HJ Assistant — DME Roadmap¶
Created: 2026-03-19T10:00:00-07:00
Customer: OrthoXpress (Jeff Salamon, Brett Varner, Kevin Johnson, Mike Wells)
Source material: 2/18 Meeting (2026-02-18) — OrthoXpress backend meeting; Jeff's TKA DME Complaint Research (ChatGPT)
Operations lead: Pete (back office/field operations)
Strategic Priority¶
Reduce equipment support calls — the DME "70%."
Same insight as doctor-side: the majority of patient calls to DME staff are routine equipment questions that don't require a human. A chat agent trained on equipment knowledge handles the routine, escalates the real issues.
"It almost is never the equipment. It is expectation mismatch, no daily plan, no hurt versus harm education."
The root cause reframe: most equipment "problems" are actually education gaps. The system addresses the education gap, not the equipment.
Current State¶
flowchart LR
A[Patient gets<br/>equipment] --> B[Rep explains<br/>once at delivery]
B --> C[Patient has<br/>question]
C --> D[Calls DME<br/>office]
D --> E{Real issue?}
E -->|70% No| F[Staff answers<br/>routine question]
E -->|30% Yes| G[Staff handles<br/>real problem]
F -.->|Repeats| C
Future State¶
flowchart LR
A[Patient gets<br/>equipment] --> B[Care plan +<br/>chat agent<br/>activated]
B --> C[Patient has<br/>question]
C --> D[Chat agent<br/>answers 24/7]
D --> E{Resolved?}
E -->|70% Yes| F[Patient<br/>continues<br/>recovery]
E -->|30% No| G[Escalates to<br/>DME staff]
B -.->|Learns| D
Feature Areas (Dependency-Ordered)¶
Tier 1 — Foundation¶
| Feature | Sub-doc | Status | Why first |
|---|---|---|---|
| Equipment Chat Agent | equipment-chat-agent.md | Planned | Core value proposition — deflect routine equipment calls. Mirrors doctor-side 70% call deflection. |
| DME Care Plan | dme-care-plan.md | Planned | Equipment-specific protocols drive the chat agent's knowledge. Can't answer equipment questions without knowing what equipment the patient has. |
Tier 2 — Learning & Reach¶
| Feature | Sub-doc | Status | Why now |
|---|---|---|---|
| Conversation Factory (DME) | conversation-factory-dme.md | Planned | Staff teach the system their real conversations without invasive call recording. Same pattern as doctor-side. |
| Companion/Caregiver Access | companion-caregiver.md | Planned | Caretakers are the real equipment operators post-op. Same content, different framing. |
Tier 3 — Operations & Modernization¶
| Feature | Sub-doc | Status | Why this tier |
|---|---|---|---|
| DME Dashboard | dme-dashboard.md | Planned | Pete's operations view — patient equipment status, rental lifecycle, proactive alerts. Requires data from Tiers 1-2. |
| OXP Live Modernization | oxp-live-modernization.md | Planned | Legacy Cold Fusion order system rewrite. Separate from patient-facing product but strategic for OrthoXpress relationship. |
Tier 4 — Deferred¶
| Feature | Sub-doc | Status | Why deferred |
|---|---|---|---|
| Business Model | business-model.md | Deferred | Pricing still open — per-patient, PMPM, per-life all discussed. Resolve after MVP validates DME value. |
Cross-Cutting Principles (From 2/18 Meeting)¶
Equipment is rarely the problem¶
Most calls are expectation mismatch, not equipment malfunction.
"Patients don't quit because devices fail. They quit because confidence fails." — Jeff's research
"70% of the calls that came in... are unnecessary, that we can triage through this platform." — 2/18 Meeting
All brands, not just one¶
The chat agent must support all equipment brands and models, not be locked to a single manufacturer. OrthoXpress works with multiple brands — the system must too.
Staff co-creation¶
DME field staff know the real conversations. They build the knowledge base by reviewing example chats and dictating real recurring issues. Not call recording — collaborative authoring.
Equipment assignment is a business decision¶
The doctor/insurance/cash "do-si-do" determines what equipment the patient gets. The system accepts the final DME decision and builds the care plan around it. It does not automate payer logic.
Prehab timing for equipment education¶
Equipment education works best before surgery, when the patient can absorb it. Post-op, the amygdala fires and retention drops.
"You're asking a 68-year-old post-anesthesia patient to run a home rehab lab. They won't." — Jeff's research
Sub-Document Index¶
- equipment-chat-agent.md — Chat agent trained on equipment knowledge, device-specific accuracy
- dme-care-plan.md — Patient-specific care plans reflecting actual equipment assignment
- conversation-factory-dme.md — Staff teach the system without call recording
- companion-caregiver.md — Caretaker access to equipment guidance
- dme-dashboard.md — Operations visibility for Pete and back office
- oxp-live-modernization.md — Legacy OXP Live system rewrite
- business-model.md — Pricing and licensing model (deferred)