| Attribute | Detail |
|---|---|
| What it is | An AI agent for hospice eligibility and admission prioritization |
| Built for | Intake coordinators, clinical leads and operations directors at hospice and palliative providers |
| Core use cases | Automated hospice eligibility check • Referral-to-admission conversion probability |
| Time per determination | Under 5 minutes, down from 30 to 90 minutes of manual review |
| How rules are set | Plain-language instructions edited by clinical staff. Live the moment you save. |
| Data layer | Snowflake read and write• Azure Document Intelligence for intelligent document processing healthcare teams already trust |
| Triggers | Intake portal, EHR, CRM, API call, webhook or scheduled task |
| Decision model | AI scores and explains. A human approves. |
A hospice referral means a patient with weeks left, not years. Admission requires confirming they meet Medicare’s hospice care eligibility standard: a physician-certified prognosis of six months or less, verified against face-to-face notes and decline indicators, for a family already in crisis.
A 25-person intake team can properly review about 75 referrals a day. Some days bring 500. Rushed reviews admit the wrong patient, risking clawbacks and hospice compliance exposure. Slow ones lose a qualifying patient to a hospice that answered first, taking the census and the referral relationship with it.
The real problem is the system. Reviewers score inconsistently, which is exactly what a CMS audit is built to catch, and eligible referrals never get prioritized against each other. Every guidance change means an engineering cycle, and the load lands on clinicians who should be at the bedside. That is the gap an hospice eligibility agent has to close.
The hospice eligibility agent screens the referral using AI to read their clinical documents, extract values such as terminal prognosis and palliative performance scale score, apply organization’s scoring rules, and return an Eligible, Pending Review or Ineligible value determination. Caregence completes this in under five minutes per referral, against 30 to 90 minutes of manual clinical review.
A referral submitted through your intake portal fires the workflow automatically within Caregence platform. Five steps run start to finish.
Eligibility criteria and thresholds
Caregence scores clinical criteria like terminal prognosis, functional decline indicators, documentation completeness, PPS, and hospitalization history based on referral documents, totals the points, and returns a status. The weights and thresholds are editable as per eligibility requirements set by the hospice management.
Eligibility tells you who qualifies. It does not tell you who converts. When hundreds of eligible referrals compete for limited slots, the second workflow scores each one on its likelihood of becoming an admission, using your own historical data.That score becomes your admission priority order; the referrals that are most likely to convert rise to the top of the queue, without anyone having to manually reorder them.
What drives the conversion probability score
The score is decided on criteria including diagnosis fit, referral source history, geography, caregiver availability, and documentation completeness, weighed against your own admission record. It flags the opposite just as clearly: unclear family decisions, capacity limits, missing paperwork. Caregence surfaces the score and the reasoning behind it; the coordinator still decides, and it only runs on referrals already marked eligible. The scoring logic lives as plain-language text inside the Intelligent Agent, where any authorised clinical lead can read it and edit it.
| Logic & rules | Process & flow | Integration & output |
|---|---|---|
| Eligibility criteria and point weights | Mandatory versus optional documents | Notification recipients and conditions |
| PPS, prognosis and decline thresholds | Workflow branching conditions | Database fields to read and write |
| Decision cut-off scores | Trigger source and escalation paths | CRM, EHR and scheduling outputs |
| Probability weighing factors | Review routing by patient type | Follow-up workflow triggers |
Caregence is portal-agnostic. It sits above your systems as an orchestration layer rather than replacing them, so there is no migration.
A button click or form submission in your existing intake portal
Any EHR, CRM or scheduling tool can POST to a Caregence endpoint
Nightly batch eligibility reviews or weekly probability recalculations
An ‘eligible’ determination can automatically trigger the probability workflow
Every workflow step is timestamped against the patient record: the inputs pulled, the logic applied, the score produced, the decision written and the email sent. Consistent, documented scoring across every referral removes the human variance that audits punish.
| Dimension | Manual intake | Caregence |
|---|---|---|
| Time per determination | 30 to 90 minutes | Under 5 minutes |
| Scoring consistency | Varies by reviewer and by day | Identical logic applied to every referral |
| Changing the criteria | IT ticket and a development cycle | Text edit, live on save |
| Referral prioritization | Intuition and whoever called last | Conversion probability score with reasoning |
| Audit evidence | Reconstructed after the fact | Timestamped at every step, automatically |
| Capacity ceiling | Fixed by headcount | Scales with referral volume |
Bring one week of real referrals. We will run them through the workflow and show you the scores against your own past decisions.
Book a 20-minute walkthrough
