HealthTeky · value model for hospital boards

What would this be worth in your hospital?

Put your own numbers in. Every multiplier below is an assumption you can change, and the model shows its arithmetic. Nothing here is a claim about results HealthTeky has already produced.

Read this before you quote a number. This is a planning model, not evidence. It multiplies figures you supply by improvement rates you choose. Use it to decide what is worth measuring in a pilot — then replace every assumption with what the pilot actually shows.

Your hospital

Defaults describe a 200-bed multi-specialty hospital. Change anything.

Size and activity
Money
Improvement assumed — the part you are really deciding
What it costs

Annual value, year one

—
after platform and setup cost
Gross value
—
Cost
—
Payback
—
Clinician hours returned
—
Annual value by lever
Annual value by lever, at the assumptions set on the left.

The arithmetic, in the open

Each line is a multiplication you can check. Where a figure is contested, the model takes the conservative side and says so.

LeverHow it is calculatedValue a yearWhere the improvement comes from

What a 90-day pilot would actually measure

A model is a hypothesis. These are the five numbers worth instrumenting from day one, each with a baseline taken before anything is switched on.

Documentation minutes per consult

Timed, on a sample of consultations, before and after the scribe. Not self-reported.

Your assumption: 4 min saved

Denial and short-payment rate

Denials as a share of claims submitted, by payer, with the reason codes.

Your assumption: 18% fewer

Length of stay, case-mix adjusted

Adjusted, or the number moves for reasons that have nothing to do with software.

Your assumption: 2.5% shorter

Time to communicate a critical result

Minutes from a critical finding to a named clinician acknowledging it. A safety measure that is also an audit finding.

Baseline today: often unmeasured

Override rate on AI suggestions

How often clinicians reject what the system prepared. Rising overrides mean the model needs re-validation — the platform reports this on itself.

Target: stable and explainable