Files
stack/tests/pay/test_ceiling.py
kert 37c732dfd0
Some checks failed
CI / lint (push) Failing after 44s
CI / notebooks-smoke (push) Successful in 1m48s
Deploy / changes (push) Successful in 18s
Infra CI / notebooks (push) Successful in 55s
Infra CI / zotero (push) Successful in 23s
CI / test (push) Successful in 4m23s
Infra CI / docs (push) Successful in 22s
Infra CI / api (push) Successful in 1m10s
Infra CI / bcda (push) Successful in 1m46s
Infra CI / llm (push) Successful in 52s
Infra CI / mc (push) Successful in 20s
Deploy / notebooks (push) Has been skipped
Deploy / zotero (push) Has been skipped
Deploy / docs (push) Has been skipped
Deploy / api (push) Successful in 2m0s
Lean Pay / pay (push) Successful in 3m26s
Deploy / bcda (push) Successful in 1m59s
Deploy / mc (push) Has been skipped
Deploy / llm (push) Successful in 1m19s
Deploy / report (push) Successful in 15s
Comment Pull / comment-pull (push) Failing after 29s
Notebooks Integration / notebooks-integration (push) Failing after 12m24s
Renovate / renovate (push) Successful in 2m31s
LLM Golden / llm-golden (push) Successful in 29m52s
Zotero Sync / zotero-sync (push) Successful in 3m3s
Package Supply Chain / pkg-supply-chain (push) Successful in 1m35s
feat(ipps): price the claim from the published factors
Operating payment is the federal rate with COLA on the nonlabor portion,
times (1 + operating IME + operating DSH), plus the uncompensated-care
add-on. The impact-file operating DSH is already the empirically justified
factor and is not multiplied by 0.25. Capital payment is the Table 1D
national rate times the hospital GAF, times (1 + capital IME + capital DSH).
The impact file is the final-rule model, not the quarterly Provider Specific
File.

An outlier is the 42 CFR 412.84 marginal factor on the excess of charge-based
cost over that claim payment and the wage- or GAF-adjusted fixed-loss share.
No charges, or no published fixed-loss threshold, produces no outlier. The
national wage-index-1 federal-rate identity is unchanged.
2026-11-18 12:51:45 -05:00

218 lines
5.9 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""Python calculations meet the Lean ceilings in ``lean/Pay``.
The committed ``cases.json`` is the output of ``lake exe pay-cases``.
Each case calls the production function on the same inputs.
"""
from __future__ import annotations
import json
import os
import shutil
import subprocess
from fractions import Fraction
from pathlib import Path
import polars as pl
import pytest
from opps.calcs.payment import payment as opps_payment
from pfs.calcs.payment import (
equipment_cost_per_use,
limiting_charge,
mppr_reduction,
)
from pfs.calcs.payment import (
payment as pfs_payment,
)
from qpp import QPP
ROOT = Path(__file__).resolve().parents[2]
PAY = ROOT / "lean" / "Pay"
CASES = json.loads((PAY / "cases.json").read_text())
BY_ID = {row["id"]: Fraction(row["num"], row["den"]) for row in CASES}
def _rat(value: float) -> Fraction:
return Fraction(value).limit_denominator(10**9)
def test_case_ids() -> None:
assert set(BY_ID) == {
"pfs-unit-gpci",
"limiting-100",
"opps-wage",
"opps-missing-wage",
"asp-100",
"reach-earn-back-85",
"qpp-2024",
"equipment-minute",
"mppr-second",
"ipps-operating-cola-one",
"ipps-operating-claim",
"ipps-capital-gaf-one",
"ipps-capital-claim",
"ipps-outlier",
}
def test_pfs_unit_gpci() -> None:
rvu = pl.DataFrame(
{
"hcpcs": ["X"],
"work_rvu": [1.0],
"non_fac_pe_rvu": [2.0],
"mp_rvu": [3.0],
"mac": ["1"],
"locality": ["00"],
}
)
gpci = pl.DataFrame(
{
"mac": ["1"],
"locality": ["00"],
"work_gpci": [1.0],
"pe_gpci": [1.0],
"mp_gpci": [1.0],
}
)
got = pfs_payment(rvu, gpci, cf=10.0)["payment_amount"][0]
assert _rat(got) == BY_ID["pfs-unit-gpci"]
def test_limiting_charge() -> None:
assert _rat(limiting_charge(100.0)) == BY_ID["limiting-100"]
def test_opps_wage() -> None:
apc = pl.DataFrame(
{
"hcpcs": ["A", "B"],
"apc": ["1", "2"],
"relative_weight": [1.0, 1.0],
"payment_rate": [100.0, 80.0],
"status_indicator": ["J1", "J1"],
"cbsa": ["10180", "99999"],
}
)
wage = pl.DataFrame({"cbsa": ["10180"], "wage_index": [1.05]})
result = opps_payment(apc, wage, labor_share=0.60)
adjusted = {
row["hcpcs"]: row["adjusted_payment"] for row in result.iter_rows(named=True)
}
assert _rat(adjusted["A"]) == BY_ID["opps-wage"]
assert _rat(adjusted["B"]) == BY_ID["opps-missing-wage"]
def test_asp_limit_inside_skin_sub() -> None:
"""The express skin-sub step sets the limit to ASP × 1.06, rounded to cents."""
from opps.express.payment import skin_sub_impact
claims = pl.DataFrame({"hcpcs": ["Q"], "payment_rate": [0.0], "units": [1.0]})
asp = pl.DataFrame({"hcpcs_code": ["Q"], "asp_per_unit": [100.0]})
result = skin_sub_impact(claims, asp)
assert _rat(result["old_payment"][0]) == BY_ID["asp-100"]
def test_reach_earn_back_matches_the_published_example() -> None:
"""85% quality × the 2% withhold = 1.7% of benchmark."""
assert BY_ID["reach-earn-back-85"] == Fraction(85, 100) * Fraction(2, 100)
def test_qpp_payment_year() -> None:
assert QPP[2024].payment_year == BY_ID["qpp-2024"]
def test_equipment_cost_per_minute() -> None:
equipment = pl.DataFrame(
{
"hcpcs": ["X"],
"unit_price": [1000.0],
"useful_life": [5.0],
"minutes_per_year": [10000.0],
"maintenance_factor": [0.0],
}
)
got = equipment_cost_per_use(equipment)["cost_per_minute"][0]
assert _rat(got) == BY_ID["equipment-minute"]
def test_ipps_claim_matches_the_ceiling() -> None:
from ipps.calcs.payment import (
OUTLIER_MARGINAL,
capital_claim,
operating_claim,
outlier_payment,
)
assert _rat(operating_claim(2, 10, 5, 1)) == BY_ID["ipps-operating-cola-one"]
assert (
_rat(
operating_claim(
2,
10,
5,
1,
cola=1,
operating_ime=0.125,
operating_dsh=0.125,
ucp_per_claim=1,
)
)
== BY_ID["ipps-operating-claim"]
)
assert _rat(capital_claim(2, 10)) == BY_ID["ipps-capital-gaf-one"]
assert (
_rat(capital_claim(2, 10, gaf=1, capital_ime=0.25, capital_dsh=0))
== BY_ID["ipps-capital-claim"]
)
got = outlier_payment(
1000,
operating_ccr=0.25,
capital_ccr=0.25,
operating_paid=30,
capital_paid=20,
fixed_loss=100,
wage_index=1,
cola=1,
gaf=1,
labor_share=0.5,
nonlabor_share=0.5,
marginal=OUTLIER_MARGINAL,
)
assert _rat(got) == BY_ID["ipps-outlier"]
def test_mppr_second_rank() -> None:
claims = pl.DataFrame(
{
"hcpcs": ["A", "B"],
"session_id": ["s", "s"],
"non_fac_pe_rvu": [4.0, 9.0],
}
)
result = mppr_reduction(claims, reduction_pct=0.50)
second = result.filter(pl.col("hcpcs") == "A")["pe_rvu_adjusted"][0]
assert _rat(second) == BY_ID["mppr-second"]
def test_committed_cases_match_the_lean_executable() -> None:
lake = shutil.which("lake")
if lake is None:
candidate = Path.home() / ".elan" / "bin" / "lake"
lake = str(candidate) if candidate.is_file() else None
if lake is None:
pytest.skip("lake is not installed")
env = os.environ.copy()
elan_bin = str(Path.home() / ".elan" / "bin")
env["PATH"] = elan_bin + os.pathsep + env.get("PATH", "")
proc = subprocess.run(
[lake, "exe", "pay-cases"],
cwd=PAY,
check=True,
capture_output=True,
text=True,
env=env,
)
assert json.loads(proc.stdout) == CASES