reduction against a cancellation problem addresses the smaller number.
03
No documented status — the group cannot establish whether the encounter happened

This is not a conversion failure. It is a measurement failure, and it makes every conversion number the group reports unreliable in both directions.

40,377 of 103,737 referral scheduling attempts (38.9%) had no documented status at all. A further 4,640 (4.4%) fell into other statuses. The system could not establish whether those patients were ever seen.

Patel MP et al., J Gen Intern Med. 2018;33(5):715-721 — read at source, denominator explicit.

NOT THIS IF — Not this if a closed referral cohort can be resolved to a terminal state for every record. That is the test, and passing it is rarer than operators expect.
04
Access friction at the moment of scheduling

The patient tries to convert and cannot — the phone queue, the callback, the wait for the first available new-patient slot. Every one of these sits between the referral and the encounter.

Medical group leaders ranked their 2026 patient-access priorities as no-shows 27%, online scheduling 24%, phone access 22%, wait times 21%, other 5% — the top four are all intake and access, not one is demand generation. And 71% of medical groups report fewer than one in four patients using digital tools to schedule appointments.

MGMA Stat poll, December 9, 2025 (236 applicable responses); MGMA Stat poll, July 2025 (response count not stated in the article citing it — treat the 71% as directional).

NOT THIS IF — Not this if third-next-available for new-patient specialty slots is short, call abandonment is low, and self-service booking is widely used. Then friction is not what is eating the referrals.
05
Ancillary services referred out of the group entirely

Imaging, lab, and therapy referrals leave because they are scheduled as a separate future event rather than captured at the point of the office visit. Same-day is the capture mechanism.

In an analysis of 3.4 million patients using commercial claims compiled by Definitive Healthcare, more than 1.9 million (55.9%) left the referring health system for physical therapy care elsewhere. MedPAC found that outpatient physical and occupational therapy is rarely provided on the same day as a related office visit, and that fewer than half of advanced imaging, ultrasound, and clinical laboratory/pathology services are performed same-day; about half of standard X-rays are.

Luna, using Definitive Healthcare commercial claims (n=3.4M, reported 2021 — vendor-commissioned; Luna sells in-home physical therapy, and that commercial interest should be stated when the figure is used). MedPAC, Report to the Congress: Aligning Incentives in Medicare, June 2010, Chapter 8 — regulatory framing current, 2010 volume statistics are not.

NOT THIS IF — Not this if the group has no in-house ancillary capability to capture into. Ancillary capture also runs through the Stark in-office ancillary services exception — it is a regulated activity before it is a growth activity.

HOW TO TELL THEM APART

How to tell these apart in your own numbers

Each of these is a measurement you can run yourself, without us.

01 · Referrals placed that never reach a documented scheduled appointment

Cohort one closed month of referral orders placed. Denominator is orders. Count how many reach a documented scheduled appointment within a defined window, by receiving service line.

CONFIRMS IF

A large minority of orders never reach a documented scheduled appointment. Patel et al. measured a 61.1% documented scheduling rate inside one integrated system.

EXCLUDES IF

Nearly every order resolves to a scheduled appointment. Then move to the completion test.

02 · Scheduled and then lost — cancellation and no-show

Of referral-originated appointments scheduled, the share completed, cancelled, and no-showed — reported as three separate numbers, never merged into an aggregate "missed appointment" rate.

CONFIRMS IF

Cancellations materially exceed no-shows, which is the published pattern (17.9% vs 4.0%). The interventions are different: cancellation is a re-booking and access problem, no-show is a reminder and confirmation problem.

EXCLUDES IF

Both are low and completion is high. Then the loss is before scheduling, not after it.

03 · No documented status — the group cannot establish whether the encounter happened

For the same closed cohort, the share of records that cannot be resolved to any terminal state. This is the single most diagnostic number available to a medical group, and almost none of them have it.

CONFIRMS IF

Any material share is unresolvable. At 38.9% — Patel et al.'s figure inside an integrated system — no conversion rate the group reports can be trusted, because the denominator of every other calculation is contaminated.

EXCLUDES IF

Every record resolves. Then the group's capture rate is real and can be acted on.

04 · Access friction at the moment of scheduling

Third-next-available new-patient appointment by service line; call abandonment rate and average speed of answer; share of referral-originated appointments booked without staff intervention.

CONFIRMS IF

Long new-patient waits or high abandonment in the same service lines showing low referral completion. That co-occurrence is the signature.

EXCLUDES IF

Short waits, low abandonment, high self-service booking, and completion still low. Then friction is not the cause and the handoff itself is.

05 · Ancillary services referred out of the group entirely

Same-day capture rate for imaging, lab, and therapy orders written during an office visit — orders placed as the denominator, in-group services performed as the numerator, by modality.

CONFIRMS IF

Same-day capture is low and the orders resolve to outside providers. MedPAC's finding that fewer than half of advanced imaging, ultrasound, and lab/pathology services occur same-day as an office visit is the structural reason this leaks.

EXCLUDES IF

Same-day capture is high, or the group has no in-house capability, in which case this is a capital decision and not an operational one.

WHAT RESOLVES EACH

What resolves this, and how you will know it resolved

Referrals placed that never reach a documented scheduled appointment	Rx 03 · patient conversion →	Online booking and routing so a referral order creates a schedulable event rather than a task in someone's queue. What does not work: buying acquisition against an unmeasured referral loop. Every dollar of new demand enters the same funnel that is already losing the demand you have.
Scheduled and then lost — cancellation and no-show	Rx 03 · patient conversion →	Same-day capacity matching, so a cancellation becomes a re-book rather than a loss. Size the two losses separately first — cancellation is the larger one in the only denominated series that exists.
No documented status — the group cannot establish whether the encounter happened	Rx 04 · marketing attribution →	The one-page monthly readout with a stated denominator. This is the first fix, not the fourth: nothing else on this list can be measured until a referral cohort resolves to terminal states. Worth naming plainly — the 40% to 70% referral-leakage range circulating in vendor material has no traceable denominator anywhere, and the widely-repeated "roughly 50% of referrals never result in a completed visit" appears to be a transformed misquote of Patel et al.'s 34.8%.
Access friction at the moment of scheduling	Rx 03 · patient conversion →	Call-centre instrumentation and front-desk feedback loop. Medical group leaders themselves rank no-shows, online scheduling, phone access, and wait times as their top four priorities for 2026 — this is not a fix we have to argue them into.
Ancillary services referred out of the group entirely	Rx 03 · patient conversion →	Same-day capacity matching for imaging and lab so the order is filled during the visit that generated it. Regulated territory: ancillary capture runs through the Stark in-office ancillary services exception, so the compliance view comes before the operational one.

WHAT "RESOLVED" LOOKS LIKE — Documented referral completion rate, and the share of referrals with no documented status, reported together with the denominator

MEDIAN

34.8% of referral scheduling attempts resulted in documented completed appointments; 38.9% had no documented status; documented scheduling rate 61.1% — n=103,737, inside a single integrated health system (Patel et al., J Gen Intern Med 2018)

TOP DECILE

None exists. No national referral-capture distribution has ever been published, which is precisely why the vendor literature invented a 40%–70% range with no denominator behind it. Anyone quoting a referral-capture decile for US medical groups is quoting a number that does not have a source.

TARGET

The first target is not a capture rate — it is a denominator. Resolve every record in a closed referral cohort to a terminal state, driving the no-documented-status share well below the 38.9% observed inside an integrated system. Only then is a completion rate above 34.8% a meaningful claim rather than a measurement artefact. Stated as a distribution position against the one peer-reviewed series that exists, not as a promise.

Patel MP, Schettini P, O'Leary CP, Bosworth HB, Anderson JB, Shah KP. "Closing the Referral Loop: an Analysis of Primary Care Referrals to Specialists in a Large Health System." J Gen Intern Med. 2018;33(5):715-721.

HOW THIS DIFFERS BY SCALE

How this differs by scale

Single site	One site with in-house specialists. The referral loop is short enough to reconcile by hand, and usually should be for a month before any system is bought. The commonest finding is that referral orders and scheduled appointments live in two places that were never joined.
Group	Three to fifteen sites. Referrals cross sites, and the cross-site handoff is where the no-documented-status bucket lives — the sending site closes the order, the receiving site never sees it as a referral. Reconcile by sending site and receiving service line, not in aggregate.
Platform	Fifty-plus providers, often multiple EHR instances from acquisitions. Referral status is not comparable across instances, and a platform-level capture rate computed across them is a weighted average of incompatible definitions. Fix the definition before computing the rate. This is also the scale at which ancillary capture — imaging, lab, therapy — becomes the larger dollar line than specialist referral capture.

OTHER PRESENTATIONS — MULTI-SPECIALTY MSO

A differential narrows the field. It does not replace the examination — that is what the six weeks are for. Every figure above is an industry reference range, not a client's numbers; those stay sealed. Sources are set out at /sources.

APPLY — 6 / QUARTER →