For Odoo 18.0 · Community & Enterprise · AminiTech HMS
Predictive Pre-Staging (Clinical AI)
$149
$99
★ 30% LAUNCH OFFER
Stop ordering. Start approving. A rule-based prediction engine watches admissions, surgery bookings, diagnoses, lab releases, age milestones and chronic follow-ups â then pre-stages the likely labs, prescriptions and referrals so the doctor approves a whole bundle in one click.
✅ Rule engine + seed library✅ 1-click bundle approve✅ Rejection feedback loop
⭐ 90 days free support
💻 Try the LIVE demo — login demo / demo
📚 Documentation
6
Trigger events fire predictions: admission created, surgery booked, diagnosis recorded, lab result released, age milestone, chronic follow-up due.
8
Predicted action types â lab test, prescription, referral, monitoring, patient education, imaging, vaccination and procedure bundles.
5
Lifecycle states per suggestion: Suggested, Approved, Rejected, Snoozed, Executed â with reviewer, timestamp and generated-record trace on every one.
📸 Live screens
Features
★ All features
Setup in 4 steps
★ Live demo ↗
HMS family
FAQ
Support & Contact
CLINICAL AI
What you get
Every feature below is live on the public demo — open hms.aminitechsolutions.com (login demo / demo) and walk it yourself.
01 · PREDICTION ENGINE
Rules that fire on real clinical events
Each hms.predictive.rule carries a trigger event, Python condition logic, a JSON list of suggested actions, a base confidence and a clinical rationale page. A seeded rule library ships with the module, and rules can be scoped to specific facilities or disabled per facility â so a district clinic and a referral hospital run different playbooks.
02 · BUNDLE APPROVE
One click creates the real records
Predictions generated together share a bundle token. Accept the bundle and the module atomically creates the underlying lab requests (with matched test lines by clinical code), prescriptions and referrals â each wrapped in a savepoint so one failure never poisons the rest. The executed record is linked back on the suggestion for a full audit trail.
03 · DOCTOR WORKFLOW
Approve, Reject or Snooze 24h
Suggestions surface in the Predicted Actions queue and the Predicted Orders panel with confidence percentage, evidence text and patient context (consultation, admission or surgery). Snoozed items are automatically revived by cron when their window expires; nothing is ever silently dropped.
04 · FEEDBACK LOOP
Rejections retire low-value rules
Every rejection records a reason and notes into hms.predictive.feedback. Rules track fired, accepted and rejected counts with a live acceptance-rate percentage, and low performers get flagged for review â so your rule library gets sharper the more your doctors use it.
05 · GOVERNANCE
Full traceability on every suggestion
Each predicted action logs who reviewed it, when, which rule generated it and which record it produced. Chatter tracking on state changes gives you a defensible record of what the system suggested and what the clinician decided.
Real screens — not mockups
See it running on the live demo
Every screen below is a real screenshot from the working system at hms.aminitechsolutions.com — log in with demo / demo and try it yourself.
Live screen recording — captured from the running demo.
Everything included
All 79 features — nothing held back
The complete capability list for Predictive Pre-Staging (Clinical AI), straight from the module. Every one of these is live on the demo at hms.aminitechsolutions.com.
Prediction Engine & Triggers10 features
✓
Rule-Based Prediction Engine — Abstract hms.predictive.engine walks active rules for a trigger, evaluates each and generates predicted-action rows for the patient.
✓
Six Clinical Trigger Events — Fires on admission created, surgery booked, diagnosis recorded, lab result released, age milestone, and chronic follow-up due.
✓
Automatic Admission Trigger — New hms.admission records auto-generate pre-op/medical bundles from primary diagnosis, admission reason and chief complaint text.
✓
Automatic Surgery Trigger — Booking a surgery auto-stages pre/post-op bundles matched against the procedure name and code.
✓
Automatic Diagnosis Trigger — Recording an hms.diagnosis stages workup bundles from the ICD-10 code and diagnosis text of the consultation patient.
✓
Automatic Lab-Result Trigger — Releasing a lab result (validation_state=released) fires abnormal-value rules using test code, numeric value and flag.
✓
Clinical Context Builder — Assembles a rich ctx dict (age, sex, diagnosis text, ICD-10 list, procedure text, lab code/value/flag, pregnancy, post-op) for rules.
✓
Automatic Patient Age Calculation — Derives patient age from date_of_birth/birthdate to drive age-gated screening and paediatric rules.
✓
Bundle Token Grouping — All actions from one prediction pass share a UUID bundle token so they can be reviewed and accepted together.
✓
Safe Sandboxed Condition Evaluation — Rule Python conditions run through Odoo safe_eval against ctx, with errors logged and treated as non-match instead of crashing.
Predicted Action Workflow10 features
✓
One-Click Approve — Doctor approves a suggested action, stamping reviewer and timestamp and triggering execution into a real clinical record.
✓
Reject with Reason — Rejecting an action records reviewer, timestamp and a structured feedback row for analytics.
✓
Snooze 24h — Defers a suggestion by a configurable number of hours (default 24), setting a snooze-until timestamp.
✓
Bundle Accept — Accept-bundle approves every suggested/snoozed action sharing a bundle token in one call.
✓
Accept-All-for-Admission — Approves all pending predicted actions attached to a specific admission at once.
✓
Eight Action Types — Supports lab, prescription, referral, monitoring/vitals, patient education, imaging, vaccination and procedure/bundle actions.
✓
State Machine with Statusbar — Actions flow through suggested, approved, rejected, snoozed and executed states shown as statusbar and colour-coded list badges.
✓
Confidence Scoring — Each action carries a 0-1 confidence, boosted for highly specific rules (per 'and' clause) up to a cap.
✓
Evidence Trail — Every generated action stores a human-readable 'why' string naming the matching rule and trigger.
✓
Context Linking — Actions link back to their patient, consultation, admission, surgery and facility for traceability.
Automatic Order Execution7 features
✓
Auto-Create Lab Requests — Approved lab actions create an hms.lab.request, matching the clinical code to a lab test and adding a request line when available.
✓
Auto-Create Prescriptions — Approved Rx actions create an hms.prescription linked to the consultation with label and dose details in notes.
✓
Auto-Create Referrals — Approved referral actions create an hms.referral carrying the action label as the reason.
✓
Generated-Record Back-Reference — Executed actions store a polymorphic reference to the lab request, prescription or referral they produced.
✓
Savepoint-Isolated Execution — Each execution runs in its own DB savepoint so one failure keeps its approved state without poisoning the rest of the bundle.
✓
Soft Cross-Module Guards — Execution checks model/field presence so the module degrades gracefully when a target app is absent.
✓
Non-Order Action Hand-Off — Monitor, education, imaging and vaccine actions are logged for nurses to pick up from the inbox rather than force-created.
Clinical Rule Library (Seeded)9 features
✓
33 Pre-Built Clinical Bundles — Ships a curated library of 33 evidence-based pre-staging rules across surgery, medicine, obstetrics, paediatrics, labs and screening.
✓
Surgical Pre/Post-Op Bundles — Hip/knee arthroplasty, appendectomy, cesarean, long-bone fracture and a generic post-op day-1 bundle.
✓
Acute Medical Admission Bundles — Pneumonia, DKA, MI, stroke, sepsis hour-1, COPD, asthma, UTI/pyelonephritis, meningitis and GI-bleed bundles.
✓
New-Diagnosis Workup Bundles — Diabetes, hypertension, CKD, hypothyroidism and anaemia diagnosis bundles trigger labs, referrals and starter meds.
✓
Abnormal-Lab Response Bundles — HbA1c>9, rising creatinine/AKI, critical potassium, severe anaemia, high troponin, critical glucose, high WBC and high INR reversal.
✓
Obstetric ANC & Post-Partum Bundles — Pregnancy registration ANC workup and cesarean post-op care including DVT prophylaxis and family-planning counselling.
✓
Age-Based Screening Bundles — Age 50+ and 65+ preventive screening plus childhood immunisation catch-up and paediatric febrile-admission care.
✓
Discharge & Chronic Follow-Up Bundles — Elective pre-discharge medication reconciliation and chronic-disease follow-up adherence bundles.
✓
ICD-10 & Free-Text Matching — Rules match on ICD-10 code prefixes and keyword scans of diagnosis/procedure text for flexible detection.
Rule Configuration & Governance8 features
✓
Editable Rule Model — hms.predictive.rule stores name, code, trigger, sequence, confidence and active flag with a unique-code constraint.
✓
In-Form Python Condition Editor — Condition logic is edited in an Ace code editor with Python syntax and validated at save via compile().
✓
JSON Action-Spec Editor — Suggested actions are authored as JSON in an Ace editor and validated to be a well-formed list on save.
✓
Clinical Rationale Field — Each rule documents its evidence-based clinical justification in a translatable rationale text page.
✓
Facility Scoping — Rules can be restricted to specific facilities or explicitly disabled on named facilities via many2many sets.
✓
Equipment Capability Gating — requires_ct auto-suppresses CT-dependent rules like stroke on facilities without a CT scanner; requires_lab flag included.
✓
Rule Sequencing — Drag-handle sequence field controls the evaluation and display order of rules.
✓
Translatable Rule Names & Rationale — Rule name and rationale are translatable, with a bundled Swahili (sw) translation catalog.
Analytics & Self-Learning8 features
✓
Per-Rule Firing Counters — Tracks how many times each rule fired, was accepted and was rejected.
✓
Computed Acceptance Rate — Stores a live acceptance percentage per rule for adoption monitoring.
✓
Auto-Flag Low-Value Rules — Weekly cron flags rules when rejection exceeds 70% over 10+ reviews, and unflags when it recovers below 50%.
✓
Rejection Feedback Log — hms.predictive.feedback captures reason (not indicated, duplicate, contraindicated, cost, other), notes, user and patient.
✓
Rule Adoption Bar Graph — Graph view compares accepted vs rejected counts across rules.
✓
Rule Performance Pivot — Pivot view breaks acceptance rate and fired count down by trigger event.
✓
Predictive Dashboard Stats API — Endpoint returns active-rule count, totals, adoption %, average actions per encounter and top rule rankings.
✓
Auto-Flag Chatter Audit — Flagging posts an explanatory message to the rule chatter recording the rejection rate and sample size.
Doctor UI (OWL Panel)7 features
✓
Predicted Actions OWL Panel — Reusable client component surfaces pending suggestions for a patient/consultation/admission with approve, reject, snooze and bundle controls.
✓
Multi-Surface Mounting — Panel is designed to embed in the Doctor Inbox 'Suggested Actions' folder, AI Draft Panel and consultation form banner.
✓
Confidence Labels & Colours — Renders High/Medium-High/Medium/Low confidence badges with matching CSS classes.
✓
Action-Type Iconography — Each action type shows a distinct FontAwesome icon (flask, medkit, syringe, camera, etc.).
✓
Inline Reject Form — Panel opens an in-place reject dialog capturing structured reason and free-text notes before submitting.
✓
Toast Notifications — Approve/reject/bundle actions raise success, info and danger toasts for immediate feedback.
✓
Loading & Empty States — Panel manages loading spinners and empty-list handling with graceful RPC failure fallback.
REST / JSON API7 features
✓
Manual Predict Endpoint — /predict_for_context lets the UI trigger prediction for a patient with optional trigger and context hints.
✓
Accept / Reject Action Endpoints — Per-action approve and reject JSON endpoints return the resulting state; reject accepts reason and notes.
✓
Accept-Bundle Endpoint — /accept_bundle approves all actions sharing a token and reports the accepted count.
✓
Accept-Bundle-for-Admission Endpoint — Bulk-approves all pending predicted actions on one admission.
✓
Inbox Endpoint — /inbox returns pending suggestions filtered by patient or admission, ordered by confidence, with label, type and evidence.
✓
Dashboard Stats Endpoint — /dashboard_stats returns aggregate adoption metrics and top-rule rankings as JSON.
✓
Authenticated JSON Contract — All endpoints are auth='user' JSON routes with consistent ok/error response envelopes.
Backend Views & Navigation7 features
✓
Predictive AI Menu Root — Adds a Predictive AI section under HMS clinical with Dashboard, Predicted Actions, Rules and Rejection Feedback items.
✓
Predicted Actions List & Form — Colour-decorated list with inline approve/reject buttons plus a form showing links, details and evidence, defaulting to Suggested.
✓
Rules List & Form — List with adoption metrics and toggles plus a tabbed form for rationale, condition logic, action JSON and facility scope.
✓
Search Filters & Group-By — Filters for active/needs-review rules and suggested/approved/rejected actions, grouped by state, type, rule or trigger.
✓
Rejection Feedback View — Dedicated list of rejection feedback records for review and pattern analysis.
✓
Chatter on Rules & Actions — Rules and predicted actions inherit mail.thread for tracked changes and message history.
✓
Admission Smart Fields — Admissions gain a predicted-actions one2many and a computed pending-suggestions count for at-a-glance workload.
Automation, Security & Packaging6 features
✓
Weekly Rejection-Analysis Cron — Scheduled job re-scores all active rules weekly to flag or unflag them for review.
✓
Snooze-Revival Cron — Hourly job returns snoozed actions to suggested state once their snooze-until time passes.
✓
Two-Tier Access Control — All internal users read rules and act on predictions/feedback; system managers get full write/create/unlink.
✓
Phase Gate Marker — Ships a phase4 gate data record coordinating install ordering with the HMS base predictive layer.
✓
Graceful Trigger Failure Handling — Every model hook wraps its predictive trigger in try/except so prediction errors never block core clinical flows.
✓
Odoo 18 App Packaging — Installable priced application (OPL-1, $99) with live-test URL, banner and dependencies on lab, pharmacy, CDS, ward and theatre.
Setup in 4 steps
From install to first live workflow
1 · Install 2 · Review the rule library
Install AminiTech HMS: Predictive Pre-Staging. Required HMS apps (base, consultation, laboratory, pharmacy, CDS, ward, theatre and more) are added to your cart automatically at checkout. Open Clinical > Predictive AI > Rules. The seeded rules ship active with trigger events, condition logic and suggested-action JSON â tune confidence, sequence and facility scope to your site.
3 · Trigger a first prediction 4 · Go live and monitor
Record a diagnosis or admission for a test patient, open Predicted Actions to see suggestions with confidence and evidence, then approve one and watch the lab request appear. Roll out to doctors, then watch acceptance rates on the Rules list. Rules with poor acceptance get flagged for review â retire or refine them using the Rejection Feedback log.
One patient record, one family
The AminiTech HMS family
60+ hospital apps that share one patient record and install together. Required apps are added to your cart automatically at checkout — you only ever download what you have paid for. Or get everything at 15% off with the Complete Suite.
FAQ
Common questions
Which apps does it depend on? Do I have to buy them separately?
It builds on the AminiTech HMS suite: base, consultation, laboratory, pharmacy, clinical decision support, concept dictionary, ward and theatre. Required apps are added to your cart automatically at checkout, so one purchase gets you a working stack.
Does it work on Odoo Community and Enterprise?
Yes â both Community and Enterprise editions of Odoo 18 are supported. It cannot be installed on Odoo Online (SaaS), since that platform does not allow third-party apps; use Odoo.sh or your own server.
Is this a black-box AI? Who is accountable for the orders?
No. Predictions come from transparent, editable rules with visible condition logic, clinical rationale and confidence. Nothing is ordered until a doctor explicitly approves â every action records the reviewer, timestamp and generated record, and clinicians can reject or snooze anything.
Can I write my own prediction rules?
Yes. Rules are regular records: pick a trigger event, write the condition logic, define suggested actions as JSON, set a base confidence and scope it to specific facilities. The engine tracks fired/accepted/rejected counts so you can measure each rule you add.
What support do you provide?
Every purchase includes a 90-day bug-fix guarantee. Write to reach@aminitechsolutions.com with your Odoo version and a description of the issue and we will get you sorted.
Support & Contact
Talk to us today
🏆
90 days of free support included with your purchase
Real people, business hours, typical first reply under 4 hours — including installation help and configuration for your hospital.
⭐ Included — no extra cost
📞
Phone / WhatsApp
+254 746 883 809
💻
Live demo
hms.aminitechsolutions.com
📚
Documentation
aminitechsolutions.com/docs
| Availability |
Odoo Online
Odoo.sh
On Premise
|
| Odoo Apps Dependencies |
•
Discuss (mail)
• Calendar (calendar) • Contacts (contacts) • Employees (hr) • Inventory (stock) • Invoicing (account) |
| Community Apps Dependencies |
Show
•
AminiTech HMS: Clinical Consultation
• AminiTech HMS: Clinical Decision Support (CDS) Engine • AminiTech HMS: Concept Dictionary (SNOMED/LOINC/RxNorm/ICD) • AminiTech HMS: Laboratory & Radiology • AminiTech HMS: Operating Theatre • AminiTech HMS: Patient Management • AminiTech HMS: Pharmacy • AminiTech HMS: Wards & Admissions • AminiTech HMS: Nursing • AminiTech HMS: Patient Billing • AminiTech HMS: Print Subsystem |
| Lines of code | 49124 |
| Technical Name |
aminitech_hms_predictive |
| License | OPL-1 |
| Website | https://aminitechsolutions.com |
Odoo Proprietary License v1.0 This software and associated files (the "Software") may only be used (executed, modified, executed after modifications) if you have purchased a valid license from the authors, typically via Odoo Apps, or if you have received a written agreement from the authors of the Software (see the COPYRIGHT file). You may develop Odoo modules that use the Software as a library (typically by depending on it, importing it and using its resources), but without copying any source code or material from the Software. You may distribute those modules under the license of your choice, provided that this license is compatible with the terms of the Odoo Proprietary License (For example: LGPL, MIT, or proprietary licenses similar to this one). It is forbidden to publish, distribute, sublicense, or sell copies of the Software or modified copies of the Software. The above copyright notice and this permission notice must be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Please log in to comment on this module