README

A reference cost model and open dataset collection for home care decisions in Canada. The model quantifies the trade-off families face when combining Personal Support Worker (PSW) hours, housekeeping services, nursing care, and informal caregiving for an older adult or person with a disability remaining at home. It spans all ten provinces and three territories, incorporates 2026 tax relief parameters (Medical Expense Tax Credit, Disability Tax Credit, Canada Caregiver Credit, Veterans Independence Program), and emits explicit scope of practice warnings when inputs imply a service substitution that is not legally permissible.

Plain-language summary

Families across Canada routinely face a choice between hiring a Personal Support Worker (PSW) and hiring a cleaner or housekeeper when a parent or spouse begins to need help at home. The two roles are not interchangeable. A PSW is trained and authorised to provide personal care — bathing, toileting, transfers, feeding, medication reminders, and (in some provinces and under delegation) basic wound and ostomy care. A housekeeper or cleaning service is authorised to clean. Asking a cleaner to help with personal care is outside their scope of practice and, in several provinces, outside the scope of their insurance.

This repository publishes a parametric model that takes a recipient's activities-of-daily-living (ADL) and instrumental-activities-of-daily-living (IADL) scores, their province, a short description of their household and diagnosis, and returns a service mix (PSW hours, housekeeping hours, nursing hours), a monthly private-pay cost, the hours covered by the applicable provincial subsidised program, and the federal and provincial tax credits that typically apply. It also flags, in writing, any combination of inputs where cleaning services cannot legally substitute for personal support.

The model is not clinical advice. It is a reference framework for comparing options before a formal care assessment. All data sources are government or peer-reviewed. All code is open source under MIT. All data is published under CC BY 4.0.

What This Does

Given an older adult or disability-care recipient's functional assessment and jurisdiction, the model computes:

Recommended weekly service hours across three categories (personal support, housekeeping, skilled nursing), derived from Katz ADL and Lawton IADL scores with diagnosis- and cognition-aware adjustments.
Scope of practice warnings when the combination of inputs implies a legally-impermissible substitution (e.g. housekeeping for a recipient whose ADL score indicates need for bathing or transfer assistance).
Private-pay monthly cost using per-province rate bands derived from Statistics Canada wage tables and Canadian Home Care Association agency markup ranges.
Subsidised hours allocation for the recipient's province, following publicly-documented eligibility rules (Ontario HCCSS, BC Health Authority home care, Alberta Continuing Care at Home, Quebec CLSC Soutien à domicile, Saskatchewan Health Authority, Manitoba Home Care, Atlantic provinces, Territories).
Federal and provincial tax credit stack — Medical Expense Tax Credit (METC), Disability Tax Credit (DTC), Canada Caregiver Credit (CCC), Veterans Affairs Canada VIP, and Indigenous Services Canada FNIHB Home and Community Care Program flags.
Out-of-pocket cost after credits, monthly and annualised.
Hybrid-versus-all-PSW comparison, showing how much a family saves by using a cleaning service for housekeeping hours rather than paying a PSW rate for those same hours.
Employment-law warnings when private-hire arrangements exceed a threshold that triggers CRA employer obligations (CPP, EI, provincial WSIB, T4 filing).
Data source citations linking every output field to the Sources manifest row that supplied its constants.

Every response includes a disclaimer field: "Reference model only. Not clinical or financial advice; consult a regulated health professional or registered tax practitioner for individual decisions."

Datasets

Three classes of data ship with this repository. All live under datasets/.

Class 1 — Sources (raw, verbatim, credited)

Fetched by datasets/pull_sources.py from authoritative portals and stored under datasets/sources/<organisation>/<filename> with a sibling <filename>.source.json that records upstream URL, retrieval timestamp, license, and SHA256. A human-readable manifest is at datasets/SOURCES.md. No Source file is ever edited after download.

Upstream organisations represented in Sources: Statistics Canada (wage, population, CPI tables), Canadian Institute for Health Information (home care indicators), Canada Revenue Agency (tax folios), Veterans Affairs Canada (VIP rate schedule), Indigenous Services Canada (FNIHB Home and Community Care), Canadian Home Care Association, the Ontario Ministry of Health (HCCSS), and the equivalent provincial and territorial ministries.

Class 2 — Reference tables (hand-curated, deterministic)

File	Rows	Description
`home_care_services_canada.csv`	~112	Scope-of-practice and rate bands for 8 service categories across 13 Canadian jurisdictions, plus federal/cross-jurisdiction rows for VAC VIP and ISC FNIHB. Every row cites a `source_url`.
`home_care_tax_parameters_2026.csv`	~40	Tax-year 2026 parameters for METC, DTC, CCC, VAC VIP, and provincial disability supplements. Every row cites a CRA folio or equivalent.
`home_care_subsidy_programs.csv`	~25	Provincial and territorial subsidised home care programs: administering body, eligibility rules, typical hours awarded, wait time, URL, retrieval date.

Class 3 — Derived scenarios (engine-generated)

These are the analysis-ready tables published to Hugging Face, Kaggle, Zenodo, Dryad, and Mendeley Data. Every row is produced by calling engines/python/engine.py with a specific input grid.

File	Rows	Description
`home_care_scenarios.csv`	5,000	Synthetic household scenarios with population-weighted province sampling, joint-distribution diagnosis/cognition/mobility sampling calibrated to CIHI home care indicators, and per-scenario engine outputs. Deterministic with `random.seed(42)`.
`home_care_per_province_rate_bands.csv`	~650	Product of the reference scope table × Statistics Canada CPI-adjusted wage data for 2019–2026, yielding implied private-pay and agency-billed rate ranges per (province, service, year).
`home_care_cost_model_archetypes.csv`	~2,500	Engine outputs for a grid of canonical household archetypes: every (province × ADL × IADL × cognitive × mobility) combination at constant household composition and income. The lookup-table form of the calculator.
`home_care_tax_relief_sensitivity.csv`	~3,500	Sensitivity grid: ~700 base scenarios × 5 net-income bands ($25k, $50k, $75k, $100k, $150k). Shows how tax-credit value shifts across the income distribution.
`home_care_subsidy_gap.csv`	~390	13 jurisdictions × 30 representative scenarios. Reports subsidised hours awarded vs. model-recommended hours vs. monthly dollar gap, exposing cross-province variation in home care program generosity.

All derived datasets are labelled "Synthetic. Generated from seed 42. Not survey data." in their header comments and dataset cards.

Calculation Model

# Inputs
adl_katz_score        ∈ [0, 6]
iadl_lawton_score     ∈ [0, 8]
province              ∈ {ON, QC, BC, AB, SK, MB, NS, NB, NL, PE, YT, NT, NU}
household_composition ∈ {alone, with_spouse, with_adult_child, multigen}
cognitive_status      ∈ {intact, mild, moderate, severe}
mobility_status       ∈ {independent, cane, walker, wheelchair, bedbound}
primary_diagnosis     ∈ {dementia, stroke, frailty, parkinson, post_surgical,
                         mobility_only, chronic_mixed}

# Hours derivation
psw_hours_per_week    = max(0, 7*(6 - adl) + cognitive_bump
                             - min(informal_hours*0.5, cap))
housekeeping_hours    = max(0, 2*(8 - iadl) + household_modifier)
nursing_hours         = diagnosis_map[primary_diagnosis]

# Gates
if adl <= 4 or cognitive in {moderate, severe}:
    require PSW or HCA for personal care         # scope gate
if diagnosis requires skilled nursing:
    require LPN/RN                               # nursing gate

# Cost stack
private_pay      = hours * province_rate_from(services_canada.csv)
subsidy          = province_subsidy_allocation(province, adl)
out_of_pocket    = private_pay - subsidy

# Tax relief (2026)
metc_qualifying  = max(0, out_of_pocket*12 - min(0.03*net_income,
                                                  METC_FLOOR_2026))
metc_value       = metc_qualifying * (0.15 + provincial_factor)
dtc_value        = DTC_BASE_2026 * (0.15 + provincial_factor)      if eligible
ccc_value        = CCC_BASE_2026 * (0.15 + provincial_factor)      if eligible
vac_vip_value    = VAC_VIP_SCHEDULE_2026                           if veteran
total_credits    = metc_value + dtc_value + ccc_value + vac_vip_value

# Comparison
all_psw_cost     = (psw_hours + housekeeping_hours + nursing_hours)
                   * psw_rate(province)
hybrid_savings   = all_psw_cost - private_pay

The reference implementation is engines/python/engine.py. Six additional language ports implement identical logic; cross-language parity is verified on three case studies from the PDF working paper.

Engines

Seven language implementations. Each exposes POST /calculate and GET /health and loads the reference CSVs at startup.

Language	Directory	Package
Python (FastAPI)	`engines/python/`	`home-care-cost-model` on PyPI
Rust (axum)	`engines/rust/`	`home-care-cost-model-engine` on Crates.io
Java	`engines/java/`	`io.github.davecookvectorlabs:home-care-cost-model-engine` on Maven Central
Ruby (Sinatra)	`engines/ruby/`	`home_care_cost_model` on RubyGems
Elixir (Plug)	`engines/elixir/`	`home_care_cost_model` on Hex.pm
PHP	`engines/php/`	`davecook/home-care-cost-model` on Packagist
Go (net/http)	`engines/go/`	`github.com/DaveCookVectorLabs/home-care-cost-model` on pkg.go.dev

Quick Start (Python)

pip install home-care-cost-model
# or run from source:
cd engines/python
pip install -r requirements.txt
python engine.py              # CLI mode — runs a sample calculation
python engine.py --serve      # HTTP server on port 8003

API Example

curl -X POST http://localhost:8003/calculate \
  -H "Content-Type: application/json" \
  -d '{
    "adl_katz_score": 2,
    "iadl_lawton_score": 1,
    "province": "BC",
    "household_composition": "with_spouse",
    "cognitive_status": "moderate",
    "mobility_status": "walker",
    "primary_diagnosis_category": "dementia",
    "informal_caregiver_hours_per_week": 20,
    "net_family_income_cad": 72000,
    "is_veteran": false,
    "has_dtc": true,
    "agency_vs_private": "private",
    "include_subsidy": true,
    "tax_year": 2026
  }'

Data Sources

Primary government and peer-reviewed sources referenced by the reference tables and cited in every derived dataset row via source_id:

Statistics Canada — wage and employment tables (NOC 44101 Home support workers; NOC 33102 Nurse aides, orderlies and patient service associates); population estimates by age and sex; health services CPI. https://www.statcan.gc.ca
Canadian Institute for Health Information — Home Care Reporting System public indicators; Your Health System home care volume and wait time tables. https://www.cihi.ca
Canada Revenue Agency — Income Tax Folio S1-F1-C1 (Medical Expense Tax Credit); S1-F1-C2 (Disability Tax Credit); Canada Caregiver Credit guidance. https://www.canada.ca/en/revenue-agency.html
Veterans Affairs Canada — Veterans Independence Program rate schedule. https://www.veterans.gc.ca
Indigenous Services Canada — First Nations and Inuit Home and Community Care Program. https://www.sac-isc.gc.ca
Ontario Ministry of Health — Home and Community Care Support Services program pages. https://www.ontario.ca/page/homecare-seniors
Government of British Columbia — Home and Community Care policy manual. https://www2.gov.bc.ca
Alberta Health Services — Continuing Care at Home. https://www.albertahealthservices.ca
Gouvernement du Québec — CLSC Soutien à domicile. https://www.quebec.ca
Provincial ministry pages for Saskatchewan, Manitoba, New Brunswick, Nova Scotia, Prince Edward Island, Newfoundland and Labrador, Yukon, Northwest Territories, and Nunavut home care programs.
Canadian Home Care Association — High-Value Home Care report series. https://cdnhomecare.ca
Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of Illness in the Aged. The Index of ADL: A Standardized Measure of Biological and Psychosocial Function. JAMA. 1963.
Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. The Gerontologist. 1969.

Full provenance is recorded in datasets/SOURCES.md, and every source_url cell in home_care_services_canada.csv points at a specific page within these portals.

Project Structure

home-care-cost-model/
├── datasets/
│   ├── pull_sources.py                         # raw source fetcher
│   ├── SOURCES.md                              # provenance manifest
│   ├── sources/                                # raw files + .source.json sidecars
│   ├── generate_home_care_services_canada.py   # reference table 2a
│   ├── generate_home_care_tax_parameters.py    # reference table 2b
│   ├── generate_home_care_subsidy_programs.py  # reference table 2c
│   ├── generate_home_care_scenarios.py         # derived 3a
│   ├── generate_home_care_per_province_rate_bands.py   # derived 3b
│   ├── generate_home_care_cost_model_archetypes.py     # derived 3c
│   ├── generate_home_care_tax_relief_sensitivity.py    # derived 3d
│   ├── generate_home_care_subsidy_gap.py               # derived 3e
│   ├── home_care_services_canada.csv           # ~112 rows
│   ├── home_care_tax_parameters_2026.csv       # ~40 rows
│   ├── home_care_subsidy_programs.csv          # ~25 rows
│   ├── home_care_scenarios.csv                 # 5,000 rows
│   ├── home_care_per_province_rate_bands.csv   # ~650 rows
│   ├── home_care_cost_model_archetypes.csv     # ~2,500 rows
│   ├── home_care_tax_relief_sensitivity.csv    # ~3,500 rows
│   └── home_care_subsidy_gap.csv               # ~390 rows
├── engines/
│   ├── python/engine.py                        # reference implementation
│   ├── rust/                                   # port
│   ├── java/                                   # port
│   ├── ruby/                                   # port
│   ├── elixir/                                 # port
│   ├── php/                                    # port
│   └── go/                                     # port
├── notebooks/
│   └── home_care_cost_analysis.ipynb           # Jupyter analysis
├── pdfs/
│   ├── generate_pdfs.py                        # PDF working paper generator
│   └── home-care-cost-model-guide.pdf          # 28-page working paper
├── kaggle/                                     # Kaggle dataset + kernel
├── observable/                                 # Observable notebook
├── scribd/                                     # Scribd metadata
├── docs/                                       # Sphinx documentation
├── public/                                     # Minimal PHP web frontend
└── scripts/
    └── parity_check.sh                         # cross-language verification

License

MIT License for code. CC BY 4.0 for datasets and the PDF working paper. See LICENSE for the MIT text; the CC BY 4.0 attribution is specified in datasets/README.md and on the PDF cover page.

Maintained by Dave Cook — Binx Professional Cleaning, North Bay and Sudbury, Ontario.

davecook / home-care-cost-model

Maintainers

Package info

Statistics

Security