53% of root canals now start with a referral

Fifty-three percent of the root canal cases an endodontist treats this year started as a referral from a general dentist, not a walk-in patient. That is up from 43% in 2012, according to the American Association of Endodontists' 2026 Endodontist-General Dentist Referral Patterns Survey Report, published May 4, 2026. The shift toward specialist referral has been building for over a decade, and for a two-site dental group weighing a third location, it lands somewhere unglamorous: the referral file that has to leave the practice correctly, on time, every single time, regardless of which site it started at.

What the 2026 survey actually found

The AAE's survey, fielded among endodontists and general dentists and commissioned by the AAE with support from Sonendo, a dental technology manufacturer with a commercial interest in higher referral volume to endodontists, found that referrals now drive nearly all specialist endodontic work. Ninety percent of root canal treatments performed by endodontists originate from a referral. The average endodontist reports 1,930 referred patients a year. And the referral relationships behind those numbers are not static: 80% of endodontists report gaining new referring dentists in recent years, while 65% report losing at least one. General dentists rated demonstrated clinical skill, prior successful outcomes, and a caring approach as the top reasons they choose one specialist over another, not price or convenience.

So what for you: referral volume to endodontists specifically is not a one-off spike. It is a decade-long structural shift, and the relationships driving it turn over often enough that your referral process needs to survive a change in which specialist gets the letter, not just how many letters go out.

Endodontics is simply the specialty with a named, dated national survey behind it. General dental practices route a comparable, if less precisely measured, stream of referrals to oral surgeons for extractions and implants, to periodontists for gum disease beyond what a hygienist can manage, and to orthodontists for cases outside general scope. No equivalent national survey with a comparable before-and-after baseline was found for those three specialties during this research, so the numbers above should be read as representative of the specific trend in endodontics, not extrapolated as an exact match for oral surgery or orthodontics referral volume. The direction, more referred cases requiring a complete, well-documented handoff, is the same across all four.

The category the ADA's own AI survey skipped entirely

The American Dental Association's Health Policy Institute published its own AI usage data in July 2026, in "Dentists Use AI to Make Appointments More Efficient, but Draw the Line at Clinical Decision Making." Among the findings: 43.3% of dentists now use AI for at least one task in their practice, and another 26.4% plan to. HPI broke that usage into specific categories: imaging and diagnostics (22.8%), insurance verification (13.6%), explaining clinical findings to patients (13.2%), business analytics (10.1%), social media (10.7%), and reception or front-desk check-in (10.1%). See our earlier look at that data for the full breakdown: 43.3% of dentists use AI. 22.8% is imaging.

Referral coordination is not on that list. Not as a current-use category, not as a planned-future category. HPI's survey tracks charting and notetaking, insurance verification, and imaging as places dentists expect AI to reduce administrative load next. Sending a diagnosis, imaging, and treatment history to an outside specialist, correctly, with the right documentation attached, apparently was not common enough or distinct enough in dentists' own minds to register as its own line item, even as referral volume to at least one specialty climbed by ten percentage points over fourteen years. Meanwhile, around one-third of dentists (32.4%) report being overworked, per HPI's Q2 2026 economic data, a figure we covered in more detail when the AI adoption spend numbers first came out.

So what for you: a workflow can be a real and growing source of administrative drag without ever showing up in industry-level AI adoption surveys. Referral coordination is one of them right now.

The calculation

Here is what that ten-point shift means in practice, using a round, illustrative caseload rather than any single practice's real numbers. Picture a two-site general dental group diagnosing 120 root-canal-indicated cases a year across both locations, a plausible figure for a group that size and used here only to show the arithmetic. At the 2012 referral rate of 43%, that group would have referred out roughly 52 cases a year. At 2026's 53% rate, the same 120 diagnoses produce roughly 64 referrals. That is 12 more referral files a year from endodontics alone, or about one additional referral letter, records package, and follow-up check every month, without counting oral surgery, orthodontics, or periodontics referrals a growing group also generates and for which comparable national referral-rate trend data was not found and is not claimed here.

So what for you: the volume growth is modest in any single month. It compounds precisely when a practice is also trying to standardize everything else, which is exactly the moment a third site tends to arrive.

Why this hits hardest exactly when you are adding a third site

At two sites, referral consistency often survives on one experienced office manager's institutional knowledge: which template to use, which specialist to call for which case type, how to confirm the letter arrived, how to chase a non-response. That knowledge rarely gets written down because it does not need to be, until a third front desk with its own habits opens and starts drafting referral letters its own way. We covered the broader version of this problem in an earlier piece on why AI tools built for a single site quietly break at multi-site scale: AI and multi-site consistency in dental groups. Referral letters are a sharp example: a specialist who receives an incomplete or inconsistently formatted referral from one of your three locations forms an opinion about your whole group, not just that site, and that opinion shapes how quickly they see your next patient.

The failure mode is rarely a lost letter. It is closer to a referral that arrives without the right radiograph, or with a clinical note that assumes context the specialist's office does not have, or with no consistent way for your team to know whether the patient actually booked and showed up. Multiply that by three front desks, each trained by a different person at a different time, and a specialist starts forming an opinion about your group's reliability from whichever site sent the messiest paperwork last.

So what for you: the fix has to happen at the process level before it happens at the tooling level. A third site does not just add referral volume, it adds a second and third way of doing the same task.

The HIPAA piece is not optional if you use a drafting tool

A referral letter is a communication about a specific patient's diagnosis, treatment history, and imaging, sent outside the four walls of the practice, which makes it protected health information under HIPAA regardless of whether it travels by fax, email, or a portal. That obligation exists today and has nothing to do with AI. What changes if a practice adopts an AI tool to draft, track, or auto-populate referral letters is that the tool itself becomes a business associate the moment it creates, receives, maintains, or transmits that PHI on the practice's behalf, and a signed Business Associate Agreement needs to be in place before the tool touches a single patient record, not after. Worth checking alongside this: most state dental boards have their own recordkeeping requirement for referral communications and specialist responses, separate from HIPAA, and it is worth confirming your own state's rule rather than assuming HIPAA compliance covers it automatically. For practices billing insurance for the referred procedure, prior authorization documentation often rides along with the referral letter itself, particularly for oral surgery, which raises the stakes on getting the record complete the first time.

So what for you: before evaluating any AI tool for this workflow, confirm the vendor will sign a BAA and can explain in plain language what happens to patient data after the letter is sent.

What to actually do about it

Do not buy a referral-drafting AI tool as the first move. Write down the current process first, at whatever site handles it best: what template gets used, what has to be attached, how confirmation of receipt gets logged, how a non-response gets escalated, and how long a referral can sit before someone follows up. That written process is what an AI tool, if one gets added later, would enforce automatically across every site. Skip that step and an AI referral tool just automates whichever site's habits happened to get copied first, and the other two keep doing it their own way. For a two-site group serious about a third, that is a smaller project than it sounds like, and it is worth finishing before, not after, the new location opens.

The 53% referral figure is not going to reverse. Referral coordination is real, growing, unglamorous administrative work that the industry's own AI-adoption data has not caught up to counting yet. Standardizing it now, before a third site multiplies the inconsistency, costs a few hours of writing down what already works. Waiting costs a specialist relationship formed on the strength of your weakest front desk, not your best one.

Related: 43.3% of dentists use AI. 22.8% is imaging.  ·  AI and multi-site consistency in dental groups  ·  the AI Opportunity & Growth Assessment

This article is for informational purposes only and does not constitute legal, clinical, or financial advice. HIPAA and state dental board recordkeeping requirements vary; confirm current obligations with your own compliance advisor before relying on anything above.