Skip to content
ParamergeParamerge
Stephen Lieberman
Through Paramerge
Contact

PAN Lab example

Arkansas ARChoices / ARIA

When the tool sets the hours: a home-care hours allocator

An algorithm sets each person's Medicaid home-care hours from a scored assessment, and nurse assessors are bound to its number with almost no room to override. Modeled on Arkansas's ARChoices/ARIA. The risk here isn't a memory loop — it's professional: a tool quietly replacing the judgment it was meant to support.

Stylized model of a documented deploymentPublic benefits & eligibility

Open this example in PAN Lab v0.1 to apply pressures and levers and watch what the system does.

What this models

This example runs on the ARIA-class home-care allocation engine network: 5 components and 13 pathways between them. Every context in the Lab is a stylized model, never a reconstruction of any actual deployment, and each assumption behind it carries a provenance label.

Evidence base: 4 assumed · 2 published baseline. In the Lab, the shaded evidence band behind each headline readout draws its width from the least-established class below.

  • assumed

    The documented consequence is drawn as the action it is: the allocation replicates directly into the hours of care a person actually receives, because assessors are bound by the engine's output - the number on the record is the number at the bedside. The reconciliation before that action is absent, and that is the case's legal core: the courts found due-process violations centred on the inability to understand or contest the determination, which is precisely the step that would put a changed determination in front of the person before it takes effect.

  • assumed

    This models the resource-allocation pattern documented in the Arkansas ARChoices/ARIA case file — not a reconstruction of the actual tool or its formula.

  • baseline

    The strong model-to-model self-loop encodes the documented correlated-error pattern: a single statewide formula meant one change (or one flaw) moved thousands of allocations simultaneously.

  • baseline

    Assessors are bound by the engine's output with only limited override at baseline, reflecting the litigation's account of a tool that constrained professional judgment.

  • assumed

    The standardized-assessment instrument is a dynamics-bearing input node: the fixed InterRAI/RUGs questionnaire the engine scores carries a real inflow into the allocation, so it enters the failure regime rather than sitting inertly on a pathway. Its baseline inflow is modest — the questionnaire is structured, not free-text.

  • assumed

    The harm in the documented case is a benefit cut to served clients — recorded externally in the case file and never computed on a diagram like this one.

What this example does not show

  • The documented harm is a reduction in care hours for disabled and elderly clients. The Lab models institutional propagation, not client outcomes, and estimates no benefit change for any served person; that harm is documented in the case file and measured outside any diagram like this one.

Sources and evidence

What this example rests on, claim by claim. Every entry resolves to the same ledger the Evidence Registry publishes.

  • A large share of Arkansas home-care recipients had care hours cut when algorithmic assessment replaced nurse judgment, and courts found due-process violations centered on the inability to understand or contest determinations.

    empirical
    • Government Arkansas Department of Human Services v. Ledgerwood, 2017 Ark. 308, 530 S.W.3d 336 (Ark. 2017) https://www.courtlistener.com/opinion/4441883/ark-dept-of-human-servs-v-ledgerwood/
    • Government Elder v. Gillespie (8th Cir. 2022) https://caselaw.findlaw.com/court/us-8th-circuit/2088858.html
    • Calo, R., & Citron, D. K. (2021). The Automated Administrative State: A Crisis of Legitimacy. Emory Law Journal, 70(4). https://scholarlycommons.law.emory.edu/elj/vol70/iss4/1/
    • Academic University of Michigan IHPI, What happens when an algorithm cuts your health care https://ihpi.umich.edu/news/what-happens-when-algorithm-cuts-your-health-care
    • Advocacy Benefits Tech Advocacy Hub, Arkansas Medicaid HCBS Hours Cuts https://www.btah.org/case-study/arkansas-medicaid-home-and-community-based-services-hours-cuts.html
    • Advocacy Center for Democracy & Technology, When computer programs cut benefits https://cdt.org/insights/what-happens-when-computer-programs-automatically-cut-benefits-that-disabled-people-rely-on-to-survive/
    • Reference AIAAIC, Arkansas DHS ARChoices RUGs algorithm https://www.aiaaic.org/aiaaic-repository/ai-algorithmic-and-automation-incidents/arkansas-dhs-archoices-rugs-algorithm

  • Clinical assessors bound by algorithmic allocation with limited override capacity form a documented constrained-judgment pattern in home-care assessment.

    empirical
    • Sutton, R. T., Pincock, D., Baumgart, D. C., Sadowski, D. C., Fedorak, R. N., & Kroeker, K. I. (2020). An overview of clinical decision support systems: Benefits, risks, and strategies for success. NPJ Digital Medicine, 3(1), 17. https://doi.org/10.1038/s41746-020-0221-y
    • Advocacy Upturn, Calculated Need: automated home-care hour allocation https://www.upturn.org/work/calculated-need/

Where this connects

Institutional pressures in this domain

  • Austerity & recovery incentives — Cost-cutting and overpayment-recovery targets tilt the system toward denial and enforcement errors.
  • Vendor opacity — The deploying institution cannot inspect the model, data, or update pipeline it is accountable for.
  • Compliance over substance — Paper controls (sign-offs, checklists) satisfy audits while the behavior they describe erodes.
  • Reviewer bottleneck — One fixed-capacity checking stage sits between AI output and consequence; everything queues behind it.

All of them in context on the Public benefits & eligibility domain page.

Levers available here and the patterns behind them

Documented case histories