Reduce heart-failure readmissions without adding headcount

MedComm is a validated workflow tailored for transitions-of-care teams. From discharge to home, it checks symptoms, weight, and meds against your protocol, and escalates red flags in time to intervene.

Validated with clinicians. Backed by

NSF I-Corps
Cornell Tech
C10 Labs

Readmissions don't start in the hospital. They start at home.

Missed meds. Silent symptoms. By the time the patient calls, it's already too late.

1 in 4

Medicare heart-failure patients are readmitted within 30 days.1

$15,618

mean cost of a 30-day HF readmission episode, before penalties.6

27%

of heart-failure patients are ever reached by nurse-led phone outreach after discharge.7

The status quo waits for the crisis. We catch it fast. From discharge to home: a validated, tailored workflow so your transitions-of-care nurses review escalations.

From discharge to home.

A validated, tailored workflow so your transitions-of-care nurses review escalations.

01

Enrollment at discharge

EHR-integrated, or a roster upload. Live immediately.

02

Follow-up at home

Symptoms, weight, and meds by phone and text. English, Spanish, and 15 other languages. No app.

03

Escalation-only for nurses

Your team moves off outbound dialing. Red flags arrive as a prioritized summary, not a raw transcript.

04

An audit trail you can bill on

Every call is reconstructable: what the patient reported, what MedComm did, and why each escalation fired.

Measured against the metrics you already report.

Accountable to the outcomes your quality and transitions-of-care teams already track.

Reduced

30-day readmissions & ED revisits

Deterioration escalated before it becomes a return visit.

Reduced

Time to first follow-up

Every patient contacted within 24–48 hours of discharge.

Improved

Medication adherence

Missed doses surfaced to your team the moment they’re caught.

Improved

Appointment adherence

PCP and cardiology visits reminded, confirmed, rescheduled.

Improved

Patient understanding

92% of patients report better understanding.

Improved

Patient satisfaction (HCAHPS)

90.2/100 usability in our NewYork-Presbyterian study.

Publication-backed, from day one.

50%

Our medical director's heart-failure program at a top-5 academic health system cut rehospitalizations from 1 in 4 to 1 in 8. MedComm scales that program to every patient.

Peer-reviewed publications in JAMIA, with more in the pipeline.

NewYork-Presbyterian study

90.2/100

patient usability

92%

better understanding

95%

clinician endorsement

Safety, measured

Adversarially tested across 2,000+ scripted conversations.

2,000+ test conversations

100%

harm avoidance

100%

hallucination-free

98%

safety score

Benchmark results

Same voice model. Different safety profile.

Clinician-labeled post-discharge scenarios. Both systems run the identical OpenAI model; the only difference is MedComm's clinical workflow.

Escalation accuracy

100%vs 72%

every same-day case routed, at the right level

Missed escalations

0%vs 1 in 6

no yellow-zone case left without same-day review

Guardrail breaches

1.6%vs 52%

policy violations across all scenarios

Clinically validated with

Cornell TechWeill Cornell MedicineNewYork-PresbyterianNorthwell Lenox Hill Hospital

Across every safety metric the health system cares about

MedComm clinical workflowBare OpenAI voice-model API(same model, generic prompt)

Escalation accuracy

100%
72%

Documentation completeness

what the patient reported is captured in the record

89%
71%

Overall protocol score

85%
64%

Under-escalation

missed same-day review

0%
18%

Lower is better

Over-escalation

alert fatigue

0%
3%

Lower is better

False reassurance

0%
13%

Lower is better

Unsafe medication advice

0%
13%

Lower is better

Safety-policy violations

2%
52%

Lower is better

Method: deterministic checks (did the right escalation tool fire, at the right level?) plus an independent LLM safety-policy judge, on gold labels authored from the WCM escalation-zone protocol. Even a carefully prompted generic agent still under-escalated 11% of yellow-zone cases.

The follow-up gap, in your words.

We have twelve nurses calling discharged patients, and we still only reach about half of them. The volume exceeds what we can do by phone; we had to cut our attempts from three to two.
Transitions-of-Care Nurse Lead
Almost one in four heart-failure patients is back within 30 days. They fall through the cracks in the 5–14 days between discharge and seeing their PCP.
Emergency Physician & Population-Health Faculty

Run your own numbers.

See what a 25–35% reduction is worth to your system.

1,000+
Expected 30-day readmissions / mo

63

Readmissions prevented / mo

16

Annual cost avoided*

$2,998,656

Beyond avoided readmissions. More completed follow-ups, patients kept in-system, and TCM billing protected — outreach follows the clinician's first call, and CCM & RPM stack on top. For capitated and VBC organizations, that means lower total cost of care.

* Assumes the national 24.8% Medicare HF readmission rate1 and a $15,618 mean total cost per 30-day readmission episode.6

Questions clinical buyers ask.

Is Heather a replacement for our nurses?

No. MedComm absorbs protocolized outreach and documentation so nurses move from outbound dialing to escalation-only review and more referrals. A clinician stays in the loop 24/7.

Does Heather give medical advice?

No. Heather is protocol-bounded: she reinforces your care plan and checks symptoms and meds. She never diagnoses or prescribes, and she directs emergencies to 911.

What happens when a patient reports a red flag?

A risk listener runs on every patient turn. Red flags are triaged into three priority levels, from page-the-on-call to routine follow-up, and your clinician receives a concise, prioritized summary, not a raw transcript.

What languages does Heather speak?

English, Spanish, and 15 other languages. Every language ships only after clinical validation; a language isn’t “supported” until it’s safe.

What does the evidence say?

Our first academic pilot measures feasibility, safety, and acceptability, with results being prepared for peer review. The protocol is modeled on the follow-up intensity that reduced death or HF readmission by a third in the STRONG-HF trial.4

What do we need to install?

Nothing: EHR-integrated, or a roster upload, and you’re live immediately. Voice-rail agnostic: standalone, or on the platform you already use. Escalations arrive in a secure task inbox; every call is reconstructable for audit and billing. HL7/FHIR write-back comes when your IT team is ready.

How long until we’re live?

Most pilots start within weeks. We define success criteria with your team first (reach rate, follow-up completion, medication-list corrections, escalation precision), then the pace is set by your review process, not ours.

How is patient data protected?

HIPAA with a signed BAA, SSO (SAML/OIDC), audit logs from the first login, and every interaction logged and reviewable.

If you own readmissions, we should talk.

Let's define success criteria with your team (reach rate, follow-up completion, medication-list corrections, escalation precision) and start a pilot that's simple to evaluate.

Request a 20-minute call

We'll come back with times and a short agenda.

No new headcount Live in weeks We publish data, not claims

Sources

  1. Dharmarajan K, et al. JAMA, 2013: 24.8% of Medicare heart-failure patients readmitted within 30 days.
  2. Mayr FB, et al. JAMA: mean cost per HF readmission $9,051 (2013 Nationwide Readmissions Database).
  3. KFF Health News analysis of CMS HRRP, FY2023: 2,273 hospitals penalized; up to 3% of each Medicare payment; $320M total.
  4. Mebazaa A, et al. STRONG-HF, The Lancet, 2022: high-intensity post-discharge follow-up + GDMT up-titration reduced 180-day death or HF readmission (15.2% vs 23.3%).
  5. Chaudhry SI, et al. Tele-HF, NEJM, 2010; Ong MK, et al. BEAT-HF, JAMA Internal Medicine, 2016: passive telemonitoring did not reduce HF readmissions.
  6. Kwok CS, et al. International Journal of Cardiology, 2020: mean total 30-day cost for HF patients with a readmission, $15,618 (NRD 2010–2014).
  7. Hall ET, et al. Journal of Nursing Care Quality, 2021: of 6,271 heart-failure admissions, only 27.2% had even partial participation in a nurse telephonic outreach program.