Why model health care costs separately?

Health care is one of the largest and most variable expense categories in retirement. A targeted model helps you: prioritize savings, test funding strategies (HSA, IRAs, Roth conversions), and avoid shocks from unexpected costs like long‑term care or high prescription drug bills. In my practice over 15 years, clients who ran a focused health‑cost projection were more durable in stress tests and avoided late‑life liquidity gaps.

Authoritative context: Fidelity and other industry estimates highlight that a 65‑year‑old couple can expect several hundred thousand dollars in health‑related spending over retirement (see Fidelity Retiree Health Care Cost Estimate). Medicare helps but does not eliminate exposure to premiums, deductibles, co‑pays, and long‑term care (CMS explains Medicare coverage limits). The CFPB provides consumer‑facing guidance about planning for medical and long‑term care costs (CFPB).

Core inputs you must collect

A useful model starts with realistic, documented inputs. Collect these for each retiree:

  • Current age and planned retirement age.
  • Present health status and known chronic conditions (hypertension, diabetes, mobility issues).
  • Current annual medical spending (premiums + out‑of‑pocket) and recent 12‑month receipts.
  • Current health insurance: employer plan details, HSA balances, Medicare enrollment expectations.
  • Expected Medicare start date(s) and whether you’ll use Medicare Advantage, Medigap, or buy Part D.
  • Long‑term care (LTC) exposure: family history, functional status, and willingness to pay for home care or facility care.
  • Geographic location (medical costs vary by state/metro).
  • Social Security or other income projections (for IRMAA exposure).

Documenting these inputs forces realistic assumptions and makes the model auditable.

Step‑by‑step modeling template (practical)

1) Set a projection horizon: typically until age 90–95 for one spouse; use joint‑life for couples.

2) Establish a baseline first‑year health cost per person:

  • Sum premiums (Medicare Part B/Part D/Medigap or employer), supplemental plans, and expected out‑of‑pocket on top of insurance (deductibles, coinsurance, routine care).
  • Example: baseline annual cost = $8,000 per person (premium $3,500 + out‑of‑pocket $4,500).

3) Choose an annual medical cost inflation rate:

  • Historically, medical inflation exceeds CPI. Use a base assumption of 3.5%–5.0% real (or 1–2% above your expected general inflation). Run sensitivity tests at ±1–2%.

4) Escalate per‑person costs each year using the chosen inflation rate.

  • Year t cost = Year 1 cost * (1 + inflation)^(t‑1).

5) Layer in discrete events:

  • Major surgery, cancer treatment, or a transition to assisted living: model as one‑time or recurring higher costs in the years they occur.
  • Long‑term care events: estimate months of home care or facility care and multiply by local rate.

6) Add tax and income interactions:

  • Consider IRMAA (Medicare income‑related adjustments) if joint or individual modified adjusted gross income (MAGI) crosses thresholds — plan withdrawals or Roth conversions accordingly (see guidance on avoiding IRMAA surprises).
  • Use HSA balances first for qualified medical costs tax‑efficiently if available; track contribution rules before Medicare enrollment (consult CMS/HSA rules).

7) Compute cumulative cost and present value (optional):

  • Sum nominal costs across the horizon for a gross liability.
  • Discount future costs to today’s dollars using a real discount rate (e.g., 2–3%) to get present value for planning comparisons.

8) Run scenarios and sensitivity tests:

  • Baseline, high‑cost, low‑cost, early LTC event, and medical‑inflation shock.

Example projection (concise, realistic)**

Assume a 65‑year‑old spouse with baseline Medicare + supplemental premiums and out‑of‑pocket = $10,000/year. Use 4% annual medical inflation and project to age 90 (25 years):

  • Year 1: $10,000
  • Year 10: $10,000 * (1.04)^9 ≈ $14,000
  • Year 25: $10,000 * (1.04)^24 ≈ $25,900

Cumulative unadjusted nominal spending over 25 years ≈ $510,000 for that person. For couples, combine both—many planners use joint‑life assumptions rather than maxing both to age 95. This example shows how compounding increases lifetime exposure and why a focused plan matters.

Note: these are illustrative calculations. Your inputs will differ by health, location, and coverage choices.

Modeling long‑term care (LTC)

Long‑term care is the most volatile and often the largest single addition to retirement health costs. Options for modeling LTC:

  • Self‑funding path: estimate probability of needing care, typical duration (months/years), and local cost (home care vs. assisted living vs. nursing home). Multiply probability × duration × cost and add to the health model as an expected value. Also run high‑impact scenarios where probability = 100%.

  • Insurance path: model the cost of a hybrid LTC policy or traditional LTC policy premiums, including inflation protection and elimination periods.

  • Hedge approach: keep a liquid bucket sized to cover expected LTC events in the early years, supplemented by long‑term care insurance for tail risk.

In practice, I build both an expected‑value line in the main projection and a separate contingency plan (liquid reserve or policy) for catastrophic scenarios.

Tax, HSA, and Medicare interplay (practical points)

  • HSAs are one of the most tax‑efficient ways to pre‑fund medical expenses before Medicare enrollment. After enrolling in Medicare, HSA contributions stop but balances remain available tax‑free for qualified medical expenses.

  • Roth conversions can reduce future IRMAA exposure because IRMAA is based on reported MAGI in prior years; however, conversions themselves can push MAGI higher and trigger IRMAA if timed poorly. Coordinate with expected Medicare Part B/D premium windows (see FinHelp article on Roth timing and IRMAA).

  • Claiming Social Security and Medicare enrollment timing affect premiums, coordination of benefits, and cash flow. See our guide on planning Medicare‑age health expenses before and after 65 for enrollment considerations.

Interlinks: read more on coordinating Medicare and retirement health costs in plan projections and strategic use of HSAs and Medicare coordination.

Sensitivity testing and scenario design

Good models don’t give a single number — they show ranges. At minimum, run:

  • Conservative case: lower medical inflation (3%), healthy baseline, no LTC.
  • Base case: mid‑range inflation (4%), modest baseline, 20–30% LTC probability.
  • Severe case: higher medical inflation (5–6%), chronic conditions, one spouse needs multi‑year LTC.

Key outputs to compare across scenarios: peak annual expense, cumulative nominal cost, present value of costs, and years where health costs exceed a defined safe withdrawal bucket.

Common mistakes to avoid

  • Treating Medicare as full coverage. Medicare has gaps; you still need to account for premiums, deductibles, cost‑sharing, and services it doesn’t cover (CMS).
  • Ignoring long‑term care or assuming family caregiving will be available indefinitely.
  • Using a single inflation assumption — health care inflation historically outpaces general inflation.
  • Failing to model the tax interactions of withdrawals and Roth conversions that affect Medicare premiums (IRMAA).

Practical tools and implementation tips

  • Spreadsheet templates: build annual rows for each spouse, columns for premium, out‑of‑pocket, LTC, and a summary column. Include toggles for inflation, LTC probability, and discount rate.

  • Use scenario toggles with simple macros or a data table instead of full Monte Carlo unless you are modeling portfolio drawdowns alongside health shocks.

  • Update annually or after major life events (diagnosis, move to a different cost‑of‑living area, or change in policy).

  • If you have an advisor, ask for stress tests that add a front‑loaded LTC event and increased medical inflation for a decade.

Frequently asked questions

Q: How often should I update the model?
A: Annually, and after any major health or insurance change.

Q: Should I assume both spouses live to 95?
A: Use joint‑life assumptions for conservative planning, but you can test median longevity scenarios as well.

Q: Is long‑term care insurance worth it?
A: It depends on family history, health, and budget. Model both self‑funding and insured paths and compare expected values and tail risk.

Sources and further reading

  • Fidelity Investments, Retiree Health Care Cost Estimate (industry estimate on lifetime retiree health spending).
  • Centers for Medicare & Medicaid Services (CMS) — Medicare coverage and premium guidance. https://www.cms.gov
  • Consumer Financial Protection Bureau (CFPB) — consumer guides on planning for medical and long‑term care costs. https://www.consumerfinance.gov

Professional disclaimer

This article is educational and not personalized financial, tax, or medical advice. Your circumstances, state rules, and plan choices differ. Consult a financial planner, tax advisor, or elder‑care specialist before implementing strategies described here.

If you’d like, I can convert this template into a downloadable spreadsheet with toggles for inflation, LTC probability, and example scenarios.