A Documentation Workflow Built for Cerbo (MD HQ)
Clinical documentation inside Cerbo (MD HQ) rarely fails because clinicians lack dedication—it fails because functional and integrative medicine generates an unusually dense record. A single visit may involve a specialty-lab panel from Genova, Vibrant, or Rupa, a reconciled supplement list, IV therapy orders, dietary and lifestyle recommendations, and a longitudinal wellness plan. Cerbo, descended from the MD HQ platform, is genuinely well suited to this complexity: it supports customizable charting templates, grouped supplement protocols, individualized treatment plans, HL7 laboratory interfaces, and portal delivery of results and plans. The bottleneck is the manual keystroke work required to move a clinician's reasoning into those structured fields.
Merry AI addresses that bottleneck without pretending to be something it is not. We want to be precise about a claim that circulates in the market: Merry AI does not maintain a verified native FHIR write integration that automatically converts laboratory results into Observation resources and maps them into Cerbo plan grids. That would require a documented, contractually assured, and independently tested write-back API—and Cerbo's published API surface centers on reading patient laboratory results, not autonomously creating plan items. Instead, Merry AI accelerates the human workflow through browser-native DOM injection, keeping the clinician firmly in the authorship loop.
This distinction matters for governance as much as for engineering. An honest architecture is one you can defend during a security risk analysis and a malpractice review simultaneously. The sections below describe how our Chrome Extension actually behaves inside a Cerbo session, why that behavior avoids the failure modes of both API marketplaces and manual copy-paste, and how the mandatory clinician-review step preserves the integrity of the legal record.
The Browser-Native Injection Model Explained
Traditional integration thinking assumes that two systems must exchange data server-to-server through an API or interface engine. For closed or semi-closed EMRs, that path is slow, expensive, and often unavailable—it requires marketplace approval, credential provisioning, sandbox testing, and ongoing rate-limit management. Merry AI takes a different route entirely by operating where the clinician already works: inside the authenticated browser tab rendering Cerbo (MD HQ).
The Chrome Extension observes the live Document Object Model of the Cerbo charting, supplement, and treatment-plan screens. Each editable field—the assessment textarea, the supplement dosing instruction input, the frequency and duration selectors, the plan-grid rows—corresponds to a stable DOM selector. When you approve a generated draft, the extension focuses each target selector, sets its value, and dispatches the native input and change events that Cerbo's front-end framework listens for. To the application, the content is indistinguishable from manual typing, so all client-side validation and formatting logic still executes normally.
Because everything happens client-side within your own session, three things follow. First, there is no external network path into the practice and no new inbound firewall exposure. Second, the extension inherits exactly the permissions your Cerbo login already grants—it cannot read or write anything you could not access manually. Third, there is no polling latency: content appears the instant you approve it, rather than waiting on an interface engine round-trip. This is why sub-second field mapping is achievable.
DOM Selector Mapping for Functional Medicine Fields
Mapping content correctly requires understanding Cerbo's field structure. A functional-medicine encounter is not a single note; it is a constellation of structured elements. Merry AI maps a generated supplement protocol into the discrete product-name, dose, route, frequency, timing, duration, and clinical-purpose fields rather than dumping unstructured text into one box. Lab interpretations are written into the clinician-interpretation narrative, deliberately kept distinct from the raw HL7-delivered values so the original report remains the source of truth.
Why This Is Not Autonomous Clinical Decision-Making
A raw laboratory value should never automatically generate a supplement order. Merry AI treats every generated protocol item as a draft that requires clinician review before it is injected, and requires clinician signature before it becomes the operative record. This mirrors the controlled transformation the informatics literature recommends: observation to validated rule to draft item to clinician review to signed plan. The system is drafting and workflow assistance—not an autonomous prescriber.
Architecture Comparison: Three Deployment Models
The following table contrasts the three ways a functional-medicine practice can move clinical content into Cerbo (MD HQ). The trade-offs are not merely cosmetic; they determine total cost of ownership, patient-safety posture, and how quickly a solo practice can go live.
| Architecture | Deployment Overhead | Latency | Annual Fee | Patient-Match Safety |
|---|---|---|---|---|
| Traditional API Marketplace Integration | High: sandbox provisioning, credential lifecycle, HL7/FHIR mapping, vendor approval | Interface-engine round-trip; polling delay | Marketplace + interface engine subscription | Automated write-back risks cross-chart errors |
| Manual Copy-Paste | None, but fully manual | Human typing speed | None | Clinician controls every field manually |
| Merry AI DOM Injection | Minimal: one-click Chrome Extension install | Sub-second, client-side | No marketplace or API fee | One-chart-at-a-time human confirmation |
Manual copy-paste is safe but slow and fatiguing—it is the status quo that drives the burnout the AMA documents in its E/M workflow analysis. API marketplace integration is fast at scale but carries setup cost, ongoing maintenance, and the genuine hazard of autonomous write-back into the wrong chart. Merry AI's DOM injection sits deliberately between them: it eliminates the keystroke burden while preserving the human confirmation step that guards patient identity.
Coding Integrity and CPT G2211 Capture
Longitudinal functional-medicine relationships are precisely the scenario CMS designed the G2211 add-on code to recognize. Per CMS Change Request 13473 (Transmittal 12461), G2211 captures the inherent complexity derived from a practitioner serving as the continuing focal point for a patient's care or managing a single serious or complex condition over time. Integrative practices managing chronic, multisystem presentations across repeated visits frequently meet this bar, yet the code is under-captured because the documentation supporting the longitudinal relationship is buried in narrative.
Merry AI surfaces the relationship context by drafting the visit note in a structure that makes the continuity of care explicit—prior lab comparisons, ongoing protocol titration, and the practitioner-patient longitudinal narrative that substantiates G2211. Because G2211 cannot be reported when the E/M is appended with modifier 25 on the same visit, the drafted note flags that interaction for clinician verification rather than assuming eligibility. The clinician confirms coding intent before signature; Merry AI never bills or codes autonomously.
Accurate E/M leveling depends on documentation that reflects true medical decision-making complexity. By structuring the assessment, data reviewed, and risk elements cleanly into their Cerbo fields, the drafted note supports defensible level selection consistent with the 2021 E/M guidelines. You can Access 1,000+ EHR Clinical Templates to standardize this structure across your practice, and clinicians can review current CMS and clinical-guideline references through resources indexed at https://www.ncbi.nlm.nih.gov/pmc/.
Compliance, BAA Coverage, and Getting Started
HIPAA responsibility remains with the practice regardless of vendor marketing, so Merry AI operates under an executed Business Associate Agreement with a zero-retention posture: no audio and no transcript persist after the draft is generated. This design directly narrows breach surface, because there is no secondary store of protected health information to compromise. The extension enforces least-privilege by inheriting your Cerbo permissions, and it requests only the DOM host access and microphone permission its function requires.
For multistate and multi-specialty practices, the segregation of sensitive content matters. Reproductive, genetic, substance-use, and mental-health information may carry stricter state requirements than HIPAA baseline, and supplement recommendations must avoid unsupported disease claims. Merry AI's field-level drafting keeps these categories in their designated Cerbo sections and preserves the clinician as author of record, so the chart clearly reflects what the patient reported, what the clinician reviewed, what was concluded, and who approved the final plan.
Getting started is deliberately unhurried and low-risk. Install the Chrome Extension, load a template, and watch a draft populate your Cerbo (MD HQ) fields under your review during a single test encounter. When you are ready to evaluate fit for your workflow, Schedule a 15-Minute Workflow Audit with our clinical informatics team, or Explore Merry AI Practice Partner Plans to see how zero-IT deployment scales across a growing functional-medicine practice.
