Pre-pilot infrastructure proposal
BRIDGE
BRIDGE is proposed buprenorphine continuity infrastructure. Candidate Mode A is designed to surface a reported medication-access interruption, route it to a named human owner, and document what happened without changing prescribing or dispensing authority.
It is not a clinic, pharmacy, prescription transmitter, benefits manager, or quality-measure engine. Prescribing remains in the treating physician's own EHR/EPCS workflow. A pharmacist independently decides whether to dispense and retains corresponding responsibility.
Report → Route → Document
Continuity problem
The gap is operational before it is measurable.
Prescription-fill claims can be used retrospectively to flag apparent gaps in buprenorphine medication possession under a specified, study-defined medication-supply gap definition: >30 days in Samples and ≥45 days in Xu (Xu et al., 2024; Samples et al., 2018). Those records are useful after the fact, but they generally cannot show the full operational history while an access problem is active: whether a pharmacy lacked stock, clarification was pending, authorization stalled, the treatment plan changed, or the patient obtained medication through another lawful path.
BRIDGE starts from a narrower hypothesis: continuity improves only if an interruption is reported early enough, routed to a named human owner, and documented at event level. Pharmacy barriers were reported in Hendy's 2024 cross-sectional, self-reported survey of 601 adults receiving treatment from one telemedicine practice across Florida, Michigan, New Jersey, Ohio, and Texas (Hendy et al., 2025), but BRIDGE does not assume that every barrier has the same cause or that software can resolve it automatically.
The practical target is not a universal explanation for discontinuation. It is an earlier, auditable handoff when a patient or participating professional reports a refill, dispensing, authorization, clarification, stock, or administration problem. The design treats uncertainty as a state to document, not a reason to infer fault. A closed event should say who owned the next step and what happened; an unresolved event should remain visibly unresolved instead of being converted into a silent gap.
Candidate Mode A
A bounded path from signal to accountable follow-up.
Candidate Mode A is the only public operating model described here. It is designed for feasibility review before any patient-level deployment.
The model is intentionally smaller than a care-management platform. It does not try to become the clinical chart, the pharmacy system, the claims system, or the regulator's record. It asks whether a narrow event workflow can keep a continuity problem from disappearing between those systems.
Reported interruption
A patient, clinician, pharmacy, or authorized operational source reports that medication access may be interrupted. The report is a coordination signal, not a prescription, order, dispensing instruction, clinical directive, or benefits determination.
Named human owner
The event is assigned to a human owner who gathers the minimum operational facts needed for follow-up. The owner may coordinate among permitted parties, but BRIDGE does not make clinical, medication-access, payment, or pharmacy decisions.
Event documentation
The outcome is documented as an event history: what was reported, who owned follow-up, what coordination occurred, and how the interruption resolved or remained unresolved. Scores, raw PDMP results, LLM output, and unrelated clinical narrative are excluded from the pharmacy-facing documentation packet.
A candidate event must have a permissible source, a timestamp, an accountable owner, and a closing state. If those facts are not available under an approved data flow, BRIDGE should not infer them.
Authority boundaries
BRIDGE does not replace the people who carry legal and clinical responsibility.
The authority boundary is part of the proposed product definition, not a marketing caveat. Candidate Mode A can be evaluated only if the workflow keeps prescriptions, dispensing decisions, payer determinations, regulatory classifications, and clinical judgments with the actors who already hold those responsibilities.
BRIDGE clinical and medication decisions are physician-only. Prescribing remains in the treating physician's own EHR/EPCS workflow.
The pharmacist independently decides whether to dispense and retains corresponding responsibility under the applicable legal and professional framework, including the federal rule described at 21 CFR 1306.04.
BRIDGE has no prescribing, dispensing, EPCS, or DEA function. It does not transmit prescriptions, select pharmacies, commit pharmacies to dispense, substitute for a clinical record, or issue medication-access directives.
Candidate live Mode A permits no automated medication approval, denial, interruption, or score-driven medication-access consequence. Any automation must remain subordinate to human review and documented authority.
Pilot status
Before live use, the prototype has to earn a narrower shape.
BRIDGE is pre-pilot, has not been validated in live clinical operations, and will not use live patient data until legal, security, technical, and operational requirements are satisfied. No improvement in access, persistence, retention, cost, or clinical outcomes has yet been demonstrated.
The existing prototype is not ready or authorized for live clinical use; conflicting automated-decision, score, PDMP, and LLM functions must be corrected or excluded and invariant-tested before any live pilot.
A bounded feasibility review should therefore start with synthetic scenarios, written role definitions, permitted data fields, security requirements, and stop criteria. The review should also identify what a partner would refuse to accept: a missing authority owner, an unverifiable source, a data element that should not be visible, or any rule that could affect medication access without human review.
Source ledger
Evidence supports the problem, not BRIDGE performance.
The sources used here establish that buprenorphine continuity gaps and pharmacy access barriers occur, and that pharmacist dispensing authority carries independent responsibility. None evaluated BRIDGE, candidate Mode A, or any proposed pilot outcome.
The ledger is deliberately short. A future feasibility protocol would need its own source list for every numerical baseline, endpoint, operational definition, privacy rule, and jurisdiction-specific authority question. Until then, source evidence is used only to justify why continuity interruption reporting may be worth reviewing, not to claim that BRIDGE works.
| Homepage claim | Source | Scope limitation and access date |
|---|---|---|
| Prescription-fill claims can be used retrospectively to flag apparent gaps in buprenorphine medication possession under a specified gap definition. | Xu et al., 2024; Samples et al., 2018 | Definitions varied: >30 days in Samples and ≥45 days in Xu. These retrospective claims analyses do not establish cause, preventability, live operational history, clinical discontinuation, or BRIDGE effectiveness. Accessed . |
| Hendy's 2024 cross-sectional, self-reported survey of 601 adults receiving treatment from one telemedicine practice across Florida, Michigan, New Jersey, Ohio, and Texas reported pharmacy-related barriers to buprenorphine access. | Hendy et al., 2025 | Survey evidence is not nationally representative or an objective pharmacy audit and does not show that BRIDGE can prevent or resolve those barriers. Accessed . |
| A pharmacist has corresponding responsibility when dispensing controlled substances. | 21 CFR 1306.04 | The federal rule supports the authority boundary; it is not a BRIDGE legal opinion and does not resolve state-specific requirements. Accessed . |
Founder and contact
Bounded feasibility review is the next step.
BRIDGE is developed under the clinical direction of Edward Ratush, MD — board-certified psychiatrist and addiction medicine specialist. Early conversations should focus on the narrow candidate Mode A boundary: report, route, document, and validate before live use.
The useful output is a written sense of fit: what review pathway would apply, which operating facts must be fixed first, which data flows are unacceptable, and which safeguards have to be proven before a limited pilot could be considered.
Last reviewed . Do not send protected health information or patient-specific information by email.