The Elder Care Index is a composite score that ranks all 50 states and Washington, D.C. across five dimensions of elder care. This page explains how each score is calculated, where the data comes from, how sensitive the rankings are to our choices, and what changed between versions. The current version is 2.0, published November 2026.
Overview
Each state is scored on five pillars. The pillars are put on a common scale and combined with these weights:
| Pillar | Weight | What It Measures |
|---|---|---|
| Affordability | 25% | The cost of care compared with the income of seniors in the state |
| Quality | 25% | Five CMS measures of what happens to nursing home residents |
| Staffing | 20% | Nurse hours per resident and staff turnover |
| Access | 15% | Nursing home beds, home care aides and nursing workforce per senior |
| Medicaid | 15% | How much of the state's Medicaid long-term care for seniors happens at home, and the asset limit |
The composite is shown on a 0 to 100 scale where 50 is the average state. Each pillar score is also shown 0 to 100, where 100 is the best state on that pillar.
Data Sources
- Genworth Cost of Care Survey (2024): median costs for home care, adult day care, assisted living, memory care and nursing homes in every state.
- US Census Bureau, American Community Survey 2025: median income of people 65 and older who live alone (table B19215), and the 65+ population (table B01001).
- CMS Provider Data Catalog, nursing home state averages (October 2026 release): resident-outcome quality measures, nurse staffing hours and staff turnover. Certified bed counts come from CMS Nursing Home Compare.
- CMS / Mathematica, Medicaid Long-Term Services and Supports Expenditures and Users, calendar year 2023 (published October 2025): the share of Medicaid long-term care spending and users aged 65+ in home and community-based services, with CMS data-quality ratings for each state.
- Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025: employment of home health and personal care aides, registered nurses and nursing assistants.
- State Medicaid agencies: asset limits for long-term care Medicaid.
Pillar 1: Affordability (25%)
For each state we compare the annual cost of five kinds of care with the median income of people 65 and older who live alone (the average of the figures for women and for men). We use people living alone rather than all households headed by someone 65+, because household income is inflated wherever many older people live with working family members.
Each care type's cost burden is compared with the burden in the median state, then the five are combined: home care 25%, assisted living 25%, nursing home 25% (the average of semi-private and private rooms), memory care 15%, adult day care 10%. Home care assumes 44 hours a week; adult day care assumes five days a week.
Pillar 2: Quality (25%)
Quality uses five CMS measures, weighted equally, where a lower number is better:
- Long-stay residents who received an antipsychotic medication
- Long-stay residents with pressure ulcers
- Long-stay residents with one or more falls with major injury
- Hospitalizations per 1,000 long-stay resident days
- Short-stay residents who were readmitted to hospital
We do not use CMS health inspection star ratings. CMS grades inspections against other nursing homes in the same state, so every state's average lands near three stars by design, and the stars cannot be used to compare states. To compare individual facilities, use Medicare Care Compare.
Pillar 3: Staffing (20%)
Registered nurse hours per resident per day (50%), total nurse hours per resident per day (25%) and total nursing staff turnover (25%, lower is better), from CMS payroll-based staffing data.
Pillar 4: Access (15%)
Per 1,000 people aged 65+: certified nursing home beds (35%), home health and personal care aides (40%), and registered nurses plus nursing assistants (25%). We count beds rather than facilities, because a count of facilities favours states with many small homes.
Pillar 5: Medicaid Support (15%)
Medicaid is the largest payer for long-term care. This pillar has two parts:
- Home and community care balance (60%): the share of Medicaid long-term care for people 65+ that goes to home and community-based services instead of institutions, measured both by spending and by number of people served. CMS rates the quality of each state's data. We use a measure only where CMS does not rate that state's data "high concern". Where neither measure is usable, the state is scored at the national average for this part, and its page says so. In this release that applies to Colorado, Hawaii, Massachusetts, Minnesota, New Hampshire, New Jersey, New York, Ohio, Vermont.
- Asset limit (40%): the countable assets a single applicant may keep and still qualify for long-term care Medicaid (on a log scale, so the step from $2,000 to $4,000 counts as much as the step from $65,000 to $130,000).
How the Scores Are Combined
Each input is capped at the 5th and 95th percentile across states, so one outlier cannot stretch the scale, then scaled 0 to 100. Inputs are averaged into pillars. The five pillars are then standardised so each has the same spread before the weights are applied, which means the stated weights are the real weights: affordability drives 25% of the differences between states, quality 25%, and so on.
The pillars are combined with a weighted geometric mean rather than a simple average. A geometric mean penalises a very weak pillar, so a state cannot make up for failing one pillar entirely by doing well on another.
Grading
Letter grades reflect how far a state's composite sits from the average state, measured in standard deviations: A at 1.0 or more above average, B from 0.4 to 1.0 above, C within 0.4 of average, D from 0.4 to 1.0 below, and F more than 1.0 below. There is no fixed quota of each grade. Each pillar gets its own grade on the same basis, shown on every state page.
How Sensitive Are the Rankings?
Any index depends on its weights. To show how much, we re-ran the index 2,000 times with each pillar weight shifted at random by up to five percentage points. Each state page and the rankings table show the range of ranks a state lands in across 90% of those runs. In a typical run, 6 of 51 states change letter grade, almost always by one step and almost always states near a grade boundary. A narrow range means the ranking is robust; a wide one means the state's position depends on how much weight you give each pillar, and its pillar grades are the better guide.
Limitations
- State averages hide local variation. A low-scoring state can have excellent individual providers, and the reverse.
- Quality and staffing measure nursing homes. Assisted living, memory care and home care quality are not reported consistently at state level.
- Cost data is a snapshot. Genworth medians are from 2024, and prices vary widely within states.
- Medicaid data lags. The newest federal long-term care spending data covers 2023, and CMS flags some states' data as unreliable (handled as described above).
- Weights are a judgment. They reflect what we think matters most to families choosing care. The sensitivity ranges above show how much that judgment moves each state.
Changelog
Version 2.0.1 (November 11, 2026). We checked every state's long-term-care Medicaid eligibility figures against the state agency's 2026 standards or federal CMS/SSA figures, and corrected 117 of 612 values. Two inputs to the index changed: New Hampshire's asset limit ($7,500 effective for nursing-home and home-care applicants since 2024, not $2,500) and Missouri's ($6,220.50 from July 2026). Because the asset limit is scored relative to all states, every score moved slightly. Grades that changed: New Hampshire D to C overall (Medicaid pillar C to A), Colorado B to C (57.9, just under the B line), and Maine's Medicaid pillar B to C.
Version 2.0 (November 2026). A researcher's question about how the index was built prompted a full stress test of version 1. It found problems, and we rebuilt the scoring:
- Quality: version 1 used CMS inspection star ratings. Because CMS curves them within each state, the state averages differed by less than half a star and quality, nominally 25% of the score, drove about 4% of the differences between states. Replaced with five resident-outcome measures.
- Affordability: version 1 compared raw prices, so cheap states scored well regardless of what seniors there earn, and it counted nursing homes twice (semi-private and private rooms) with price-based weights. Now cost burden against senior income, nursing homes counted once, fixed care-type weights. Independent living, which is housing rather than care, was dropped.
- Medicaid: version 1's income-limit input held the same value for every state and so had no effect, assisted-living coverage was true for 50 of 51 states, ACA Medicaid expansion mainly affects working-age adults, and the scope of home care programs was our own rating. Replaced with measured home-care balance for seniors and the asset limit.
- Access: beds instead of facility counts; current BLS and Census data.
- Method: pillars put on equal spread, geometric mean, grades based on distance from the average instead of a fixed 10-state quota per grade, published rank ranges.
The changes moved many states substantially. Low-cost states with weak nursing home outcomes, low staffing or little home-based Medicaid care fell; several higher-cost states with strong outcomes rose.
Version 1 (2025 to October 2026). Original five-pillar index: affordability, inspection ratings, staffing, facility access and Medicaid program features, combined as a weighted average with grades by rank position.
Per-State Medicaid Eligibility Data
In addition to the composite index above, every state Medicaid page (e.g. Wisconsin, California) presents granular eligibility data — income limits, asset limits, Community Spouse Resource Allowance (CSRA), home equity caps, HCBS waiver programs, spend-down rules, and estate recovery practices. Because the underlying programs differ meaningfully state-by-state, this data is compiled from the following primary sources:
- Medicaid.gov — Federal program rules, waiver approvals, and CMS guidance documents.
- KFF State Medicaid Profiles — State-by-state eligibility thresholds, expansion status, and waiver scope.
- HHS ASPE — Federal Poverty Level guidelines, published annually each January.
- Individual state Medicaid agency publications — Each state's Department of Health Services or equivalent (e.g. Wisconsin DHS, California DHCS, Texas HHSC) publishes the authoritative income/asset thresholds, Miller Trust rules, and program documentation we cite per state.
- Federal SSI Federal Benefit Rate (FBR) — Most states' institutional Medicaid income cap is set at 300% of the SSI FBR, which the Social Security Administration publishes annually.
State-Specific Exception Handling
Several states deviate from the standard rules in ways that materially affect eligibility. We document these exceptions explicitly on the relevant state pages:
- Wisconsin — Despite not formally expanding Medicaid under the ACA, Wisconsin covers childless adults up to 100% FPL through a long-standing Section 1115 waiver (BadgerCare Plus). For long-term care, Family Care, Family Care Partnership, and IRIS are entitlement programs without HCBS waitlists.
- California — Medi-Cal eliminated the asset limit for non-MAGI Medicaid beneficiaries in 2024, then reinstated it at $130,000 for an individual from January 1, 2026. California has no income cap for nursing-home Medi-Cal: residents keep a $35 monthly allowance and pay the rest of their income toward care (share of cost). The 138% FPL limit applies to community Medi-Cal, not nursing homes, and there is no home equity limit until 2028.
- Alaska and Hawaii — Use adjusted Federal Poverty Level thresholds (1.25× and 1.15× the contiguous-48 baseline) for any FPL-indexed eligibility category. Institutional Medicaid income caps remain federal (not state-adjusted).
- Income-cap states (24 states as of November 2026, e.g. Florida, Texas, Ohio, Georgia) — Applicants whose monthly income exceeds the cap ($2,982 in 2026 in most of them) must establish a Qualified Income Trust (Miller Trust) to qualify for long-term care Medicaid. We flag this requirement on each affected state's page.
- Spend-down states — The other 27 states have no hard income cap for nursing-home care: an applicant over the limit qualifies by paying income above a small personal needs allowance toward the cost of care (medically needy, share of cost or a similar rule). This is documented per state.
Refresh Cadence & Known Limitations
- FPL guidelines — Refreshed each January when HHS publishes new poverty guidelines.
- State Medicaid eligibility — Refreshed when states change income/asset thresholds, approve new waivers, or modify expansion status. There may be a delay between a state policy change and our update.
- Cost data (Genworth) — Updated annually when the Cost of Care Survey is released.
- Program enrollment figures — Sourced from CMS reporting and individual state agencies; reporting periods vary by state and may lag the current month.
- Authoritative source. Eligibility determinations are made by your state Medicaid agency; this methodology page documents how we compile, update, and cite their data. See our disclaimer for the long-form scope statement.
Updates
We plan to update the Elder Care Index as new data becomes available from CMS, BLS, and Genworth. Version 2.0 uses Genworth 2024 costs, Census ACS 2025, BLS May 2025, CMS state averages from October 2026 and CMS long-term care spending data for 2023.
If you find an error in our data, please contact us with the page URL, the specific data point, and a source link. We take accuracy seriously and will verify and correct issues promptly. See also our Disclaimer.
Questions about our methodology? We welcome feedback at our About page. See the full rankings or explore state-by-state data.