Which AI companies serve hospital operations, and what they do
Five layers: the bed and the move, the open shift, the list and the slot, the roster, and the house-wide forecast and capacity, ordered by how far ahead each decision looks.
The layers, from the next hour to the next year
The layers on this page were not set in advance. That gave five layers, one per resource the hospital schedules, and nothing else held: sensing, per-patient prediction, monitoring, rule checking and data infrastructure all turned out to be parts of the products on these five, not products of their own.
- horizon
- minutes to 48 hours
- who buys
- bed management and patient-flow coordinators, the transport office, discharge teams
- in
- admission requests and live bed state from the hospital system (JAG Technology) source; transport requests from the HIS plus porter location (Zhifu Hospital Logistics) source; EHR data for all inpatients across seven hospitals (Hartford HealthCare) source
- out
- the most suitable waiting patient picked for each free bed (JAG Technology) source; work order assigned to the nearest available porter (St. Olavs, MazeMap) source; probability of discharge within 24h and 48h, refreshed every morning (Hartford HealthCare) source
- decision
- porters accept tasks from a prioritised list on mobile (Systematic Columna Flow, Queen Elizabeth University Hospital) source; early discharge planning on the flagged patients (Qventus) source; placements and transport coordinated by flow staff (TeleTracking) source
- core stack
- workflow software 5 · ML 3 · OR 1 (of 11)16 open tech roles at 2 companies: ML / data science 19%, OR 0%
standards
- standards
- HL7 v2 messages, IHE PAM, FHIR Encounter and Locationmixed: open shared standarda big vendor's systempatient movements travel in a neutral spec (HL7 v2, IHE PAM) and FHIR names a bed as a place; who gets which bed, which porter goes and whether a patient is ready to leave are decided inside a vendor's flow product, and no shared format for bed status or a predicted stay was found
- horizon
- hours to a few days
- who buys
- the staffing office, medical staffing and bank teams, nurse managers on the day
- in
- an open or changed shift and the staff pool's availability (CrewLinQ) source; bed demand forecast per ward, 30 days ahead, as the staffing input (Rede Mater Dei with A3Data) source
- out
- open shifts offered to doctors over WhatsApp and the roster confirmed (Revoluna) source; 171,000+ shift hours filled through a collaborative bank across trusts (North West Collaborative Bank, Patchwork blog) source; nursing technician allocation per ward, recommended every 15 minutes (Rede Mater Dei) source
- decision
- the scheduler negotiates within the hospital's limits and confirms the shift itself, the one closed loop on the map (Revoluna) source; medical staffing teams match shifts through the staff bank and float pools shared across trusts (Locum's Nest) source
- core stack
- workflow software 4 · OR 2 · agents 1 (of 7)13 open tech roles at 4 companies: ML / data science 0%, OR 0%
standards
- standards
- NHS e-rostering guidance, Working Time Directiveno shared formatan open shift, a bank booking and an agency rate have no shared format; the hours and rest rules a fill must respect are law text, coded by each vendor
- horizon
- this morning's list to next quarter's blocks
- who buys
- perioperative and surgical services leaders, theatre and clinic schedulers, outpatient managers
- in
- a posted case with its booking data (Duke University Health System) source; past operations and the day's OR plan (nextOR) source; ceiling-camera video of the operating room (Apella) source
- out
- predicted case length at posting: in 33,815 cases 11.2% fewer underpredicted than schedulers (Duke) source; the OR plan recalculated live as cases change (nextOR) source; surgical block utilisation and case volume (LeanTaaS) source; predicted missed appointments, used to plan the clinic around them (DrDoctor) source
- decision
- schedulers see the model's case length inside Epic OpTime and set the slot (Duke) source; block allocations and recovery-room plans adjusted during the day by perioperative leaders (Opmed.ai) source
- core stack
- ML 2 · CV 1 · OR 1 (of 6)1 open tech roles at 4 companies: ML / data science 0%, OR 0%
standards
- standards
- FHIR Schedule and Slot, SNOMED CTmixed: open shared standarda big vendor's systemappointment time has a neutral spec (FHIR Schedule and Slot) and procedures have shared codes (SNOMED CT); a theatre list, a block and a predicted case length live in the theatre module of the record vendor (Duke's model is shown inside Epic OpTime), with no shared output format found
- horizon
- weeks to months ahead, one planning period at a time
- who buys
- head nurses and nursing administrators, rota coordinators, medical staffing and HR
- in
- off and leave requests, night-shift distribution, consecutive-shift limits and skill levels (DutyMaker) source; demand, contracts, skills and labour rules (ORTEC Workforce) source; Working Time Directive and Junior Doctor 2016 Contract (Rotageek) source
- out
- a month's labour-law-compliant roster in about 25 to 30 seconds (Yuan Rung Medical System) source; several roster drafts with a fairness score and the 72-hour night-shift check for the Form 9 filing (Nurse is) source; rotas validated automatically against the Directive and the 2016 contract (Rotageek) source
- decision
- head nurses pick a draft and publish it (Nurse is; Yuan Rung names head nurses as the buyer) source; staff state preferences by smartphone and exchange shifts in the app, subject to the rules (C-Live Hospital Touch) source
- core stack
- OR 8 · workflow software 7 (of 15)4 open tech roles at 3 companies: ML / data science 0%, OR 25%
standards
- standards
- NHS e-rostering guidance, Working Time Directive, FHIR (no roster resource)mixed: no shared formata big vendor's systemthe roster lives in each trust's e-rostering software; the Directive and the contracts are law text each vendor codes its own way; no roster or shift resource among the 157 FHIR R5 resources
- horizon
- the next 72 hours to the next year
- who buys
- hospital operations leadership and the command centre, planners, pharmacy and procurement, regional commissioners
- in
- real-time data from multiple hospital IT systems (Bradford command centre, GE HealthCare) source; data from procurement to dispensing (Omnicell) source; admissions, discharges and length of stay across two acute trusts, a community provider and three local authorities (BNSSG IPACS) source
- out
- capacity constraints forecast up to 72 hours ahead (GE HealthCare CareIntellect for Operations) source; 480 regularly updated forecasts of census, ICU space, staffing and supply across medical centres (Kaiser Permanente) source; stock levels and order quantities from the Inventory Optimization Service (Omnicell) source; twelve demand, capacity and length-of-stay scenarios per locality and pathway (BNSSG IPACS) source
- decision
- command-centre staff decide on patient flow from one room (Bradford; Johns Hopkins) source; commissioners use the scenarios for intermediate care commissioning and a GBP 13m business case (BNSSG ICB) source; unused theatre slots reallocated across departments from the utilisation view (Solamed OpePro) source
- core stack
- analytics 6 · workflow software 4 · ML 1 (of 13)59 open tech roles at 1 company: ML / data science 66%, OR 0%
standards
- standards
- Epic's record and its command centre, GS1 in healthcare (stock)mixed: a big vendor's systemno shared formatno standard for occupancy, utilisation or capacity measures was found; the command centre sits inside the record vendor's product (Epic, Oracle and GE on the map); stock carries GS1 product numbers; regional capacity crosses walls by dashboard and letter, with no shared format
1 of the companies here carries the between-organisations flag: NHS Bristol, North Somerset and South Gloucestershire ICB. It is a flag, not a layer; the between organisations tab has the cases.
13 companies with this as their main layer →Problems
A record lists the problems its product or unit addresses, as stated on a page the researcher read; the first one is its main problem. This is the grounded view next to the layer view: for each problem, how many of the 52 companies claim it, from which group, with what technical core and sold how.
| problem | layer | horizon | companies | by group | technical core | delivery model | classic method |
|---|---|---|---|---|---|---|---|
| Ambulance dispatch and relocation | H1 | minutes to 48 hours | - | - | - | dynamic relocation policies and simulation | |
| Bed assignment and placement | H1 | minutes to 48 hours | 4 of 52 | start-up 4 (of 4) | OR 2 · analytics 2 · workflow software 2 (of 4 with a core recorded) | software only 3 (of 3 with a delivery model recorded) | heuristics, tabu search and rules |
| Discharge readiness and discharge planning | H1 | minutes to 48 hours | 6 of 52 | vendor 4 · in-house 2 (of 6) | ML 4 · analytics 4 · workflow software 4 (of 6 with a core recorded) | software only 2 · in-house, not sold 2 (of 4 with a delivery model recorded) | classification and rules |
| ED triage queue ordering and patient routing | H1 | minutes to 48 hours | - | - | - | queueing policies and machine learning | |
| Length-of-stay prediction | H1 | minutes to 48 hours | 2 of 52 | vendor 1 · in-house 1 (of 2) | analytics 2 · ML 1 · streaming 1 (of 2 with a core recorded) | software only 1 · in-house, not sold 1 (of 2 with a delivery model recorded) | regression and gradient boosting |
| Porter and patient transport dispatch | H1 | minutes to 48 hours | 4 of 52 | start-up 2 · vendor 2 (of 4) | workflow software 3 · ML 1 · OR 1 (of 4 with a core recorded) | software only 2 · software + hardware 1 (of 3 with a delivery model recorded) | dispatching rules and assignment |
| RTLS and bed sensing | H1 | minutes to 48 hours | 2 of 52 | start-up 1 · vendor 1 (of 2) | ML 1 · OR 1 · analytics 1 (of 2 with a core recorded) | software only 2 (of 2 with a delivery model recorded) | sensor infrastructure |
| Agency, bank and float-pool allocation | H2 | hours to a few days; day of operations | 5 of 52 | start-up 4 · acquired 1 (of 5) | workflow software 5 · OR 1 · agents 1 (of 5 with a core recorded) | software only 5 (of 5 with a delivery model recorded) | matching, auctions and rules |
| Re-rostering and shift fill | H2 | hours to a few days; day of operations | 11 of 52 | start-up 8 · vendor 1 · in-house 1 · acquired 1 (of 11) | workflow software 8 · analytics 5 · ML 4 (of 11 with a core recorded) | software only 9 · in-house, not sold 1 (of 10 with a delivery model recorded) | rules, matching and heuristics |
| Overtime and agency spend attribution | H2 | hours to a few days | - | - | - | descriptive analytics | |
| Infusion centre chair and nurse scheduling | H3 | this morning's list to next quarter's blocks; tactical (months) | 1 of 52 | vendor 1 (of 1) | ML 1 · analytics 1 · workflow software 1 (of 1 with a core recorded) | software + heavy services 1 (of 1 with a delivery model recorded) | integer programming and heuristics |
| Outpatient clinic template planning | H3 | this morning's list to next quarter's blocks; tactical (months) | 1 of 52 | start-up 1 (of 1) | ML 1 · analytics 1 · workflow software 1 (of 1 with a core recorded) | software only 1 (of 1 with a delivery model recorded) | simulation and integer programming |
| Theatre master surgical schedule and block allocation | H3 | this morning's list to next quarter's blocks; tactical (months) | 4 of 52 | start-up 3 · vendor 1 (of 4) | ML 3 · OR 3 · workflow software 2 (of 4 with a core recorded) | software only 3 · software + heavy services 1 (of 4 with a delivery model recorded) | integer programming and simulation |
| Same-day theatre re-scheduling | H3 | this morning's list to next quarter's blocks; day of operations | 3 of 52 | start-up 3 (of 3) | ML 2 · OR 2 · agents 1 (of 3 with a core recorded) | software only 3 (of 3 with a delivery model recorded) | rules, heuristics and real-time optimisation |
| Appointment no-show prediction | H3 | this morning's list to next quarter's blocks | 1 of 52 | start-up 1 (of 1) | ML 1 · analytics 1 · workflow software 1 (of 1 with a core recorded) | software only 1 (of 1 with a delivery model recorded) | logistic regression and gradient boosting |
| Elective waiting-list balancing and mutual aid | H3 | this morning's list to next quarter's blocks | - | - | - | rules and manual coordination | |
| Surgical case duration prediction | H3 | this morning's list to next quarter's blocks | 4 of 52 | start-up 3 · in-house 1 (of 4) | ML 4 · OR 2 · CV 1 (of 4 with a core recorded) | software only 2 · software + hardware 1 · in-house, not sold 1 (of 4 with a delivery model recorded) | gradient boosting |
| Nurse rostering | H4 | weeks to months ahead, one planning period at a time; tactical (months) | 19 of 52 | start-up 11 · vendor 5 · in-house 2 · acquired 1 (of 19) | workflow software 14 · OR 9 · ML 4 (of 19 with a core recorded) | software only 14 · software + hardware 1 · software + heavy services 1 · in-house, not sold 2 (of 18 with a delivery model recorded) | integer programming and metaheuristics |
| Physician and junior doctor scheduling | H4 | weeks to months ahead, one planning period at a time; tactical (months) | 5 of 52 | start-up 3 · vendor 1 · acquired 1 (of 5) | OR 4 · workflow software 3 · ML 1 (of 5 with a core recorded) | software only 5 (of 5 with a delivery model recorded) | constraint programming and rules |
| Self-rostering and shift swap | H4 | weeks to months ahead, one planning period at a time | 5 of 52 | start-up 2 · vendor 3 (of 5) | workflow software 4 · OR 2 · analytics 1 (of 5 with a core recorded) | software only 4 · software + heavy services 1 (of 5 with a delivery model recorded) | preference elicitation and rules |
| Working-time and collective-agreement rule checking | H4 | weeks to months ahead, one planning period at a time | 8 of 52 | start-up 5 · vendor 1 · in-house 1 · acquired 1 (of 8) | OR 6 · workflow software 6 · ML 1 (of 8 with a core recorded) | software only 7 · in-house, not sold 1 (of 8 with a delivery model recorded) | rule engines |
| Capacity and bed planning | H5 | the next 72 hours to the next year; strategic (years) | 2 of 52 | start-up 1 · in-house 1 (of 2) | OR 1 · analytics 1 · simulation 1 (of 2 with a core recorded) | software only 1 · in-house, not sold 1 (of 2 with a delivery model recorded) | queueing models and simulation |
| Pharmacy and supply ordering | H5 | the next 72 hours to the next year; tactical (months) | 3 of 52 | start-up 1 · vendor 2 (of 3) | analytics 2 · workflow software 2 · ML 1 (of 3 with a core recorded) | software only 2 · software + hardware 1 (of 3 with a delivery model recorded) | inventory theory, (s,S) and par-level rules |
| Sterile goods (CSSD) capacity and instrument tray planning | H5 | the next 72 hours to the next year; tactical (months) | - | - | - | goal programming, scheduling | |
| Emergency department arrival forecast | H5 | the next 72 hours to the next year; day of operations | 1 of 52 | start-up 1 (of 1) | ML 1 · analytics 1 (of 1 with a core recorded) | software only 1 (of 1 with a delivery model recorded) | time series (SARIMA) and gradient boosting |
| ICU inflow forecast for staffing | H5 | the next 72 hours to the next year; day of operations | 1 of 52 | in-house 1 (of 1) | analytics 1 · simulation 1 (of 1 with a core recorded) | in-house, not sold 1 (of 1 with a delivery model recorded) | machine learning on admissions data |
| Ambulance diversion and regional capacity visibility | H5 | the next 72 hours to the next year | - | - | - | dashboards and rules, no optimisation reported | |
| Discharge pathway capacity across organisations | H5 | the next 72 hours to the next year | 1 of 52 | in-house 1 (of 1) | simulation 1 (of 1 with a core recorded) | in-house, not sold 1 (of 1 with a delivery model recorded) | discrete-event or discrete-time simulation with scenario analysis |
| Occupancy and utilisation dashboards | H5 | the next 72 hours to the next year | 13 of 52 | start-up 4 · vendor 4 · in-house 4 · acquired 1 (of 13) | analytics 12 · workflow software 7 · ML 6 (of 13 with a core recorded) | software only 8 · software + heavy services 1 · in-house, not sold 4 (of 13 with a delivery model recorded) | descriptive analytics |
| EHR and operations data layer | no layer | - | - | - | data engineering |
Every name links to its row in the Companies tab; the count opens the Companies tab filtered to that problem. A dash means no record has the field recorded.
Where the standard lives
A decision's data format lives in one of three places: an open standard, a big vendor's system, or no shared format. Where an open standard exists, a new tool can plug in quickly. Where one vendor owns the record or the control interface, such as the electronic health record (EHR) or the bed-management or e-rostering system, that vendor tends to keep the margin. Where nothing exists, every customer is a custom integration project.
Terms used here (31), one line each
Systems, data and ward words
- EHR
- electronic health record: the hospital's patient record system.
- HIS
- hospital information system: the hospital's administrative and patient record system, the usual name in East Asia and Europe.
- RTLS
- real-time location system: tags on equipment, staff or patients that show where they are.
- ADT
- admission, discharge, transfer: the events that move a patient into, through and out of a hospital.
- e-rostering
- software that builds and publishes staff rotas against working-time rules.
- command centre
- a control room that shows a hospital's or region's beds, admissions and staffing live.
- Grand Central
- Epic's patient flow module (bed capacity, admissions, discharge planning).
- roster
- the schedule of who works which shift.
- theatre list
- the ordered list of operations planned for one operating theatre and day.
- block
- a recurring slot of theatre time allocated to one specialty or surgeon.
- staff bank
- a hospital's or region's own pool of temporary staff, booked shift by shift instead of through an agency.
- float pool
- staff who are not fixed to one ward and are sent where the day's gaps are.
- ICB
- integrated care board: the NHS body that plans and commissions care for an area of England.
- mutual aid
- one hospital taking waiting-list patients from another to even out waiting times.
- UDI
- unique device identification: a code that identifies a medical device, required by law in several regions.
Standards named in this section and on the layer cards
- HL7 v2
- the message format hospital systems have used since 1987 to send admissions, orders and results to each other.
- FHIR
- HL7's newer format that sends health data as web resources (Patient, Encounter, Location); R4 is the release the US rule names, R5 the current one.
- IHE profile
- an agreed recipe for using standards such as HL7 v2 for one job, for example moving patients between wards (PAM).
- ICD-11
- the WHO's disease classification, in effect since 1 January 2022.
- GTIN, GS1
- GS1's product number and barcode; GS1 is also an issuing agency for device identification rules.
- Working Time Directive
- EU directive 2003/88/EC on weekly hours and rest periods.
- Cures Act rule
- the 2020 US rule that put FHIR Release 4 APIs into EHR certification.
Who publishes them
- HL7
- Health Level Seven International, a not-for-profit standards body founded in 1987; it publishes HL7 v2 and FHIR.
- IHE
- Integrating the Healthcare Enterprise, an initiative of healthcare professionals and industry that publishes profiles.
- SNOMED International
- the London-based body that maintains SNOMED CT.
- WHO
- the World Health Organization, which publishes ICD.
- GS1
- the global body behind barcodes and product numbers.
- NHS England and NHS Improvement
- the bodies that issued the 2020 e-rostering guidance for NHS trusts.
- ONC
- the US Office of the National Coordinator for Health IT, within HHS.
- Epic
- a US vendor of hospital record software; a vendor, not a standards body.
HL7 v2 (messages between hospital systems), FHIR (health data over web APIs), IHE PAM (patient movements between beds and wards), SNOMED CT and ICD-11 (shared clinical and disease terms)
for whoever builds on top Anyone can plug in on the same terms. HL7 says 95% of US healthcare organizations use its version 2 messages, first released in 1987. These formats move patients and records between systems, not staff hours or free beds. hl7.org
Epic's hospital record, with its bed and capacity module (Epic: US vendor of hospital records; Grand Central is its patient flow module), the e-rostering software each NHS trust runs (e-rostering: software that builds and publishes staff rotas)
for whoever builds on top Owned, not agreed. Epic says more than 325 million patients have a current record in Epic, and its bed and capacity tools sit inside that record. A new tool goes through the owner, who sells its own flow tools too. epic.com epic.com england.nhs.uk
the roster itself, and the hours and rest rules a roster must meet, which beds are free right now, and what counts as full, how many beds, ambulances and waiting patients a whole region has, what a bed sensor or a location tag reports
for whoever builds on top Each deployment starts by collecting the data by hand or by reading rule text. Whoever gets a format adopted would own it, but only if competitors adopt it too. hl7.org researchinformation.umcutrecht.nl england.nhs.uk
Layer by layer
For each layer: which standards already exist, and in plain words where the agreement lives.
| layer | standards that already exist | where it lives |
|---|---|---|
| H1 The bed and the move | HL7 v2 messages (messages between hospital systems, since 1987), IHE PAM (which messages move a patient between beds and wards), FHIR Encounter and Location (a patient's stay and the places in it, including a bed) | mixed: open shared standarda big vendor's system patient movements travel in a neutral spec (HL7 v2, IHE PAM) and FHIR names a bed as a place; who gets which bed, which porter goes and whether a patient is ready to leave are decided inside a vendor's flow product, and no shared format for bed status or a predicted stay was found |
| H2 The open shift | NHS e-rostering guidance (the NHS's maturity ladder for electronic rosters), Working Time Directive (EU limits on weekly hours and rest) | no shared format an open shift, a bank booking and an agency rate have no shared format; the hours and rest rules a fill must respect are law text, coded by each vendor |
| H3 The list and the slot | FHIR Schedule and Slot (free and busy time that appointments book), SNOMED CT | mixed: open shared standarda big vendor's system appointment time has a neutral spec (FHIR Schedule and Slot) and procedures have shared codes (SNOMED CT); a theatre list, a block and a predicted case length live in the theatre module of the record vendor (Duke's model is shown inside Epic OpTime), with no shared output format found |
| H4 The roster | NHS e-rostering guidance (the NHS's maturity ladder for electronic rosters), Working Time Directive (EU limits on weekly hours and rest), FHIR (no roster resource) (modern web-API format for health data) | mixed: no shared formata big vendor's system the roster lives in each trust's e-rostering software; the Directive and the contracts are law text each vendor codes its own way; no roster or shift resource among the 157 FHIR R5 resources |
| H5 The house-wide forecast and capacity | Epic's record and its command centre (one vendor's record that also exposes standard interfaces), GS1 in healthcare (stock) (product numbers and barcodes on medicines and devices) | mixed: a big vendor's systemno shared format no standard for occupancy, utilisation or capacity measures was found; the command centre sits inside the record vendor's product (Epic, Oracle and GE on the map); stock carries GS1 product numbers; regional capacity crosses walls by dashboard and letter, with no shared format |
Each standard links to a page it was checked on; hover for the body and the year. Every acronym is spelled out under "Terms used here" above.
What drives standardization
Three things push a health system onto a standard: a law, a dominant vendor whose record everyone else reads, or an industry body such as HL7 publishing a format. Across the 3 cases below, the push came 1 from a dominant vendor, 1 from regulation, 1 from a standards body. The 3 clearest:
The US wrote FHIR version 4 into the certification that hospital record systems must hold. It covers patient data, not staff rosters or bed capacity.
evidence The rule names FHIR Release 4 as the standard; it covers patient data access, not rostering, beds or capacity. No count of live hospital APIs was found.
who US HHS Office of the National Coordinator for Health IT
federalregister.gov federalregister.gov healthit.gov open.epic.com
One vendor's hospital record holds the patients, and its flow module holds the beds. Other tools connect through the interfaces that vendor publishes.
evidence Epic says more than 325 million patients have a current record in Epic and that its capacity command centre is part of that record; it also publishes interfaces for outside tools. These are Epic's own figures. No count of hospitals using the flow module was found.
who Epic Systems Corporation
A neutral standard from 1987 now moves patient events almost everywhere in US healthcare. Nothing similar was found for staff rosters, so each rostering product writes its own rules.
evidence HL7 says 95% of US healthcare organizations use HL7 v2.x, a standard for moving patients between systems. Its FHIR resource index has no roster or shift resource, and rostering rules are coded by each vendor.
who HL7 International
hl7.org hl7.org hl7.org profiles.ihe.net england.nhs.uk rotageek.com
What should be standardized next
The fourth has no layer on this map: it was written for the old "knowledge and people" layer, which no company on the map sells as a product, so the layer was folded away and the gap is kept as a gap.
- H5 Hospitals telling each other how full they are
Today a region sees capacity through one dashboard or one letter. There is no shared format for how many beds, ambulances and waiting patients each hospital has.
Dutch emergency departments use a traffic-light diversion dashboard (62 of 82 EDs); NHS England asks trusts to refer suitable waiting-list patients to another provider by letter. No shared format found; one in-house model on this map crosses organisations (BNSSG IPACS), no product does.
On this map 1 of 52 companies carry the between-organisations flag (NHS Bristol, North Somerset and South Gloucestershire ICB, 1 of them established); no product does.
- H4 Rosters and the hours rules, written so any tool can read them
The law says how many hours and how much rest. Every rostering product turns that into its own rules. No shared format for a roster or a shift was found.
Law: 48 hours average weekly, 11 hours daily rest, 24 hours weekly rest (Directive 2003/88/EC). NHS England's 2020 guidance has trusts put these rules into rostering software. FHIR's 157 resources include no roster or shift resource.
eur-lex.europa.eu england.nhs.uk hl7.org hl7.org rotageek.com
- H1 Which beds are free, defined the same way everywhere
Standards describe a patient moving between rooms. They do not define a free, dirty or blocked bed, so each flow product defines its own.
FHIR can name a bed as a place and follow a stay from admission to discharge; IHE PAM sends the movements. No shared definition of bed status or occupancy was found.
- no layer Hospital protocols a program can read and cite
A computable form for protocols exists in FHIR, with the guideline guide still in draft. No example of a tool that reads protocols this way was found.
FHIR PlanDefinition covers protocols and decision rules; the CPG guide is a ballot build. No retrieval tool over protocols found.
This item has no layer on the map: no company sells retrieval over protocols, which is why the old 'knowledge and people' layer was folded away.
Who sits where
Each company appears once, on its main layer; tick the box to also see it, faded, on its secondary layers. The chip style says how large the company is and how public its customer record is, by the rules in the legend; the small tag says which group it belongs to.
| group | H1 | H2 | H3 | H4 | H5 | all |
|---|---|---|---|---|---|---|
| startup | 6 | 5 | 4 | 8 | 3 | 26 |
| incumbent vendor | 4 | - | 1 | 5 | 5 | 15 |
| operator in-house unit | 1 | 1 | 1 | 1 | 4 | 8 |
| acquired startup | - | 1 | - | 1 | 1 | 3 |
| all groups | 11 | 7 | 6 | 15 | 13 | 52 |
Counts by main layer (the first layer listed). A dash is a zero.
Between organisations is a flag on a company, not a layer: 1 of 52 companies carry it (NHS Bristol, North Somerset and South Gloucestershire ICB on H5). A layer needs at least five companies or three startups; this has one in-house team and no product, so it stays a flag and a section.
Layer × who closes the loop
Each dot is one company at its main layer. Dot size and fill follow the same size tiers as the map.
Companies by layer
| National University Health System (ENDEAVOUR AI platform) | H1H5 | Singapore public health cluster's in-house AI platform that predicts hospital length of stay, supports discharge and shows the bed situation across five institutions in a command-centre view. | in-house operator in-house unit |
| patient flow and beds | ML · streaming · analyticswhyML models for stay duration and discharge on an EHR data stream (Kafka) with Spotfire command-centre views. sources spotfire.com | Established | reads EHR, workforce or theatre systems | - | spotfire.com | |
| Oracle Health health system operations (Patient Flow, Command Center) | H1H5 | Oracle Health suite for patient throughput, clinical location awareness (tracking patients, clinicians and equipment) and capacity visibility. | vendor incumbent vendor |
| patient flow and beds | analytics · MLwhyPatient throughput, location awareness and capacity visibility. hiring note Oracle careers are company-wide; no postings scoped to this product line counted. sources oracle.com | detaildelivery Suite sold as software within Oracle Health; no hardware or services model read. | Established | reads RTLS, bed or device signals | - | oracle.com |
| Qventus | H1 | AI software that automates hospital care operations: surgical growth, perioperative care coordination, inpatient capacity and discharge. | vendor incumbent vendor |
| patient flow and beds | ML · agents · workflow softwarewhyAI, machine learning and behavioural science applied to care operations (surgical, inpatient capacity, discharge); the about page describes automation delivered to frontline teams and AI agents. hiring 8 open tech roles: data science 2, data eng 1, software 1, solutions / deployment 1, other 3 boards-api.greenhouse.io. Counted: Senior Data Scientist x2, Senior Data Engineer, Sr. Software Engineer PAT Assist, Forward Deployed AI Solutions Architect; other: AI Tooling Engineer, Senior Analytics Specialist, Technical Product Architect. No role titled ML engineer or optimisation.) sources qventus.com qventus.com | - | Established | reads EHR, workforce or theatre systems | - | qventus.com qventus.com leadiq.com |
| Systematic Columna Flow | H1 | Porter and service task management where staff see a prioritised task list on mobile and accept tasks, with management reports for resource planning. | vendor incumbent vendor |
| patient flow and beds | workflow softwarewhyMobile prioritised task list for porters and service staff with management reports; no optimiser claimed in the case read. hiring note jobs.systematic.com lists 54 roles for the whole company (). sources systematic.com | detaildelivery Software product in a case study; no hardware or services model read. | Established | reads EHR, workforce or theatre systems | - | systematic.com |
| Navenio | H1 | Smartphone location and tasking for hospital teams: real-time location of staff, patients and equipment with ride-share-like task allocation, used for porter dispatch. | start-up startup |
| patient flow and beds | edge / IIoT · ORwhySmartphone-sensor indoor location of staff and equipment (no new infrastructure) with ride-share-like task allocation for porters. | detaildelivery Site states location works from smartphone sensors with no investment in new infrastructure; OEM licensing also offered. | Customer track record | reads RTLS, bed or device signals | - | pulse2.com navenio.com |
| Zhifu Hospital Logistics (智辅医院后勤) | H1H4 | Hospital logistics platform whose central transport module takes requests from the HIS, automatically matches porters, plans routes and tracks every delivery, with an AI staff-scheduling module. | start-up startup |
| supply and pharmacy | workflow softwarewhyWork-order and dispatch platform with automatic porter matching, route planning, GPS tracking and HIS integration. sources hsmsx.com | detailbuyer hospital logistics departments and facility service companies The site says the system is integrated software and hardware (Bluetooth, NFC, QR confirmation) and offers a download and price page. | Customer track record | assigns a bed, dispatches a porter or re-sequences a list | - | hsmsx.com hsmsx.com hsmsx.com |
| JAG Technology | H1H3H5 | Taiwanese smart-hospital vendor whose AI admission scheduling and bed control system picks the most suitable waiting patient for each free bed, alongside an operating-theatre management system with scheduling, dynamic dispatch and BI dashboards. | start-up startup |
| several | workflow software · analyticswhyAdmission and theatre management workflows with ranking of waiting patients per bed and BI dashboards; the algorithm is not described. sources jag-tec.com | Early | reads EHR, workforce or theatre systems | - | jag-tec.com hst.org.tw | |
| Medigital (rebranded Auta Health in May 2024) | H1H5 | Hospital capacity management software whose bed-allocation module uses AI algorithms to assign patients to beds. | start-up startup |
| patient flow and beds | ORwhyhiring note No careers page found. medigital.de resolves to an unrelated one-person software shop in Balingen. sources munich-startup.de | Early | reads EHR, workforce or theatre systems | - | munich-startup.de | |
| UnityAI | H1 | AI platform that predicts patient needs, finds flow bottlenecks and manages bed assignments, using reinforcement learning and classical optimisation. | start-up startup |
| patient flow and beds | ML · ORwhyPress articles state it uses reinforcement learning and classical optimisation to predict patient needs, find flow bottlenecks and manage bed assignments. sources hitconsultant.net pulse2.com | - | Early | reads EHR, workforce or theatre systems | - | hitconsultant.net pulse2.com |
| WiseMed (BedWise) | H1H5 | Chilean hospital software: BedWise bed management with real-time occupancy and a traceable bed-assignment workflow, deployed in a public hospital of more than 400 beds; QxWise surgical scheduling is in development. | start-up startup |
| patient flow and beds | workflow software · analyticswhyBed-management workflow with real-time occupancy visibility and traceability; AI is claimed only in general terms. hiring note No careers page found on wisemedchile.com. sources expo-salud.cl | detailThe site offers Chilean public hospitals help applying through the Compra Publica de Innovacion route. | Early | reads EHR, workforce or theatre systems | - | expo-salud.cl wisemedchile.com |
| TeleTracking Operations IQ | H1 | Privately held patient-flow platform: referrals, placements, transport, discharge planning and network coordination across inpatient and ambulatory settings. | vendor incumbent vendor |
| patient flow and beds | workflow software · MLwhyPatient-flow platform for placement, transport and discharge coordination. hiring 8 open tech roles: ML 1, software 4, solutions / deployment 1, other 2 boards-api.greenhouse.io. Counted: Machine Learning Engineer III; Director Software Engineering, Software Engineer I, Software Engineer II, Software Engineer II Quality Engineering; Forward Deployed Engineer; other: Platform Architect, QA Release Manager. sources teletracking.com teletracking.com | - | reads EHR, workforce or theatre systems | - | teletracking.com teletracking.com | |
| Rede Mater Dei (Gestao de Escala project with A3Data) | H2H4H5 | Hospital network's nurse staffing system, built with the data consultancy A3Data, that forecasts bed demand per ward up to 30 days ahead and recommends nursing technician allocation per ward every 15 minutes from an optimisation model. | in-house operator in-house unit |
| workforce and rostering | OR · ML · analyticswhyOptimisation model proposing rosters under business rules, a probabilistic bed-demand forecast for 1 to 30 days, and dashboards. sources sindhosfil.com.br | detailbuyer nursing management | Established | reads EHR, workforce or theatre systems | - | sindhosfil.com.br |
| Hublo | H2H4 | HR platform for hospitals and care homes: planning and time management, replacement shift filling from a staff pool, recruitment. | start-up startup |
| workforce and rostering | workflow softwarewhyStaff planning, replacement shift filling from a staff pool and SIRH integration. hiring 4 open tech roles: data eng 1, software 3 careers.hublo.com.Counted: Engineering Manager x2 and Tech Lead SRE (software 3), Data Engineer Interop. sources hublo.com | detailbuyer health facility HR and care-unit managers (site copy on staffing visibility for managers) sources hublo.com | Customer track record | reads EHR, workforce or theatre systems | 190 company site: 'plus de 190 collaborateurs en France et a l'international' (lower bound, https://hublo.com) | hublo.com tech.eu |
| Revoluna | H2 | Physician shift platform for Brazilian hospitals whose AI scheduler (Julia) offers open shifts to doctors over WhatsApp, negotiates within the hospital's limits and confirms the roster. | start-up startup |
| workforce and rostering | agents · workflow softwarewhyMarketplace for physician shifts with an LLM-style agent (Julia) that negotiates and confirms shifts over WhatsApp. sources revoluna.com.br | detailbuyer hospital managers who post shifts (gestores) | Customer track record | reads EHR, workforce or theatre systems | - | revoluna.com.br |
| Chromie Health | H2 | AI workforce software for healthcare: automated shift filling (Dispatch), staffing forecasts (Predict) and workforce analytics. | start-up startup |
| workforce and rostering | workflow software · ML · analyticswhyDispatch automates shift-fill outreach and confirmation, Predict forecasts patient demand and coverage gaps, Analytics reports workforce data; Epic and ADT feeds. hiring 0 open tech roles chromiehealth.com. sources chromiehealth.com | detailbuyer nurse managers and nursing leadership (site quotes an associate CNO) sources chromiehealth.com | Early | reads EHR, workforce or theatre systems | - | chromiehealth.com |
| CrewLinQ | H2H4 | Absence and float-pool management app for hospitals and care providers: fills open shifts in minutes, steers an internal float pool, handles shift swaps and generates rosters with AI. | start-up startup |
| workforce and rostering | workflow softwarewhyApp-based absence, float-pool and swap workflows with an AI roster add-on; the method behind the roster generator is not described. sources crewlinq.com | detailA pricing page (Preise) is linked from the menu. | Early | reads EHR, workforce or theatre systems | - | crewlinq.com crewlinq.com de.linkedin.com |
| Vitalize | H2H4 | AI platform for health systems that unifies staffing, scheduling, flow and capacity, with a command center for staffing offices. | start-up startup |
| workforce and rostering | OR · MLwhyUnifies staffing, scheduling, flow and capacity with optimisation and time-series prediction; careers page states 30x growth in under 24 months, dozens of hospitals. hiring 6 open tech roles: software 4, solutions / deployment 2 api.ashbyhq.com. Counted: Staff Engineer, Principal Engineer, Senior Full Stack Engineer, Senior Product Engineer; two Enterprise Solutions Engineer roles (clinical and EHR; workforce management and payroll). No ML, data or optimisation roles open.) sources vitalize.care vitalize.care | -detaildelivery No delivery model asserted. | Early | reads EHR, workforce or theatre systems | - | vitalize.care pulse2.com |
| Locum's Nest | H2 | Mobile shift-matching and staff-bank platform for NHS trusts, with digitally-powered float pools shared across trusts and regions. | acquired acquired startup |
| workforce and rostering | workflow software · analyticswhyShift-matching and staff-bank platform with an Intelligence product for spend analytics and integrations to e-rostering and payroll. hiring 3 open tech roles: software 3 locumsnest.co.uk (Careers page lists 3 vacancies: Web Engineer (Frontend), Backend Engineer, Full Stack Engineer.) sources locumsnest.co.uk | detailbuyer NHS trust medical staffing, rota and HR teams (quotes from a rota coordinator, a service manager, a CEO and an HR director) Home page says the software pays for itself through savings; no price is stated. sources locumsnest.co.uk | reads EHR, workforce or theatre systems | - | locumsnest.co.uk silicon.co.uk | |
| Duke University Health System (Department of Surgery with Duke Health Technology Solutions) | H3 | Duke's own surgery and data teams built and run a gradient-boosting surgical case length model, shown to schedulers inside Epic OpTime. | in-house operator in-house unit |
| theatre, outpatient and infusion | MLwhyGradient-boosting surgical case length model with a similarity cascade and custom loss function, deployed in scheduling and evaluated prospectively on 33,815 cases. hiring note No postings naming the unit were found. sources pmc.ncbi.nlm.nih.gov | detailbuyer Duke operating-room schedulers and perioperative leadership (the model output is shown to schedulers) sources pmc.ncbi.nlm.nih.gov | Established | reads EHR, workforce or theatre systems | - | pmc.ncbi.nlm.nih.gov |
| LeanTaaS | H3 | Prescriptive analytics for hospital capacity: operating rooms, infusion centers and inpatient flow. | vendor incumbent vendor |
| theatre, outpatient and infusion | analytics · ML · workflow softwarewhyiQueue products for ORs, infusion centres and inpatient flow described as predictive and prescriptive analytics, ML and generative AI. hiring 0 open tech roles leantaas.com. sources leantaas.com leantaas.com | detailHomepage offers 'Transformation as a Service' with a dedicated engagement team that implements the technology; ROI figures per OR, bed and chair per year are claims, not prices. No buying role named. sources leantaas.com leantaas.com | Established | reads EHR, workforce or theatre systems | - | leantaas.com leantaas.com leadiq.com |
| DrDoctor | H3 | Outpatient scheduling and patient engagement on top of the hospital EPR: self-booking into live slots, prediction of missed appointments to plan clinics around it, waiting list validation. | start-up startup |
| theatre, outpatient and infusion | workflow software · ML · analyticswhyScheduling Suite writes self-booked slots into the EPR, Smart Centre predicts missed appointments to plan clinics, plus waiting-list validation workflows. hiring 0 open tech roles apply.workable.com. sources drdoctor.co.uk drdoctor.co.uk | detailbuyer NHS trust operational leaders; site has role pages for ops leaders, admin teams, clinical teams and data/IT teams The site also sells an on-site Delivery Partner service where part of the fee depends on outcomes; treated as an add-on to the software, not the main product. sources drdoctor.co.uk | Customer track record | reads EHR, workforce or theatre systems | - | drdoctor.com |
| Apella | H3 | Operating-room intelligence platform using ceiling-mounted cameras and ML to detect perioperative events and forecast case duration. | start-up startup |
| theatre, outpatient and infusion | CV · ML · workflow softwarewhyComputer vision models on live video from ceiling-mounted cameras detect 14 perioperative events, and ML case-duration predictions feed live boards and EHR write-back. hiring 0 open tech roles jobs.ashbyhq.com. | Early | reads RTLS, bed or device signals | - | apella.io | |
| nextOR | H3 | Self-learning operating-room scheduling software that predicts case duration and recalculates the OR plan live. | start-up startup |
| theatre, outpatient and infusion | ML · ORwhySelf-learning OR-scheduling software that predicts case duration and recalculates the OR plan live (next-or.de). hiring 1 open tech role: software 1 next-or.de. sources next-or.de ki-syndikat.de | detaildelivery Software platform for OR resource planning per next-or.de. | Early | reads EHR, workforce or theatre systems | - | next-or.de ki-syndikat.de |
| Opmed.ai | H3 | Agentic AI platform that forecasts case duration, demand and staffing needs and builds and adjusts operating-room schedules, block allocations and recovery-room plans during the day. | start-up startup |
| theatre, outpatient and infusion | OR · ML · agentswhyForecasts of case duration and demand feed a scheduler that searches very large numbers of schedule permutations, wrapped as AI agents that adjust the day. sources opmed.ai finance.yahoo.com | detailbuyer perioperative and surgical services leaders | Early | reads EHR, workforce or theatre systems | - | opmed.ai finance.yahoo.com technologyrecord.com |
| HealthStream ShiftWizard | H4 | Nurse and staff scheduling with self-scheduling and shift swaps on mobile, demand forecasting and multi-facility dashboards, with Workday-certified integration. | vendor incumbent vendor |
| workforce and rostering | workflow software · analyticswhyScheduling with self-scheduling, mobile shift swaps, predictive scheduling tools and Workday-certified integration. sources healthstream.com | detailsources healthstream.com | Established | reads EHR, workforce or theatre systems | - | healthstream.com |
| UKG (Pro Workforce Management for healthcare) | H4H2 | Workforce management suite with healthcare scheduling modules: Advanced Scheduler (forecasting, rule-based scheduling and real-time shift management under ratio and fatigue rules), Clinical Scheduling Extensions (staffing to patient workload) and EZCall for physician schedules. | vendor incumbent vendor |
| workforce and rostering | workflow software · analyticswhyWorkforce management suite with forecasting and rule-based scheduling modules for healthcare. sources finance.yahoo.com | detailsources finance.yahoo.com | Established | reads EHR, workforce or theatre systems | - | finance.yahoo.com ukg.com ukg.com |
| Yuan Rung Medical System (nurse rostering system, intrii) | H4 | Taiwanese hospital group's in-house nurse rostering system that produces a month's labour-law-compliant roster in about 25 to 30 seconds, now being spun out as the product intrii. | in-house operator in-house unit |
| workforce and rostering | OR · workflow softwarewhyA patented rostering algorithm computing a full month against labour-law and rest rules. hiring note No careers listing on intrii.com. sources intrii.com | detailbuyer head nurses (護理長) | Established | reads EHR, workforce or theatre systems | - | yuanrung.com.tw intrii.com intrii.com |
| C-Live (Hospital Touch) | H4 | AI SaaS (Hospital Touch) that collects staff preferences by smartphone, proposes shift schedules and supports shift exchange in the app, for healthcare and welfare facilities. | vendor incumbent vendor |
| workforce and rostering | OR · workflow softwarewhyCompany describes itself as an AI and mathematical-optimisation firm; Hospital Touch auto-composes shift tables with a patented engine (patent no. 7324260) and collects wishes and swaps by smartphone. hiring note No postings page.The recruiting page (https://www.c-live.jp/aiキャンプ-ブログ) links to external Mynavi listings and a note.com blog. | detailbuyer hospital nursing departments and other shift managers (site has pages for hospital nursing, ER physicians, care facilities, municipalities) hospi.ai says C-Live does all system operation and delivers the finished shift table (BPO), so delivery is software plus services. sources hospi.ai | Customer track record | reads EHR, workforce or theatre systems | - | c-live.jp aspicjapan.org |
| Epigno (EpiTal HR) | H4 | Cloud shift-scheduling tool for hospital nursing departments with AI shift creation, shift-wish collection and automatic generation of the Japanese staffing-standard report (Form 9). | start-up startup |
| workforce and rostering | workflow software | detailbuyer hospital nursing managers (page targets nursing administrators' shift-making time) sources epigno.jp | Customer track record | reads EHR, workforce or theatre systems | 20 company profile page, 20 employees including external board members (https://epigno.jp/company) | epigno.jp epigno.jp aspicjapan.org |
| Patchwork Health | H4H2 | Workforce platform for NHS trusts: job planning, rota with preference-based scheduling, staff bank and agency management, collaborative banks across trusts. | start-up startup |
| workforce and rostering | workflow software · ORwhyRota and job-planning platform with preference-based scheduling, staff bank and agency management. | detaildelivery Platform for NHS trusts; no services-only or hardware model read. | Customer track record | reads EHR, workforce or theatre systems | - press: patchwork.health Series B release says over 100 employees (2022) | patchwork.health patchwork.health |
| Planerio | H4H2 | Automatic shift planning, time tracking and payroll preparation for hospitals, practices and care facilities, with labour-law and collective-agreement checks. | start-up startup |
| workforce and rostering | OR · workflow softwarewhyAutomatic shift planning with labour-law checks, time tracking and payroll preparation. hiring note Careers sit on the parent doctari group portal (karriere.doctarigroup.com, filter company=Planerio GmbH); the list is rendered by script and returned no readable postings, so no count. sources planerio.de planerio.de | detaildelivery Software (shift planning, time tracking, payroll preparation, 100+ interfaces) per planerio.de. | Customer track record | reads EHR, workforce or theatre systems | 50 press: https://www.munich-startup.de/news/planerio-7-fragen | planerio.de munich-startup.de |
| Rotageek | H4 | Rota software with auto-scheduling, demand forecasting and automatic validation against the Junior Doctor 2016 Contract and the Working Time Directive; began as a scheduling tool for A&E, now multi-sector. | acquired acquired startup |
| workforce and rostering | OR · MLwhyAuto-scheduling, demand forecasting and automatic validation against the Junior Doctor 2016 Contract and Working Time Directive. hiring 3 open tech roles: software 1, data eng 1, other 1 careers.rotageek.com. sources rotageek.com rotageek.com | detaildelivery Rota software; no hardware or services-only model read. | Customer track record | reads EHR, workforce or theatre systems | - | rotageek.com breathehr.com |
| DutyMaker (Answer Inc.) | H4 | AI nurse duty-roster generator for hospital wards that applies off and leave requests, night-shift distribution, consecutive-shift limits and skill levels. | start-up startup |
| workforce and rostering | OR · workflow softwarewhyAutomatic roster completion under ward rules (requests, night distribution, consecutive-shift limits). sources dutymaker.com | detailbuyer ward head nurses A pricing page (요금제) is linked; it rendered only a login prompt. | Early | reads EHR, workforce or theatre systems | - | dutymaker.com |
| Medical Strategy & Innovation (ScrubEdge) | H4 | Cloud SaaS for operating-room management linking shifts, staff allocation, experience tracking, skills and training. | start-up startup |
| theatre, outpatient and infusion | workflow softwarewhyCloud SaaS for operating-room shifts, staff allocation and skills; staff submit wishes by smartphone and managers set rules. The press release names no algorithm. hiring note No careers page found. sources prtimes.jp | detailbuyer operating-room managers and staff allocation leads (the release addresses administrators and allocators) sources prtimes.jp | Early | reads EHR, workforce or theatre systems | - | prtimes.jp |
| Nurse is (Nurse Is Inc.) | H4 | Free AI nurse roster tool for Japanese hospitals that proposes several roster drafts balancing staffing, night shifts and staff requests, shows a fairness score and checks the 72-hour night-shift rule for the Form 9 filing. | start-up startup |
| workforce and rostering | workflow softwarewhyRoster drafts generated by AI with a fairness score and regulatory checks for the Form 9 filing; the method is not described. hiring note Recruitment link on the page is for nurse matching, not company jobs. sources shift.nurseis.jp | detailFree at present; paid plans are planned, per the FAQ on shift.nurseis.jp. | Early | reads EHR, workforce or theatre systems | - | shift.nurseis.jp |
| RosterLab | H4 | Rostering software for healthcare clinical teams built on mathematical optimisation (integer programming and metaheuristics) that generates fair, rule-compliant rosters and cuts locum and overtime use. | start-up startup |
| workforce and rostering | OR · workflow softwarewhyRostering engine built on mathematical optimisation (integer programming and metaheuristics) per the open role, inside a rostering application. hiring 1 open tech role: OR / optimisation 1 rosterlab.com (One open role on the careers page: Optimisation Engineer, Mathematical Optimisation & AI (Auckland).) sources rosterlab.com rosterlab.com | detailA pricing page and a "start for free" option are linked from the menu. | Early | reads EHR, workforce or theatre systems | - | rosterlab.com rosterlab.com rosterlab.com |
| rotable | H4 | SaaS that automates rotation planning for physicians in training across specialties and clinics. | start-up startup |
| workforce and rostering | ORwhyhiring note No company site found: rotable.app and rotable.io redirect to a domain-sale lander; no careers page. sources trendingtopics.eu | Early | reads EHR, workforce or theatre systems | - | trendingtopics.eu | |
| ORTEC Workforce | H4 | Healthcare workforce scheduling software (ORTEC Workforce) that balances demand, contracts, skills and labour rules, with self-scheduling; the Dutch hospital VieCuri bought the ORTEC product named Harmony (2012 report). | vendor incumbent vendor |
| workforce and rostering | ORwhyWorkforce scheduling that balances demand, contracts, skills and labour rules; ORTEC states it was founded in 1981 on the conviction that mathematics can solve operational problems (ortec.com/en/about-us). hiring note ORTEC careers (ortec.com/careers) is group-wide across routing, workforce and other lines. | detaildelivery Software product per the industry page; no services model read. | reads EHR, workforce or theatre systems | - | ortec.com ortec.com computable.nl | |
| RLDatix Allocate Optima (e-rostering) | H4 | E-rostering platform for NHS trusts (Allocate Optima, formerly Allocate Software), with SafeCare acuity-based staffing and the Loop staff app. | vendor incumbent vendor |
| workforce and rostering | workflow softwarewhyE-rostering platform with SafeCare acuity-based staffing and the Loop app. hiring note careers.rldatix.com is company-wide (safety, workforce, other products); no postings scoped to Allocate Optima counted. sources rldatix.com rldatix.com | detaildelivery Platform software per rldatix.com; no hardware model read. | reads EHR, workforce or theatre systems | - | rldatix.com rldatix.com | |
| Cetrea (Getinge INSIGHT) | H5 | Clinical logistics and patient flow system, now sold by Getinge as INSIGHT, showing real-time clinical and resource status and bed availability across departments and facilities. | acquired acquired startup |
| patient flow and beds | workflow software · analytics | Established | reads EHR, workforce or theatre systems | - | getinge.com clearwatercf.com tracxn.com | |
| Epic Hospital Patient Flow (Grand Central) | H5H1 | Capacity command centre, length-of-stay and discharge planning, and predictive analytics (predicted stay, expected ED admissions, next-day census) inside the Epic record. | vendor incumbent vendor |
| patient flow and beds | workflow software · analyticswhyCapacity command centre, discharge planning and transfer centre inside the Epic record, with predictive analytics (predicted length of stay, ED admissions, next-day census). sources epic.com | detailsources epic.com | Established | reads EHR, workforce or theatre systems | - | epic.com |
| GE HealthCare Command Center | H5H1 | Command Center software for hospital capacity and patient flow, extended in 2026 by CareIntellect for Operations, which forecasts capacity constraints up to 72 hours ahead. | vendor incumbent vendor |
| patient flow and beds | analytics · ML · workflow softwarewhyCommand Center is real-time operations analytics; CareIntellect for Operations is an AI SaaS forecasting capacity constraints up to 72 hours ahead with an Estimated Day of Discharge model. hiring note careers.gehealthcare.com is a JavaScript job search (Phenom); postings for this product line could not be isolated or counted. sources gehealthcare.com hfmmagazine.com | Established | reads EHR, workforce or theatre systems | - | gehealthcare.com hfmmagazine.com gehealthcare.com | |
| Hospital Authority Hong Kong (Queen Elizabeth Hospital Command Centre) | H5H1 | Public hospital operator's command centre at Queen Elizabeth Hospital tracking bed capacity, occupancy, length of stay, admissions, discharges and transfers from emergency arrival to discharge, with AI discharge prediction in the Authority's smart hospital programme. | in-house operator in-house unit |
| patient flow and beds | analytics · data eng | Established | reads EHR, workforce or theatre systems | - | www3.ha.org.hk healthcareimprovementforum.bmj.com cloudera.com | |
| Infor CloudSuite Healthcare | H5 | Healthcare ERP and human capital suite with workforce scheduling and time tracking and AI-assisted supply chain forecasting for hospitals. | vendor incumbent vendor |
| supply and pharmacy | workflow software · agents · analyticswhyHealthcare ERP and HCM suite with scheduling, supply chain and AI agents and AI-supported analytics, per the healthcare page. sources infor.com | detailsources infor.com | Established | reads EHR, workforce or theatre systems | - | infor.com |
| Johns Hopkins Hospital (Judy Reitz Capacity Command Center) | H5 | Operator-run command centre where referral line, ambulance retrieval, admitting and bed placement staff work from one room using real-time predictive analytics. | in-house operator in-house unit |
| patient flow and beds | analytics · streamingwhyReal-time predictive analytics on about 500 messages per minute from 14 hospital IT systems, shown on a Wall of Analytics. hiring note No postings naming the command centre were found. sources hfmmagazine.com | detailbuyer hospital administration (article quotes the hospital's chief administrative officer); the 2016 article says GE built the system sources hfmmagazine.com | Established | reads EHR, workforce or theatre systems | 24 press: HFM Magazine 2016-11-17, about 24 staff consolidated in the centre (staff, not engineers) | hfmmagazine.com |
| Kaiser Permanente (COVID-19 Forecasting and Surveillance Guidance Group, predictive analytics group) | H5 | Physician and analyst group that produced 480 regularly updated forecasts of hospital census, ICU space, staffing and supply across medical centres during COVID-19. | in-house operator in-house unit |
| patient flow and beds | analytics · simulationwhyForecasting models and simulations of regional case load and its effect on hospital census, ICU space, staffing and supply; later regional leading indicators. hiring note No postings naming the group were found. sources permanente.org | detailbuyer Kaiser Permanente regional and medical centre operations leaders (the forecasts served each region) sources permanente.org | Established | reads EHR, workforce or theatre systems | - | permanente.org |
| NHS Bristol, North Somerset and South Gloucestershire ICB (Modelling and Analytics, IPACS) | H5 | Regional NHS body's in-house modelling team whose open-source simulation (IPACS) models discharge-to-assess capacity across two acute hospital trusts, one community provider and three local authorities and runs demand, capacity and length-of-stay scenarios. | in-house operator in-house unit |
| regional and ambulance | simulationwhyOpen-source discrete-time simulation in R with scenario analysis over demand, capacity and length of stay. sources theorsociety.com | detailbuyer ICB commissioners and D2A service planners | Established | reads EHR, workforce or theatre systems | - | theorsociety.com bnssghealthiertogether.org.uk |
| Omnicell (hospital pharmacy automation and Inventory Optimization Service) | H5 | Medication automation (dispensing cabinets, central pharmacy robots, IV room) plus an Inventory Optimization Service and the OmniSphere data platform. | vendor incumbent vendor |
| supply and pharmacy | OR · workflow softwarewhyInventory Optimization Service and OmniSphere platform on top of dispensing cabinets and pharmacy robots. hiring note careers.omnicell.com is company-wide (cabinets, robots, services); no postings scoped to the inventory optimisation unit counted. sources omnicell.com omnicell.com | detaildelivery Dispensing cabinets, central pharmacy robots and IV room automation sold alongside the software and services (omnicell.com). revenue Annual total revenue USD 1.185B as of 2025-12-31, stated at https://www.omnicell.com/about-us | Established | reads EHR, workforce or theatre systems | - | omnicell.com |
| Philips Patient Flow Capacity Suite | H5 | Cloud patient logistics software (HealthSuite) that combines clinical and operational data across a health system, its affiliates and post-acute settings to show capacity, predict demand and spot patient flow bottlenecks. | vendor incumbent vendor |
| patient flow and beds | analytics · ML · workflow softwarewhyCloud HealthSuite application combining clinical and operational data, with visualisation and machine learning supported analytics for demand and bottlenecks. sources usa.philips.com | detailsources usa.philips.com | Established | reads EHR, workforce or theatre systems | - | usa.philips.com mobihealthnews.com ifdesign.com |
| DaphOS | H5H2H4 | AI forecasting platform for hospital and care managers: emergency-department arrival and ward occupancy forecasts, staffing shortfall detection, absence management, automatic rosters and supply needs. | start-up startup |
| several | ML · analyticswhyForecast models trained on historical data and fine-tuned per site, combined with weather and epidemiological data, feeding manager dashboards and suggestions. sources trendingtopics.eu daphos.ai | detailbuyer hospital and care managers (Fuehrungskraefte); the site also notes eligibility for the German KHVVG transformation fund | Early | reads EHR, workforce or theatre systems | - | trendingtopics.eu daphos.ai daphos.ai |
| Solamed (OpePro) | H5H3 | AI SaaS for acute-care hospitals that unifies surgery schedules across departments, shows operating-room utilisation and supports reallocating unused slots. | start-up startup |
| theatre, outpatient and infusion | workflow software · ORwhyUnifies surgery schedules across departments, shows operating-room utilisation and supports reallocating unused slots (ascii.jp article). hiring note No company site reachable (solamed.co.jp and solamed.jp returned HTTP 000); no careers page found. sources ascii.jp | Early | reads EHR, workforce or theatre systems | - | ascii.jp | |
| Faculty (Frontier, system coordination centre use) | H5 | Frontier is marketed for NHS integrated care board system coordination centres: capacity visibility across ambulance, 111, acute, social care, virtual ward, GP and mental health data, forecasting and scenario simulation. | start-up startup |
| regional and ambulance | analytics · ML · data engwhyFrontier pulls ambulance, 111, acute, social care, virtual ward, GP and mental health data into one capacity view with AI models forecasting pressure points; simulation is described only in conditional terms. hiring 55 open tech roles: ML 27, data science 12, CV 1, software 15, other 4 jobs.ashbyhq.com. None of the titles names NHS or hospital operations. The Faculty site links the vacancies through faculty.ai/en-gb/job-board.) sources faculty.ai | detailbuyer NHS integrated care boards running system coordination centres Faculty's navigation lists AI Services, Frontier and a Fellowship; the delivery label is from that and the NHS positioning in the blog, not from a Frontier pricing page. sources faculty.ai faculty.ai | reads EHR, workforce or theatre systems | - | faculty.ai faculty.ai |
Group: startup, incumbent vendor (over 200 people, listed, or a product line of such a company), operator in-house unit (an operator's own team building decision software for itself) or acquired startup; the parent is named where there is one. Problems: the named decision tasks the product addresses, main one in bold, hover for the definition. Technical core: the one to three competencies the product stands on, ranked, from open roles, engineering posts and product pages; open "why" for the reasoning, the role counts and the sources. Money signals: deal-size band and what it is per, then delivery model and pricing as stated; open "detail" for revenue, funding, buyer, named competitors and sources. Customers alone do not count. Insurers count only when they operate hospitals. None were gathered for this sector. Size: Established: more than 200 employees, or publicly listed, or a product line of such a company. Customer track record: 200 or fewer employees, and the sources name 5 or more customer companies or state a count of 50 or more customer companies or sites. Early: 200 or fewer employees, with 4 or fewer customers named and no stated count of 50 or more: pilots, a few logos, or none public yet. A dash means nothing was found in a public source. Amounts in other currencies are converted to US dollars only to place them in a band, at approximately 1 EUR = 1.15 USD, 1 GBP = 1.33 USD, 1 CHF = 1.25 USD, 1 AUD = 0.65 USD, 1 CAD = 0.72 USD, 1 CNY = 0.14 USD, 1 TWD = 0.032 USD, 1 JPY = 0.0066 USD, 1 KRW = 0.00071 USD, 1 INR = 0.0113 USD, 1 CZK = 0.047 USD, 1 NOK = 0.098 USD, 1 SEK = 0.105 USD; bands are rough.
How regional capacity crosses hospital walls, and who carries it
Between organisations is a flag on a company, not a layer: 1 of 52 companies carry it (NHS Bristol, North Somerset and South Gloucestershire ICB on H5). A layer needs at least five companies or three startups; this has one in-house team and no product, so it stays a flag and a section.
1. The signal
In this sector the signal that crosses organisation walls is regional spare capacity: which hospital has a bed, which emergency department (ED) is full, which provider can take a patient who has waited too long. Three cases.
Patient distribution in the Netherlands (COVID, 2020 onward). The Landelijk Coordinatiecentrum Patienten Spreiding (LCPS) is the Dutch centre that relocates patients between regions to balance bed pressure [1][2]. Since 1 July 2023 its tasks sit inside the Landelijk Netwerk Acute Zorg (LNAZ). It has an information part (data management and analysis) and a coordination part; if a capacity problem needs spreading across regions, the coordination and escalation plan of the LNAZ allows it to distribute care demands over regions, run from the Patient Evacuation Coordination Centre (PECC), which also handles medical evacuations, for example from Ukraine [2]. LNAZ itself lists ten regional acute care networks [3]. The model behind the daily decisions: in autumn 2020 Bekker and Koole (VU Amsterdam), with uit het Broek (Groningen), built a statistical arrivals model plus a queueing-theory model that turns arrivals into occupancy for COVID patients, for LCPS. Koole says all hospitals deliver patient numbers and inflow every day, the model forecasts one week ahead new patients on the ICU and in the ward, and "on that basis the hospitals transfer patients every day"; LCPS built it into a fully automated data flow that is "still used" [4]. The paper describes forecasts at national level and regional level through LCPS, and no scenario analysis [5]. The VU page also records a design rule: a hospital that can cope with COVID inflow again should not stop transferring at once, so that postponed care (knees, hearts, oncology) recovers evenly across regions [4].
Ambulance diversion dashboard, Netherlands. A nationwide survey of Dutch EDs between January and October 2023 found 62 of 82 EDs (75.6%) using a real-time dashboard that shows regional ambulance diversions as a traffic light; 56 of those answered (90.3%). 69.7% of ED managers reported crowding more than three times a week; of dashboard users, 52.8% said it only occasionally eases inflow, because few patients are affected by the red light and because crowding is regional. The authors conclude crowding is "largely unchanged" since the dashboard was introduced (published 21 January 2025) [6]. The dashboard shows the present state of neighbours. It does not forecast and does not test what happens if one hospital changes its rule.
NHS England mutual aid for waiting lists, 16 August 2022. Letter PR1830 asks providers to validate outpatient lists and identify patients who could be seen by another provider, with deadlines: over 78 weeks and no first appointment booked within four weeks, by 16 September 2022; over 52 weeks, by 25 November 2022; over 60 weeks with a diagnostic done and at risk of breaching 78 weeks at 31 March 2023, by 3 February 2023. Mutual aid runs by "the escalation process of system to region, and region to national level". Providers must agree funding and reimbursement, patient movement onto the other provider's patient tracking list, transfer of requests and results, and next steps after a decision to admit. If systems cannot place a patient, regions are expected to arrange transfer within or to another region. The letter says providers should use "technology and AI" where possible, and writes the outcome of validation into the waiting list minimum data set from 1 September 2022 [7]. No simulation is named.
2. Who carries it today
- A national coordination body with a data feed and a model: LCPS, now in LNAZ [1][2][3].
- A shared dashboard: the Dutch ambulance diversion traffic light [6].
- A letter with deadlines and bilateral agreements between providers, escalated to region and nation: NHS England PR1830 [7].
- A purchasing contract: HealthTrust SIMS.
- NHS system coordination centres: NHS England issued a specification in August 2023 asking for three outcomes: enhance operational visibility of capacity and flow, enable real-time coordination, improve clinical outcomes.I have this from a vendor blog quoting it (dated 2023-10-20) [8].
The Dutch survey says hospitals' own internal agreements drive the response to the orange light [6].
3. Who sells there
The vendor article (2023-10-20) says AI simulations "can show the consequences of your teams' decisions before they are made" and running real-time scenarios will help staff "understand how adjustments in specific areas will affect the broader healthcare system", for ICB coordination centres [8]. That is a statement of what such a tool can do. So BNSSG as a user is not shown.
Near misses.
- LCPS forecast model: crosses organisations and drives transfers every day, but it is a forecast, not scenarios over the parties [4][5].
- Neither is an operating cross-organisation unit.
5. The one company carrying the flag
Its IPACS simulation, written in R with the universities of Bath and Exeter, models the discharge-to-assess pathways across two acute trusts, one community services provider and three local authorities, and runs twelve demand, capacity and length-of-stay scenarios for each locality and pathway. The paper says responsibility for the pathway is "not the remit of any single organisation", that the model is now used for intermediate care commissioning, and that its estimates supported a business case of about GBP 13m. It fits the reading in section 4: a public coordination body with an in-house model built with a university, not a product bought from a vendor.
Sources
[1] https://lcps.nu/ , LCPS home: task transfer to LNAZ from 1 July 2023, information and coordination parts, PECC.
[2] https://lcps.nu/ (same page), cited for the escalation plan and PECC wording.
[3] https://www.lnaz.nl , LNAZ: ten regional networks listed; no forecasts, scenarios or tools mentioned.
[4] https://www.vu.nl/nl/onderzoek/vu-wiskundigen-maken-model-voor-spreiding-coronapatienten , VU news item: one-week forecast, daily transfers, automated data flow, postponed-care rule.
[5] https://ar5iv.labs.arxiv.org/html/2102.11021 , Bekker, uit het Broek, Koole, modelling COVID-19 admissions and occupancy; fetched through a summarising tool, which reported national and regional forecasts, no scenarios, no operational figures.
[6] https://researchinformation.umcutrecht.nl/en/publications/emergency-department-crowding-in-the-netherlands-evaluation-of-a-/ , survey figures (62 of 82, 90.3%, 69.7%, 52.8%), published 21 January 2025.
[7] https://www.england.nhs.uk/coronavirus/wp-content/uploads/sites/52/2022/02/B1830-letter-delivering-the-validation-of-non-admitted-outpatient-waiting-lists-patients-suitable-for-mutual-a.pdf , PR1830, read as full text.
[8] https://faculty.ai/blog/why-an-ai-powered-scc-is-key , vendor article, 2023-10-20: SCC specification outcomes and simulation wording.
Also opened today without usable result: https://www.regioner.dk/ (home page only), https://bnssg.icb.nhs.uk/ (home page only), https://www.england.nhs.uk/ourwork/clinical-policy/ipacs/ (empty), https://www.bsw.icb.nhs.uk/.
method
Companies were found by web search, and each row rests on the sources linked in its last column. The five layers were derived from what the companies sell, each company's services listed from its own pages, and are ordered by how far ahead each decision looks. A dash means nothing was found in a public source; it does not mean zero.