For: RWE, HEOR, clinical development, feasibility, and data science teams at pharma, biotech, device, payer, and AI companies
From: Rabiel Amirian, DDS, Founder, Bridge Health Syndicate
Site: https://bhsyndicate.com
Contact: dra@bhsyndicate.com
Every serious real world evidence program now starts with the same question: is this fit for purpose real world data for the decision we need to make? Regulators expect a fit for use assessment before the protocol is locked. Clinical development teams want real world data that tests eligibility criteria and enrollment assumptions upstream. HEOR teams want evidence a payer reviewer cannot pick apart.
The fit for purpose grid is now familiar. Patient reach. Clinical specificity. Recency. Representativeness. Provenance traceable to source. Feasibility, cost, and timeline. Sources are scored, ranked, and licensed against it.
Almost nobody scores the dental record, because almost nobody has it on the grid.
The fit for purpose real world data menu stops at the mouth
Ask an RWE team which data types they most want linked to a trial or a cohort and the answers are predictable: EHR, claims, trial data, social determinants of health. Dental rarely appears on the list. That is not because it lacks value. It is because the standard tokenization and linkage stack was assembled around medical claims and hospital records, and the dental chart was never connected to it.
The result is a quiet assumption inside every linked dataset: whatever happened in the dental chair either did not matter or did not happen. For a growing set of questions, that assumption fails the fit for purpose test on its own terms.
Four buyer questions where the gap decides the answer
1. Protocol feasibility and eligibility criteria
Feasibility work increasingly runs on linked real world data rather than site surveys, because every extra trial day is expensive and every unrealistic eligibility criterion shows up later as a slow enrollment curve. Yet several common criteria live in the dental record first: active oral infection, planned extractions, periodontal status, recent invasive dental procedures before bone modifying or antiangiogenic therapy. A feasibility count built without the dental chart is guessing at how many patients actually clear screening.
2. External control arms and natural history
External control arms and natural history cohorts live or die on complete event capture and correct timing. When the outcome or a key confounder has an oral component, such as osteonecrosis of the jaw, oral mucositis, or infection risk during immunosuppression, the earliest documentation sits in the dental chart. A control arm that cannot see it will undercount events, misdate onset, and invite exactly the bias questions reviewers now ask first.
3. Patients who fall out of claims
Claims based evidence assumes the patient stays inside the benefit. Many no longer do. Self pay channels for GLP-1 and other obesity therapies have created populations that are partly or entirely invisible to claims, with switching and pauses that never generate a pharmacy record. Those same patients still keep their dental recalls. The dental chart sees them on a regular cadence, with an updated medical history and oral findings at each visit. For an obesity or metabolic evidence gap, that continuity is an asset most linked datasets cannot offer.
4. Post market safety signals
Post market safety work is moving toward more granular clinical data. Oral adverse events are a clear example of where granularity matters, because they are often noticed, documented, and managed chairside long before they reach a medical claim, if they ever do. A safety program that monitors only medical encounters is measuring the escalations and missing the signal.
Phenotype validation and provenance: what buyers should demand of dental data
The fit for purpose standard cuts both ways. Dental data earns a place on the grid only if it meets the same bar as any other source. Buyers should ask four things.
A validated dental phenotype. How is periodontitis severity, osteonecrosis of the jaw, or a completed extraction defined from chart fields, and how was that definition checked against the clinical record?
Provenance traceable to source. Can each delivered field be traced back to the practice system it came from, with dates precise enough to sit before or after a medical event?
Linkage without identity exposure. Is the dental record joined to the medical record at the patient level through privacy preserving record linkage, so identity never enters the analytic environment?
A plain statement of limits. What the data captures, what it does not, and where coverage is thin. To a regulator or a sponsor, an honest account of what a dataset cannot see is worth more than another claim about scale.
If a source cannot answer these crisply, it is not fit for purpose, however large it is. That applies to dental data, and it applies just as much to the medical datasets that silently leave dental out.
Freshness is part of fitness
The best received real world data programs share one more trait: recency. Buyers increasingly favor data that is months old rather than years old, refreshed on a cadence that lets a team rerun a metric and see whether an intervention is working. Dental practices generate structured records at every recall visit, which makes the dental chart one of the most frequently updated longitudinal records a patient has. Linked properly, that cadence carries into the evidence.
What Bridge Health Syndicate operates
Bridge Health Syndicate operates the member governed dental data layer for healthcare. Contributing dental organizations set research use policy and approve partner categories. De identified dental records link to medical records at the patient level through the HELIX Protocol (patent pending). Identity never crosses the boundary.
For RWE, HEOR, and clinical development teams, that means the dental chart can be assessed on the same fit for purpose grid as every other source: reach, specificity, recency, provenance, and documented permission for the intended research use. Feasibility counts reflect who actually clears screening. Control arms see oral events when they happen. Self pay populations stay visible. Safety work captures the signal at the chair.
Questions for your next data source assessment
Which eligibility criteria in my protocol depend on information that lives in the dental chart?
Does my external control arm capture oral events with the same completeness and timing as the trial?
How many patients in my obesity or metabolic cohort disappear from claims but keep visiting a dentist?
Has any source I license told me plainly what it cannot see?
Soft next step
Fit for purpose real world data was always supposed to mean fit for the question. For a growing set of questions, the dental record is part of the answer, and it now links to the medical record at the patient level.
For the data brief on the Bridge Health Syndicate linked dental medical layer, visit https://bhsyndicate.com or email dra@bhsyndicate.com.
Rabiel Amirian, DDS
Founder, Bridge Health Syndicate
https://bhsyndicate.com
dra@bhsyndicate.com