CIFoundry

Evaluation packet

TREWS (Targeted Real-time Early Warning System)

Bayesian Health (originally developed at Johns Hopkins University by Suchi Saria and colleagues) · Hospital Medicine & Critical Care · Sepsis & deterioration early warning

Prepared November 22, 2026 from the public listing on CIFoundry. It shows what the public record says and what to ask the vendor. It is not a recommendation or an endorsement.

1. What the public record shows

Regulatory status
Not on record
Published evidence
Prospective / randomized study · Linked study
Named health-system users
Integrations
Not on record
Data inputs
Not on record
HIPAA BAA
Not on record
SOC 2 / HITRUST
Not on record
Deployment model
Not on record

Integrations, data inputs, BAA, SOC 2, HITRUST and deployment model are reported by the vendor and not checked by us. Confirm them in writing.

From public directories

Third-party directories checked by our team. These are not the vendor's own pages and not an endorsement.

Listed in Epic Showroom
Connection Hub: Bayesian Health (Frontend) by Bayesian Healthsource, checked Nov 21, 2026Epic’s public app directory. Listed developers pay Epic for services; Epic does not endorse, certify or validate listed products.

2. Evaluation checklist for your role

Each question shows what the public record says: on record (with its source), reported by the maker, or not on record. It is a list of things to check, not a score.

For every reviewer

Physician

Your ticks and notes stay in this browser. Nothing is sent to us.

3. Gaps to raise with the vendor

Based on what is missing or worth confirming in this tool’s public record.

  1. No regulatory status is listed. Ask whether the tool is regulated by the FDA as a medical device and, if so, for the clearance number and the exact cleared intended use.
  2. Will the vendor sign a HIPAA Business Associate Agreement?
  3. Does the vendor have a current SOC 2 Type II report or HITRUST certification you can review?
  4. Where is data stored and processed (cloud region, on-premises), and which subcontractors, including any large-language-model provider, touch patient data?
  5. It is listed in Epic Showroom, which says nothing about how it is configured. Ask which Epic integration it uses at live sites and where results appear for the clinician.
  6. Is your data used to train the vendor’s models, and can it be deleted when the contract ends?

4. Questions for this kind of tool

From our Predictive AI for deterioration and sepsis guide.

Questions for every predictive tool

Applies to deterioration, sepsis and condition-specific models.

Questions to ask

  • At the threshold you recommend, what are the positive predictive value and the number of alerts per 100 patient-days?
  • How was the outcome defined: by clinical criteria or by billing codes? Billing-code outcomes can inflate apparent accuracy.
  • Is the model calibrated, not just good at ranking (AUC)? Ask for a calibration plot.
  • Was it validated externally, at hospitals other than the developer's, and on outcomes measured after the alert time?
  • If the model is supplied in a certified EHR, ask for its ONC HTI-1 source attributes.
  • How do we retune the threshold, and who monitors drift?

Test it locally

  • Run the model silently for 2 to 3 months before anyone sees alerts.
  • At several thresholds, measure alerts per day per unit, PPV, sensitivity and median lead time.
  • Choose the threshold together with nursing and physician leaders, and write down the response expected for each alert.

Red flags

  • Outcomes defined by billing codes
  • No calibration data
  • Alerts with no recommended action
  • No estimate of alert volume per unit

General deterioration warnings

Use this section if the tool gives a general early-warning score for ward or ICU patients.

Questions to ask

  • Does it beat the score you already use (for example a standard early-warning score) on your data?
  • Who receives the alert: bedside nurse, rapid-response team or a remote monitoring center?
  • How is alert burden balanced across shifts and units?

Sepsis prediction

Use this section if the tool predicts or helps diagnose sepsis.

Questions to ask

  • Is it FDA-authorized? If so, read the labeling: authorized sepsis tools are cleared as aids, not as the sole basis for diagnosis.
  • Which sepsis definition was used for training and testing?
  • How many alerts fire after treatment has already started? Late alerts add workload without benefit.

Single-condition predictions

Use this section if the tool predicts one specific event, such as acute kidney injury, falls or readmission.

Questions to ask

  • Is there an action that changes the outcome once a patient is flagged, and do you have the staff to take it?
  • How does it perform in subgroups that matter for that condition (for example baseline kidney disease)?

5. Questions for every tool

The 12 questions that apply to every tool

Fit and evidence

  • Intended use. Does the stated use match the patients, setting, user and decision you plan to use it for?
  • Evidence level. From weakest to strongest: vendor-internal test, retrospective external validation, prospective "silent" run, prospective live study, randomized trial, proven change in outcomes. Ask which level the claims rest on and look for peer-reviewed external validation.
  • Population and equipment. Ask for results by age, sex, race and ethnicity, language and payer, and, where relevant, skin type and body habitus. Ask for results on your scanner or device vendors, EHR and protocols.
  • Local validation. Before go-live, run the tool on your own recent cases and compare it to ground truth. Agree in the contract on what counts as passing.

Workflow and people

  • Workflow cost. Count clicks, alerts per clinician per shift, minutes added per case, and who acts on the output. A strong tool that nobody acts on has no value.
  • Integration. Native EHR or PACS integration or a separate window? Single sign-on? Where do results land?
  • Human oversight. Who has the final say, how are disagreements with the tool handled, and how do you guard against over-reliance? Train users on known failure modes, not just features. Explanations matter most for risk scores and decision support.

After go-live

  • Monitoring. Who checks for drift, how often, and what triggers a pause? Re-check after vendor updates, scanner or EHR changes, and shifts in your patient population. Agree in advance on the results that would retire the tool.
  • Model updates. Version control, advance notice of changes, and the right to stay on a version you validated.

Data, liability and money

  • Data rights and security. BAA, where data is stored and processed, encryption, retention, SOC 2 or HITRUST, whether your data trains the vendor's models, the list of subcontractors that touch patient data (including any large-language-model provider), and whether data can be deleted.
  • Liability and patient notice. Who is liable when AI-assisted care goes wrong is still unsettled, and responsibility currently stays with the clinicians and organizations using the tool. Ask your malpractice carrier, and cover indemnification, audit rights and incident reporting in the contract. Joint Commission and CHAI guidance says patients should be told when AI directly affects their care, with consent where relevant.
  • Money and exit. Pricing model, any real payment pathway, the return you will measure, termination for performance failure, data return, and no auto-renewal without a performance review.

6. Notes and decision

Problem we want to solve and how we will measure it

Vendor answers and follow-ups

Local test plan and what counts as passing

Decision, owner and next step

CIFoundry is an independent, vendor-neutral directory. Listing is free, and no one can pay to change a listing or this packet. Check every fact with the vendor and the primary source before relying on it.