Rate a Device by What the Hospital Can Do With Its Evidence
Connectivity Maturity Model
Companion to Medical Device Connectivity · Updated 2026-09-29
Use this to say in one line what the institution can do with a device’s evidence today, and what the next level would take. Buyers use it to set a class minimum for each device class in a request for proposal (RFP); builders use it to plan a roadmap. “Many Devices Keep a Record of Their Work That the Hospital Can’t Use” works it on the sterilizer. The levels apply to every institutional device class, so the examples below come from across the hospital: the laboratory, the operating room, dialysis, pharmacy, radiation oncology, and sterile processing.
A level describes what the institution can do with the device’s evidence, not what radio the device carries. Levels are cumulative and describe the device as a whole: each level assumes the ones below it. Rating each kind of evidence separately, as “Many Devices Keep a Record of Their Work That the Hospital Can’t Use” does for the sterilizer, is a diagnostic that shows where the gaps are. The device’s level is the lowest level reached by the evidence the institution needs from it. Level 1 is a real level: a device may be offline and still not data-dead, if its file export is complete, is documented well enough that someone other than the manufacturer can parse it, and identifies the unit that produced each record.
The Connectivity Maturity Model: six levels, each defined by what the institution can do with the device’s evidence.
| Level | Name | What the institution can do with the evidence | Typical evidence path | Example |
|---|---|---|---|---|
| 0 | Silent | Nothing machine-readable. The institution knows about the device’s work only from its screen, a printout, staff memory, or what staff copy into another system by hand. | Display, printed tape or report, handwritten log | A medication refrigerator whose temperature record is a paper log that staff initial each shift |
| 1 | Exportable | Retrieve complete machine-readable records after the fact, by a person, for audit, investigation, or analysis | Universal Serial Bus (USB) drive, file download from a local interface, service port | A washer-disinfector that writes each cycle’s program, temperatures, operator, and outcome to a file that staff copy off the device on a schedule |
| 2 | Network Observable | See device status and records over the network as they happen, without walking to the device. No workflow acts on them automatically. | Network interface to a local server, gateway, or viewer; status and event feeds | An automated endoscope reprocessor that sends every cycle record to a server where infection prevention can look up any endoscope’s history |
| 3 | Workflow Integrated | Move device events and state into the workflows that use them, with the right patient, specimen, or asset context; where the design allows, send orders or settings back | Documented or standards-based interfaces, through integration middleware, to the electronic health record (EHR), a tracking system, or an alarm system | A chemistry analyzer that receives its orders from, and returns results to, the laboratory information system with the specimen identity attached; an anesthesia workstation whose delivered agent and ventilation values reach the anesthesia record |
| 4 | Fleet Managed | Inventory, configure, update, monitor, and retire every unit as a fleet, with evidence of each unit’s version, configuration, and state | Device-management interface, inventory feed to the computerized maintenance management system (CMMS), configuration and update service | A fleet of hemodialysis machines whose software versions, configurations, disinfection cycles, and locations the institution can see and change across every unit |
| 5 | AI-Ready and Evidence-Producing | Rely on a device that meets the twelve AI-ready capabilities that “An AI-Ready Device Is Mostly a Well-Engineered, Interoperable Device” defines (the AI-Readiness Procurement Checklist), and whose evidence for quality, safety, security, and compliance work falls out of normal operation, with enough provenance that an authorized reviewer, human or machine, can reconstruct what happened | Everything above, plus provenance on every record, logged automated actions, and assurance data | A linear accelerator whose treatment records, interlock events, software version, and calibration status arrive with provenance, so an artificial intelligence (AI) system can observe and recommend while control stays local and clinical |
Set a class minimum for each device class, based on the work the class does. A low-volume device doing simple work, such as a patient lift, may be right at Level 1. An Institutional Process Device doing repeated, audited work, such as a washer-disinfector or a blood bank refrigerator, usually isn’t. A device below its class minimum gets either a dated roadmap commitment in the contract or an exception record under the No Connectivity, No Purchase Checklist. When no bidder meets the class minimum and the institution buys anyway, it needs both: one exception record for the class, and a roadmap commitment with a remedy if the date slips. Use the roadmap clause in the Connected Device RFP Language.