How Primary Care Referrals Decide Specialist Quality and Cost

Key Takeaways

• A primary care doctor's referral usually determines which specialist a patient sees, and that choice shapes care costs and outcomes.
• Referrals tend to go to specialists the doctor already knows or works with, often inside the same health system, because referring doctors rarely see cost or outcome data for the specialists they recommend.
• Employers can steer referrals to the best-performing specialists by having a referral team use doctor-level cost and quality data to recommend specialists to each patient, rather than the primary care doctor naming one.

Primary care made up less than 5% of US health care spending in 2023, according to the Milbank Memorial Fund's 2026 primary care scorecard. Specialty care takes a far bigger share. McKinsey found that direct spending in just six specialties, including orthopedics, oncology, and cardiology, reached 38% of commercial and Medicare medical spending in 2023. Much of that spending starts with a primary care referral to a specialist, because primary care doctors make 30 to 33 referrals a year for every 100 patients, according to a 2025 study of doctors in Ontario, Canada.

Employers have spent years adding primary care access through onsite clinics, near-site centers, and virtual visits. But access hasn't changed where patients go once the doctor says "you should see a specialist." A referral decides who delivers the care that follows. Yet primary care doctors usually make that choice without seeing what the specialist costs or how their patients fare. That’s why, earlier this year, Garner hosted leaders from Marathon Health, Teladoc Health, and Garner to discuss how referrals get decided.

What happens during a primary care referral

A referral starts when a primary care doctor decides a patient needs care outside their scope, such as a cardiologist visit for chest pain or a gastroenterologist visit after an abnormal cancer screening. The doctor writes the order, sends the patient's records and the reason for the visit, and either names a specialist or hands the patient a list of providers.

Whether the patient needs a formal referral depends on the plan. Health maintenance organization (HMO) and point-of-service plans usually require one before they'll cover a specialist visit. Some also require prior authorization. Preferred provider organization (PPO) plans generally let members book specialists directly, though a referral still tells the specialist why the patient is there.

After the visit, the specialist is supposed to send the findings back to the primary care doctor so the patient's record stays complete. That step is called closing the loop, and it is a common point of failure.

How primary care doctors actually choose a specialist

Doctors tend to send patients to specialists they already know personally or have worked with. A systematic review in the Journal of General Internal Medicine pooled studies of 1,575 referring providers and found that personal knowledge of the specialist and past experience with them were the most common reasons for choosing one. Perceived skill and a wish to keep the patient inside the doctor's own practice or hospital system came next. Whether the specialist accepted the patient's insurance came up in only four of the seven studies. It never ranked among the top five factors in any of them.

Chad Ashcraft, chief growth officer at Marathon Health, which runs employer primary care clinics, described the same pattern during the panel discussion. He said most primary care doctors can't tell you what a specialist costs, so they refer to whoever their health system employs or to a friend or colleague. So the referral reflects who the doctor knows, not how good that specialist is at the procedure the patient needs.

Why the referral decides specialist cost and quality

Specialists in the same network charge very different prices. Their patients' outcomes differ just as much, so the specialist a patient is referred to sets the total cost of care for everything that follows, from the first consult through any repeat procedures.

Those differences show up between individual doctors, not just between hospitals. Ashcraft described an orthopedic group where the surgeon who performs a given back surgery most often can have different outcomes than a partner down the hall. He said facility-level data can't tell those two surgeons apart, so a referral to a strong group can still land the patient with a surgeon who has weaker results.

Bryce Miller, a general manager at Teladoc Health, which uses Garner's data in its referral tools, said that in Dallas-Fort Worth, referrals made with Garner's data saved about $2,700 per gastroenterology referral on average. Mammogram referrals saved about $300 each, roughly half the local price of the scan.

Cost data alone isn't enough, though. Ashcraft said the lowest-priced specialist often isn't the best-performing one, so a referral needs both cost and quality data.

How often doctors refer matters too

A practice that handles more care itself makes fewer referrals. Each problem it solves in the clinic is one less specialist visit for the plan to pay for. Marathon Health refers out about 10% of primary care visits, compared with 15% to 20% at the typical hospital-owned practice, Ashcraft said. He credited the lower rate to longer primary care visits and to labs, physical therapy, and behavioral health services inside the clinic.

Rates also differ widely between practices. A study of family medicine clinics in one academic health system found average referral rates ranging from about 10% to 28% of visits, nearly a threefold gap.

The hospital-owned primary care trap

Health systems can bid low to run an employer's onsite or near-site clinic. Ashcraft said a hospital's bid can come in at about half an independent group's price, because the system makes the money back when the clinic refers patients to its own specialists. So the employer saves on the clinic contract but pays the system's specialist prices on the referrals that follow. For example, nearly 50% of routine imaging is done in acute care facilities, because hospital systems generally encourage their doctors to refer internally. Hospital-based imaging costs on average 3.7 times as much as freestanding centers for the same procedure, with no discernible difference in quality, according to Garner's data methodology white paper.

Ashcraft's fix is a contract performance guarantee that caps referrals inside the hospital's own system at 10% to 15% of all referrals. He said health systems often walk away from the bid when asked to hold to that cap. In his view, that shows where they planned to make their money.

Why referrals break down after the doctor makes them

Even a well-chosen referral can fail in three places:

• The provider directory lists specialists who aren't taking new patients
• Nobody tracks whether the patient ever booked the visit
• The specialist's findings don't make it back to the primary care doctor

Kirk Czonstka, senior vice president of strategic alliances at Garner, said only about one in three directory listings leads to a bookable appointment. Miller called it a common problem, with patients finding out only when they try to book that the specialist isn't taking new patients for months.

Garner checks this in claims data. A specialist bills different codes for a new patient's first visit than for a returning patient, so when the new-patient codes stop appearing, the specialist has likely stopped taking new patients. Garner then calls the office to confirm.

The other two failures happen after the doctor places the referral. The second failure is that nobody tracks whether the patient ever booked the visit. The third failure is that the specialist's findings often don't make it back to the referring doctor. In a survey of 4,754 doctors in a federal primary care program, about a third said they only sometimes, seldom, or never get useful information back from the specialist after a consult. Without a report back, the primary care doctor can't tell whether the problem was resolved.

What a better referral process looks like

Marathon Health trains its doctors not to name a specialist during the visit. Once a doctor says a name in the exam room, Ashcraft said, patients fix on it and call the referral team asking for that doctor. Marathon's referral process runs in five steps.

1. The primary care doctor orders the referral without naming a specialist in the exam room
2. A referral team checks doctor-level cost and quality data for the patient's plan and condition
3. The team confirms the specialist is taking new patients
4. The team contacts the patient within 24 hours with options
5. The team tracks the visit and gets the results back into the patient's record

Of the about 50,000 referrals Marathon made in 2025, about 85% went to top-performing doctors Garner recommends, up from 65% to 70% when it started. Garner's data shows why that matters: the top-performing doctors in every geography add three healthy days to their patients' lives each year while lowering total cost of care by 32%, according to Garner's data methodology white paper. Ashcraft said Marathon’s referral share to Top-Performing doctors climbed because the team got better at explaining each doctor's cost and quality. For an employer with a Marathon clinic, that means most specialist referrals now go to doctors chosen on cost and quality data, not on whoever is closest or most familiar.

The doctors came around too. Ashcraft said seeing the data convinced Marathon's doctors there was "a science behind this, not just my personal preference."

Before signing with any primary care vendor, ask for three numbers that show whether it works this way:

• The referral rate, as a share of primary care visits
• The share of referrals that go to best-performing specialists
• The share of referrals that stay inside the health system that owns the practice, if one does

How Garner helps primary care send patients to the right specialist

A process like Marathon's depends on having doctor-level data to check. Garner's cost and quality data on individual doctors, built from the care journeys of 310 million patients, powers the referral tools that Marathon Health and Teladoc Health use. Marathon's referral coordinators work from that data under Garner's primary care partnership with Marathon. Employers can also give members the same doctor search, so someone who books a specialist without a referral team still sees the Top Providers in their existing network. Learn how Garner works.

Where specialist spending actually gets decided

A primary care strategy that ignores referrals leaves much of the specialist spending it was meant to control untouched, because the referral dictates the doctor who delivers the care.

Give your doctors the data to steer members to the best-performing specialists, and see how Garner works.

FAQs

How do primary care doctors decide which specialist to refer you to?

Primary care doctors most often choose a specialist based on personal relationships, past experience, and whether the specialist works in their own practice or hospital system. Insurance acceptance rarely ranks among the top reasons. Few primary care doctors see data on what a specialist charges or how their patients do. So a primary care referral to a specialist usually reflects who the doctor knows, not which specialist performs best.

Does a referral from my doctor mean the specialist is high quality?

No, a referral doesn't guarantee quality. It means your doctor thinks you need a specialist and knows of one, but primary care doctors rarely see data on specialists' costs or outcomes. Specialists in the same network, and even the same group practice, can have very different outcomes at very different prices. A tool that rates individual doctors on cost and quality tells you more than the referral alone.

Why do hospital-owned primary care practices refer inside their own system?

Hospital-owned practices refer inside their own system because the system earns money when patients see its specialists. According to Ashcraft, health systems sometimes price onsite and near-site clinics at about half an independent group's price, because referrals into their specialists make up the difference. Employers can limit this by capping in-system referrals at 10% to 15% of all referrals in the contract.

What is a normal referral rate for primary care?

No single rate counts as normal, but Chad Ashcraft of Marathon Health puts the typical hospital-owned primary care practice at 15% to 20% of visits ending in a referral. Marathon Health's clinics offer longer visits and more in-clinic services. They refer about 10% of visits. Rates also differ widely between practices, with clinics in one academic health system averaging from about 10% to 28% of visits.

How can employers influence where primary care doctors refer patients?

Employers can influence referrals by setting up a referral team that uses doctor-level cost and quality data to recommend specialists to each patient, so the primary care doctor doesn't name one in the exam room. Vendor contracts can require reporting on referral rates and cap referrals inside an owning health system. Employers can also give members a search tool that shows the best-performing doctors in their existing network when they choose a specialist on their own.

Do I need a referral to see a specialist?

Whether you need a referral depends on your plan. HMO and point-of-service plans usually require a referral from your primary care doctor before they cover a specialist visit. Some also require prior authorization. PPO plans usually let you book a specialist directly, though a referral still gives the specialist your history.