cost_of_care_exposure
Size an exposure over a declared population: what a condition accounts for across a book, over a declared period, on figures the caller declares. Route here when the question is how big something is on a plan's book, and read the last paragraph before quoting anything downstream of the answer. Declare EITHER an `affected_members` count OR a `population` with a `prevalence`, never both routes at once, together with `monthly_cost_per_affected_member`, `period_months` and `currency`. The exposure is their product and every input is echoed with its denominator named. Figures and shape travel differently. A figure may arrive as a direct argument or from a `declared_inputs` handle passed as the `declared_inputs` argument; shape, meaning the period and the currency, arrives only directly, because those declare what the output means rather than carrying a figure of their own. There is no basis argument to spell: a direct figure argument IS the caller's own declaration on this call and enters as an assumption by construction, while a figure with a source behind it travels through `declared_inputs`, where the basis and the source are declared once and read back here. When the handle and a direct argument fill the same role the call is refused as ambiguous rather than one winning silently. The answer's basis is the weakest basis among the figures that built it. The money input is a cost per AFFECTED member per month, and a plan-book PMPM is not one. A PMPM already spans every member, so multiplying one by a prevalence takes a slice of a slice, and a `pmpm` key is refused by name whether it arrives as an argument or inside a handle row. AN EXPOSURE IS NOT A SAVING. It says what a condition costs a book as declared; what an intervention would change about that is a savings figure, and a plan's Cost of Care standard scores savings on paid amounts, which no compartment of this world produces. The payload carries that caveat beside the figure so the number cannot travel without it, and carries no field a savings figure could be written into at any depth. A caller who asks for the saving anyway is met at that wall before any missing-input check, because a missing-input refusal there would read as an invitation to supply the inputs and be served the saving.
Use this form for new arguments. Follow concrete links for known values. Fetch-only readers may be unable to submit new URLs.
Exact input contract
{
"additionalProperties": false,
"properties": {
"affected_members": {
"description": "How many of the book have the condition, as a count. The other dimensional route; declaring it alongside a population or a prevalence is refused.",
"type": "number"
},
"currency": {
"description": "What the money inputs are denominated in, echoed onto the answer. Shape rather than figure, and this world converts nothing.",
"type": "string"
},
"declared_inputs": {
"description": "A handle returned by the declared_inputs tool, whose rows supply any of the figure roles above on the basis and source declared there.",
"type": "object"
},
"monthly_cost_per_affected_member": {
"description": "Cost per AFFECTED member per month, in the declared currency. Not a plan-book PMPM, which spans every member and would take a slice of a slice.",
"type": "number"
},
"period_months": {
"description": "How many months the exposure covers. Shape rather than figure, so it arrives directly and never through a handle.",
"type": "number"
},
"population": {
"description": "How many people the book covers. Declared with a prevalence, and the two together are the alternative to an affected-member count.",
"type": "number"
},
"prevalence": {
"description": "The share of that population with the condition, between zero and one.",
"type": "number"
}
},
"required": [
"period_months",
"currency"
],
"type": "object"
}