A patient can have years of periodontal measurements, restorative history, missing teeth, emergency dental visits, medications prescribed by dentists and repeated evidence of poor oral health while the medical record used by the patient’s physician contains almost none of it. That separation feels increasingly strange in a healthcare system that has spent more than a decade trying to make information follow patients across hospitals, physician practices, pharmacies and laboratories. The problem is not that dentistry generates too little information. It is that the information has historically lived in a parallel clinical and financial system. 6
For most of modern U.S. healthcare, dentistry has operated with different benefits, different software, different claims, different provider networks and often a different definition of what belongs in the longitudinal patient record. Dental offices can treat teeth without seeing every medical detail, and medical practices can manage chronic disease without opening a periodontal chart. But the more healthcare moves toward whole-person risk management, medication safety, screening and longitudinal care, the more consequential that separation becomes.
The data exists. It is just somewhere else.
Dental records can contain highly structured information: tooth-level diagnoses, periodontal probing depths, gingival margins, attachment loss, radiographs, procedures, medications, treatment plans, frequency of preventive care and patterns of emergency utilization. Even ordinary dental claims create longitudinal signals about whether a patient is receiving preventive care, experiencing advanced restorative burden, undergoing periodontal therapy or losing teeth. None of those signals automatically predicts a medical outcome. But they are still healthcare data, and in many organizations they remain difficult for medical teams to access in the same workflow where other clinical information is reviewed.
The technical argument for separation is getting weaker. Epic now publicly documents FHIR interfaces that can return dental findings as Condition resources, periodontal measurements as Observation resources and dental treatment plans through CarePlan resources. Its periodontal API can return probing depth, gingival margin and clinical attachment measurements attached to individual teeth. Those are structured clinical data that can be exchanged using the same broader interoperability standard increasingly used elsewhere in medicine. 234
Interoperability is not the same as integration.
It is tempting to conclude that once a FHIR endpoint exists, the problem has been solved. It has not. Interoperability describes the ability for systems to exchange information. Integration means that the information arrives in a place, format and moment where it actually changes a decision. A primary-care clinician does not need every dental chart detail on every patient. A dentist does not need the entire medical record reproduced inside a dental workflow. The real design question is which oral-health information is important enough to surface, to whom, and under what circumstances.
Epic says organizations using Care Everywhere exchange more than 30 million patient records daily, with around half of those exchanges occurring with organizations using a different interoperable EHR. It also exposes more than 1,000 public API and interface specifications. That is a massive infrastructure layer. The oral-health opportunity is therefore not to invent interoperability from scratch, but to define how dental information should participate in an ecosystem that already exists. 1
“The oral-health data gap is no longer mainly a technical problem. Standards and APIs increasingly exist; the harder problems are workflow, incentives, benefit separation and deciding which oral signals are useful enough to act on.”
Oral Signal analysis
Medication safety is the clearest near-term use case.
The most intuitive benefit of a shared record is not a speculative prediction model. It is avoiding mistakes when clinicians do not know what other clinicians know. Dentists routinely make treatment decisions affected by anticoagulants, immunosuppressants, diabetes medications, cancer therapies, allergies, kidney function, pregnancy status and other medical factors. Medical teams likewise may not know about antibiotics, analgesics or other medications prescribed during dental care unless that information is reconciled elsewhere.
Epic reported that more than 2,300 dental clinics were using its platform to bring dental and medical care closer together. In describing the shared-record model, Epic highlighted medication safety, systemic-risk identification and referrals as practical benefits. The company also said integrated records allow dentists to tailor treatment for patients managing cancer, diabetes, complex pregnancy and other medical conditions. Those examples matter because they do not require proving that periodontal treatment prevents a heart attack. They show value from giving clinicians a more complete patient story. 5
The dentist may see patients medicine misses.
There is another reason the oral record matters: dental care can be a healthcare touchpoint for people who are not consistently interacting with the medical system. Epic cited an estimate that roughly 9% of Americans — more than 28 million people — had seen a dentist in the prior year but not a medical doctor. In that population, the dental encounter can become a screening and referral opportunity for hypertension or other risks that might otherwise go unnoticed. 5
The key word is opportunity. A blood-pressure reading taken in a dental office only matters to the broader healthcare system if the result can be routed, interpreted and acted on. A referral only matters if the patient gets connected to care. A flagged medical risk only matters if someone owns the next step. That is why the missing-data-layer problem is as much about workflow as data exchange.
Oral data should earn its place in the medical record.
The argument for integration becomes weaker when it turns into an assumption that every dental variable belongs everywhere. Healthcare already suffers from information overload. Clinicians face long problem lists, duplicated alerts and data that are technically available but practically unusable. Oral-health advocates should resist the idea that success means dumping the dental chart into the medical chart. Success means identifying a small number of oral signals that improve a decision enough to justify the complexity they introduce.
A useful framework is incremental value. Suppose a health plan already has age, diagnoses, pharmacy history, laboratory data and prior utilization. Does adding periodontal status improve identification of people at risk for uncontrolled diabetes? If it does, by how much? Does that improved prediction lead to earlier screening or a different intervention? Does the intervention improve outcomes? The same sequence applies to pregnancy, cardiovascular risk, oncology and older adults with complex medication burdens. The burden of proof should increase as the claim becomes more ambitious.
Claims data may move faster than chairside data.
The first scalable oral-health data layer may not come from full clinical interoperability at all. It may come from dental claims. Claims are imperfect: they reflect billing rather than the complete clinical picture, and procedure codes are not equivalent to disease severity. But they are longitudinal, structured and already familiar to payers. They can reveal preventive-visit patterns, periodontal procedures, extractions, emergency care and treatment intensity across large populations.
That makes claims particularly attractive for payer analytics. A medical insurer that also administers dental benefits can ask whether dental utilization adds signal to existing risk models without first solving every EHR-integration problem. The standard should be empirical. If oral features improve calibration, identify a population earlier or help route a useful intervention, they may have value. If they do not, the hypothesis should be discarded rather than defended simply because the mouth and body are biologically connected.
Benefit design is part of the data architecture.
One reason the data remains fragmented is that the financing remains fragmented. Medical and dental benefits are often sold, administered and contracted separately. The entities holding dental claims may not be the same entities managing medical risk. Even when a parent insurer owns both businesses, operational teams, data warehouses, provider networks and product incentives may remain separate. The result is a structural version of the clinical silo: the information exists, but no one is paid to combine it in a way that changes care.
That is why medical-dental integration cannot be reduced to EHR interoperability. A shared record is useful infrastructure, but durable integration also requires incentives. Who pays for the oral intervention? Who captures the medical savings if savings exist? Who is accountable for the referral? Who measures outcomes? Until those questions are answered, technical connectivity can remain impressive but underused.
The Epic example matters because it collapses two worlds.
Epic’s dental work is strategically interesting because it places oral and medical information inside the same broad record architecture rather than asking healthcare to build an entirely separate interoperability universe for dentistry. Its public FHIR documentation treats dental findings, periodontal observations and care plans as resources that can live within the broader health-information model. That is a subtle but important shift: dentistry becomes another clinical domain within a comprehensive record instead of an adjacent software category. 234
This is also why the PDS Health story matters beyond one dental organization. PDS has spent years operating dentistry on Epic and is now extending that experience through PDS Health Technologies. The broader question is whether other DSOs, dental schools, health systems and integrated care organizations follow. If they do, oral-health interoperability could advance not because a new dental standard wins, but because dentistry joins healthcare’s existing infrastructure.
What should actually flow between dentistry and medicine?
A pragmatic exchange model would start with information that has obvious cross-setting relevance: medications, allergies, major diagnoses, pregnancy status, cancer therapy, immunosuppression, anticoagulation, recent procedures, relevant laboratory findings and blood pressure from the medical side; significant infections, medication prescribing, periodontal disease severity, tooth loss, urgent dental needs and selected screening findings from the dental side. More detailed information can remain available on demand rather than being pushed into every workflow.
The design should also distinguish between raw data and interpretation. A probing depth is a measurement. A periodontal diagnosis is a clinical interpretation. A claim for scaling and root planing is a billing event. Those should not be treated as interchangeable signals. Building useful cross-domain data products will require discipline about provenance, timing and meaning — especially if the information is used for analytics rather than direct care.
There is a research opportunity hiding inside the integration problem.
Better linked records would not only help care coordination; they would improve the evidence base. Many oral-systemic questions are difficult because dental and medical outcomes are studied in different datasets. Linking longitudinal oral findings to laboratory values, diagnoses, medication use and utilization could make it easier to test whether oral variables provide incremental information and whether interventions change outcomes. The same infrastructure that supports care can support more rigorous causal and predictive research.
Larger integrated datasets can still produce highly significant associations that are confounded. Smoking, socioeconomic status, diet, healthcare access, age and chronic-disease burden can affect both oral and systemic outcomes. Better data increases statistical power; it does not automatically create causality. The value of integration is that it allows better questions to be asked, not that it guarantees the answers oral-health advocates may want.
The next competition may be over workflow, not standards.
If the standards are increasingly available, the strategic battleground shifts. Which organization can turn oral data into a useful clinical workflow? Which health plan can identify a population worth intervening on? Which technology can route a referral and close the loop? Which integrated group can demonstrate better medication safety or earlier detection? Those are more defensible problems than merely transporting data from one database to another.
Companies that solve this well may come from several directions: EHRs, dental software, payer analytics, referral platforms, benefits administrators, diagnostics or integrated provider organizations. The winning product may not advertise itself as an oral-systemic platform. It may simply make dental information visible at the moment it matters and make the next action easy.
Bottom line
The surprising thing about the medical record’s oral-health blind spot is no longer that dentistry has historically been separate. It is that the technical rationale for keeping it separate is eroding. Epic publicly exposes dental FHIR resources. Large dental groups are operating on shared medical-grade records. Interoperability networks already move enormous volumes of patient information. The remaining challenge is deciding what oral data deserves to move, where it should appear and what action should follow. 1235
The mouth does not need to become the center of the medical record. It needs to stop being invisible when its information can improve a decision. That is a narrower claim than saying oral health determines overall health — and a much more actionable one.
NOTES & SOURCES