A tauSpan platform
About tauSpan
Physician-designed. Engineered to be checked.
tauSpan Technologies builds High Definition Care: clinical AI for hospitals that leaders can hold to account. We ship what works, we describe it plainly, and every answer the platform produces points back to a document or a note you already own.
What we build
Clinical AI you can argue with.
High Definition Care was designed by a physician, and it shows in what the platform refuses to do. It doesn't return a score without the evidence behind it. It doesn't answer a policy question from general knowledge. When the record genuinely doesn't support a call, it says so and sends the case to a person rather than picking the likelier option.
That posture costs us some impressive-sounding claims. A system that says “ambiguous, a human should look” reports fewer confident detections than one that guesses. We think that trade is obviously correct in a hospital, and that the people evaluating this software think so too.
It works less like a black box and more like a good resident: it tells you what it found, where it found it, and when it isn't sure.
How we work
How a pilot actually goes.
-
Step 01
Scope
One module, and one question you actually have. Not a platform evaluation — a specific thing you want to know.
-
Step 02
Load
Ten policies, or one month of inpatient notes. Enough of your own material that the results mean something.
-
Step 03
Pilot
Your people, your material, two to four weeks. We stay close enough to fix things and far enough that you form your own view.
-
Step 04
Decide
You've seen it work on your own data, or you haven't. Either answer is a useful outcome for both of us.
Data handling
How your data is handled.
The platform has passed an external HIPAA compliance risk assessment. Below are the controls behind that, stated plainly — and we're happy to walk your security team through the architecture in detail.
-
Tenant-scoped isolation. Your data is scoped to your tenant and enforced at the data-access layer, so isolation doesn't depend on each query remembering to ask for it.
-
Role-based access per module. Viewer, user and admin roles with granular permissions, and module administration you can delegate to the people who own the work.
-
A managed access lifecycle. Invitation, activation, password reset and session authentication, so access is granted and removed deliberately.
-
Secrets and transport. Credentials held in AWS Secrets Manager rather than configuration files, and HTTPS end to end.
-
Audit logging. Analysis runs and review actions are recorded against the user who performed them.
-
Ingestion that fails safely. Each import is validated in full before anything is written, is idempotent on re-send, and alerts on failure rather than failing quietly.
Straight answers
Questions your IT team will ask.
Do you integrate with our EHR?
Not through an interface, and the two modules differ. HDC:Policy touches your EHR not at all — policies are uploaded through the browser. HDC:Quality needs exactly one thing: a scheduled report against your clinical data warehouse — Clarity, if you're an Epic shop — returning six columns, one row per clinical note, written daily to a storage location scoped to your tenant.
That is a reporting job your analysts write with the tools they already use, not an interface engine, an HL7 feed or a field-mapping exercise. We hand your team the column specification and work through the first export with them. It is usually the difference between starting this quarter and starting next year.
How does analysis get triggered?
A person starts a run. Imported encounters wait until an analyst chooses to analyse them. Runs that start themselves as each export lands are on the roadmap, not in the product.
Do you support single sign-on?
Not today. Access uses session-based authentication with per-module role-based access control, an invitation and activation flow, and password reset. If SSO is a requirement for you, tell us — it's a reasonable thing to need and we'd rather hear it early.
Are you HIPAA compliant?
We have completed an external HIPAA compliance risk assessment and passed it. Alongside that, the controls listed above are what the platform does today, and we're happy to walk your security team through the architecture in detail rather than asking them to take a summary on trust. If a signed agreement is needed before any data moves, let's talk about it up front.
Do you have SOC 2 or HITRUST?
Not yet — we haven't sought either attestation. They are separate exercises from the HIPAA risk assessment, and we'd rather say so plainly than let one credential imply another. If a formal attestation is a procurement requirement for you, tell us early and we'll be straight about where that sits.
What file formats can we load into HDC:Policy?
PDF, Word and plain text. Scanned documents that contain no extractable text are not supported, since there is no OCR step — if much of your library is scanned paper, tell us before a pilot so we can scope it honestly.
Can we see it without sending real patient data?
Yes. Both modules can be demonstrated end to end without exposing any PHI. HDC:Policy works on policy documents, which carry no patient data to begin with. HDC:Quality includes a demo mode that de-identifies clinical text on the fly, so a case can be walked through in front of a room — or a committee — without the patient being visible. Nothing has to leave your building before you've decided the product is worth a pilot.
Contact
Talk to us.
We'd rather show you than tell you.
Useful to include: which module you're interested in, roughly how big your hospital is, and what you'd want to test first. That's enough for us to tell you whether a pilot makes sense before either of us spends much time on it.
High Definition Care
Two modules live. One platform underneath.
HDC:Quality reviews your inpatient notes with cited evidence. HDC:Policy answers questions from your approved policies with citations back to the source.
See the platform