Golden Horizons
CAPABILITY · VERTICAL-SPECIFIC
Automated Patient Follow-Up
Post-visit recap, HEP adherence sequence, and no-show recovery without staff involvement.
Audit this workflow →What it does
Sends post-visit recaps and home exercise program reminders keyed to each patient plan. Runs no-show recovery sequences and routes NPS responses to Google review requests. Writes outcomes back to WebPT or Prompt EMR. HIPAA-covered on HIPAA-eligible AI infrastructure.
Physical therapy has a completion problem. A patient comes in after a knee replacement, makes real progress through weeks two and three, starts feeling better, and quietly stops scheduling. They got enough relief to feel functional. The rest of the prescribed course, the visits that would have gotten them to full range of motion and lasting strength, never happens. The clinic loses the remaining billable visits. The patient loses the outcome their surgeon actually ordered.
Most PT practices know this dropout curve by feel. Almost none have beaten it with process. Manual follow-up calls are the usual fix, and they don't scale. The front desk is buried in check-in queues, rescheduling calls, and insurance follow-ups. A personal call to every patient who missed a visit, or every patient whose home exercise compliance is slipping, never happens consistently because there aren't enough hours to make it happen.
Golden Horizons builds a HIPAA-covered outreach cadence that runs on HIPAA-eligible AI infrastructure and connects straight into the clinic's EHR: WebPT, Clinicient, or TheraOffice. After each visit, the patient gets a plain-language recap of what was worked on and what to do at home before the next appointment. If a scheduled visit comes and goes with no check-in, a recovery sequence runs. A text first, a second touch if there's no reply, then a flag to the front desk with the patient context already pulled, so the staffer making the call has what they need in ten seconds instead of two minutes of chart hunting.
The cadence keys to each patient's clinical protocol. A post-surgical patient on a twelve-week course gets a different message rhythm and a different set of check-in questions than someone on a six-visit sports-injury program. If a patient reports a symptom change, like increased pain, swelling, or unexpected limitation, the system flags the record for clinician review the same day. The automation owns the communication layer. Clinical judgment stays with the therapist.
Consent for SMS and email outreach gets collected at intake and stored in the EHR. The system operates under a Business Associate Agreement on HIPAA-eligible AI infrastructure. No PHI reaches any model endpoint without the contractual protections HIPAA requires. The clinic's compliance team gets the full data flow diagram before go-live.
Use cases
- Post-surgical knee rehab: a patient finishes three of eight prescribed visits, then cancels twice without rescheduling. The system sends a check-in on day five of silence, hands the front desk the gap with a reschedule link pre-populated, and the patient books back in the same day.
- Chronic low-back program: a patient reports big pain reduction after four visits and stops booking. Outreach asks two questions on home exercise compliance. The clinician sees the flagged responses in the EHR that afternoon, calls with an adjustment, and the patient re-engages.
- Sports injury return-to-play: a high school athlete on a six-week hamstring protocol gets weekly check-ins confirming home exercise completion and flagging any new soreness for therapist review, keeping athlete and parent informed without piling onto front desk call volume.
- Stroke rehab outpatient: the caregiver joins the cadence alongside the patient. Post-visit recaps go to both contacts, reminders use plain language sized for varying health literacy, and any reported functional change routes to the treating therapist within the hour.
- Balance and falls program for seniors: a weekly check-in confirms balance exercise completion and flags any reported fall for same-day clinician review. Non-responders after 48 hours trigger a front desk escalation with the patient's emergency contact already surfaced.
What’s included
- Fixed scope with written acceptance criteria before any build starts
- Customization layer for your brand voice and business rules
- Clean handover with documented runbook and live training
- Monthly ROI report for three months post-delivery
- Source code delivered to your GitHub on handover
What’s NOT included
- Third-party API subscription costs (billed to your accounts)
- Data migration from legacy systems
- Ongoing infrastructure costs after handover
HIPAA-covered when sold to a clinical entity. Pinned to approved cloud model access with executed BAA before go-live.
How clients use this
Fixed-scope build with clean handover, documented ownership, and optional support for monitoring, maintenance, and minor changes.
Questions Patient Follow-up (Physical Therapy) clients ask
Is this system actually HIPAA-compliant, and do we need to sign a BAA?
Yes to both. You sign a Business Associate Agreement before any PHI touches the system. The outreach infrastructure runs on HIPAA-eligible AI infrastructure, which falls under approved cloud's HIPAA-eligible services program and BAA. SMS provider, which delivers the SMS, offers a BAA for covered entities as well. We configure every integration under a signed BAA, map every PHI data flow in writing before the build starts, and hand that documentation to your compliance contact for review. The workflow never passes patient health information to any model endpoint that hasn't been scoped under HIPAA-eligible terms. If your compliance officer wants to review the architecture diagram or the BAA language before you commit, that happens before the build starts, not after.
Which EHR systems does the follow-up cadence integrate with?
The primary integrations are WebPT and Prompt EMR, both of which expose API access for appointment data, patient demographics, and visit notes. Clinicient and TheraOffice are supported too, though Clinicient's API access depends on your contract tier with Clinicient directly. We audit that access during the scoping call before we quote the build. The system reads the appointment schedule and clinical protocol from the EHR, writes outreach event logs back to the patient record, and surfaces flagged responses as task items inside the EHR where the platform allows it. We don't build against screen-scraping or unsupported workarounds. If your EHR doesn't expose the access we need, you hear it before the build, not after.
What does the AI actually do versus what the therapist still handles?
The system owns the communication layer: scheduling the outreach, sending the messages, collecting responses, logging events to the EHR, and routing flags to the right person. It makes no clinical decisions. When a patient reports a symptom change, like increased pain, unexpected swelling, or a fall, that response routes to the treating therapist as a same-day flag, and the therapist decides what to do with it. The system never tells a patient whether a symptom is normal or concerning. It never modifies a care plan. It never advises a patient to stop or change an exercise. Clinical judgment stays with the licensed practitioner. The automation's job is to put the information in front of the therapist without a phone-tag cycle in the way.
How is patient consent for SMS and email outreach handled?
Consent gets collected at intake, before any automated outreach runs. The intake form includes explicit opt-in language for both SMS and email, consistent with Telephone Consumer Protection Act requirements for automated messaging. The consent record is stored in the EHR against the patient's chart. Patients can opt out at any time by replying STOP to any SMS, which immediately suppresses outreach for that patient and logs the opt-out in the EHR. The system sends no automated messages to patients who haven't opted in or who have revoked consent. If your current intake process doesn't capture this consent, we help scope a compliant intake update before the follow-up build goes live.
What does better visit completion actually do for a clinic's revenue?
The revenue math is direct: every patient who finishes their prescribed course represents the gap between what they actually paid and what a full course generates. For a practice running forty to sixty active patients on multi-visit protocols, even a modest bump in completion rate across the roster adds up inside a single month. We don't quote a fixed lift percentage, because every clinic starts from a different baseline. A practice with strong front desk bandwidth and an existing recall process sees a smaller delta than one where follow-up is entirely inconsistent. The scoping call reviews your current visit completion rates, average prescribed course length, and existing follow-up process before we put a realistic range on it. What we can say plainly: clinics running no systematic follow-up are leaving the most on the table, and the build closes that gap head-on.