What Is Health-Adjusted Life Expectancy?

Beyond the Average

When a financial advisor tells a 65-year-old client that their life expectancy is 84, what does that number actually mean? In most cases, it means the advisor looked up the client's age and sex in a population life table and read off the result. It is a starting point, and for many advisors, it is also the ending point.

The problem is that population life tables describe populations, not people. They aggregate the mortality experience of an entire country: every smoker and non-smoker, every diabetic and marathon runner, every person with congestive heart failure and every person in perfect health. The resulting number is an average across an enormous range of individual outcomes.

Health-adjusted life expectancy is a fundamentally different approach. Rather than asking "how long does the average person this age live?", it asks "how long is this specific person likely to live, given what we know about their health?" The distinction matters for every time-dependent financial decision an advisor makes.

What Health-Adjusted Life Expectancy Is

Health-adjusted life expectancy is an individualized longevity estimate that starts from actuarial mortality tables and then adjusts for the specific health conditions, lifestyle factors, family history, and functional status of the person being evaluated. (You will sometimes see the acronym HALE attached to this phrase — be careful, because the WHO uses HALE for a different population statistic, disambiguated below.)

The output is not a single number. A properly constructed HALE estimate includes a central estimate (the expected age at death, often expressed as both mean and median), a confidence interval that reflects the uncertainty inherent in any longevity projection, and a probability distribution showing the likelihood of survival to various ages.

This richer output gives the advisor something a point estimate never can: a way to plan for uncertainty. Instead of designing a financial plan around a single assumed lifespan, the advisor can stress-test the plan against the full distribution of possible outcomes.

Not the Same as the WHO's HALE

A terminology trap worth clearing up front: the World Health Organization publishes a population statistic called HALE — healthy life expectancy (historically also "health-adjusted life expectancy"), and it measures something entirely different.

The WHO's HALE answers: of the years the average person lives, how many are lived in full health? It starts from total life expectancy and subtracts a disability-weighted allowance for years lived with disease. Under the WHO's 2019 estimates, US HALE at birth is roughly 66 years against a total life expectancy in the high 70s — meaning the average American spends their last dozen or so years managing meaningful health impairment. Women's HALE runs a few years above men's, and the figure is a population average, not a statement about anyone in particular.

The individualized measure this article is about answers a different question: how long is this specific person likely to live, given their actual health? It adjusts the length of life for individual health status; the WHO metric adjusts the quality-weighted portion of an average life. Same words, different objects:

• WHO HALE: population-level, quality-adjusted years in full health, published by country and sex. Useful for public-health comparisons, useless for planning one client's retirement.
• Health-adjusted life expectancy (individualized): person-level, total expected lifespan conditioned on that person's diagnoses, severities, and lifestyle. This is the number a financial plan needs.

If you searched for HALE tables by country or "peak HALE," you want the WHO's Global Health Observatory data. If you want to know how long you — with your specific health profile — are likely to live, you want an individualized estimate, and the rest of this article explains how those are built.

How It Differs From Population Tables

To understand why health-adjusted estimates are more useful, it helps to understand what population tables actually measure and where they fall short.

Period vs. Cohort Tables

Most published life tables are period tables. They capture mortality rates observed in a single calendar year (or short period) and apply those rates across all future ages. They answer the question: "If mortality rates stayed exactly where they are today at every age, how long would this person live?"

But mortality rates do not stay constant. They have been declining for decades across most age groups and most causes of death. A period table systematically underestimates the life expectancy of anyone who will benefit from future mortality improvements, which is nearly everyone.

Cohort tables attempt to correct for this by projecting future mortality improvements. The SOA's MP-2021 mortality improvement scale, for example, models expected annual reductions in mortality rates by age and sex, based on historical trends and expert judgment. Health-adjusted longevity models that incorporate mortality improvement scales produce more realistic projections than those that rely on static period tables alone.

One Size Does Not Fit All

Population tables have exactly two inputs: age and sex. Some tables add a smoking/non-smoking distinction. None account for the dozens of health conditions, lifestyle factors, and family history variables that materially affect an individual's mortality.

Consider the range of conditions that affect longevity:

• Cardiovascular: Hypertension, coronary artery disease, heart failure, atrial fibrillation, history of stroke or MI
• Metabolic: Type 1 and type 2 diabetes, obesity, metabolic syndrome
• Respiratory: COPD, asthma, pulmonary fibrosis, history of pulmonary embolism
• Oncological: Cancer history by type, stage, treatment status, and years since diagnosis
• Neurological: Dementia, Parkinson's disease, ALS, multiple sclerosis
• Renal and hepatic: Chronic kidney disease by stage, cirrhosis, hepatitis
• Lifestyle: Smoking status and pack-year history, alcohol use, physical activity level, BMI
• Family history: Longevity patterns, heritable conditions

Each of these conditions affects mortality, and many of them interact with one another in ways that are not simply additive. A person with both diabetes and coronary artery disease has a mortality profile that is worse than the sum of the two conditions considered independently, because the conditions share pathophysiological pathways that amplify each other's impact.

A health-adjusted model captures these individual conditions, their severities, and their interactions. A population table captures none of them.

The Role of the SOA 2015 VBT

The foundation of any credible health-adjusted longevity model is the underlying mortality table. In the United States, the gold standard for individual life mortality is the Society of Actuaries (SOA) 2015 Valuation Basic Table (VBT) — see our full explainer on the 2015 VBT for how the table itself works.

The 2015 VBT is based on the mortality experience of insured lives, people who have undergone medical underwriting to obtain life insurance. This is an important distinction from population tables, which include uninsured lives. Insured populations tend to be healthier than the general population (a phenomenon called "select" mortality), so the 2015 VBT starts from a higher baseline of health than a CDC or Social Security table.

The 2015 VBT provides mortality rates by:

• Age (from 18 to 120)
• Sex (male and female)
• Smoking status (smoker, non-smoker, and aggregate)
• Underwriting class (preferred best, preferred, residual standard, and others)
• Duration (years since policy issue, capturing the wear-off of underwriting selection)

This granularity makes the 2015 VBT a far more appropriate starting point for individual longevity estimation than a population table. When combined with the MP-2021 mortality improvement scale (which projects annual mortality rate reductions based on historical trends), the result is a forward-looking baseline that accounts for both current mortality patterns and expected future improvements.

But the 2015 VBT, even with mortality improvement, still does not capture individual health conditions. That is where condition modeling comes in.

How Health Conditions Are Modeled

Translating a client's health profile into a mortality adjustment requires three layers of modeling: individual condition assessment, severity grading, and comorbidity interaction analysis.

Individual Condition Assessment

Each health condition is evaluated for its impact on mortality. This is not guesswork. The actuarial and medical literature provides extensive data on the mortality implications of specific diagnoses. Conditions like type 2 diabetes, COPD, and coronary artery disease have been studied across large populations with long follow-up periods, and their effects on mortality are well-characterized.

The key is that not all conditions affect mortality equally. Well-controlled hypertension in a 65-year-old has a modest mortality impact. Congestive heart failure with reduced ejection fraction has a severe one. A health-adjusted model must account for these differences, which requires comprehensive condition modeling that reflects the actuarial evidence.

Severity Grading

The same diagnosis can mean very different things for different patients. Type 2 diabetes controlled with metformin and diet has a different mortality implication than type 2 diabetes with insulin dependence and peripheral neuropathy. Early-stage prostate cancer has a different impact than metastatic pancreatic cancer.

A credible health-adjusted model grades conditions by severity (mild, moderate, severe, or a more granular classification) and adjusts the mortality impact accordingly. This severity grading is one of the areas where the quality of the model matters most, because the difference between severity levels can be the difference between a modest and a dramatic mortality adjustment.

Comorbidity Interactions

Perhaps the most important, and most often overlooked, aspect of health-adjusted longevity modeling is the interaction between multiple conditions.

Human mortality is not a linear sum of independent condition effects. Conditions interact through shared biological pathways, through their combined impact on functional capacity, and through their implications for treatment options. Diabetes and cardiovascular disease interact because hyperglycemia accelerates atherosclerosis. COPD and heart failure interact because both compromise cardiopulmonary reserve. Obesity interacts with nearly everything because it amplifies inflammation, insulin resistance, and mechanical stress.

A model that simply adds up independent condition effects will underestimate mortality for people with multiple interacting conditions. Comprehensive comorbidity interaction modeling is essential for producing accurate estimates in the multi-morbid populations that are most common among the age groups financial advisors serve.

The Calculation, in Plain English

People often ask for "the formula" for health-adjusted life expectancy. There is no single closed-form formula — the calculation is a pipeline, and it helps to see it as four steps:

1. Start from a baseline. Look up the person's annual probability of death at every future age from an actuarial table (age, sex, smoking status).
2. Adjust each year for health. Shift those annual probabilities up or down based on the person's conditions, their severities, lifestyle factors, and the interactions between conditions.
3. Project improvement. Apply a mortality improvement scale so future years reflect expected medical progress rather than frozen current rates.
4. Build the survival curve. Chain the adjusted probabilities together — the chance of reaching age 80 is the product of surviving every year in between — and life expectancy falls out as the probability-weighted sum of future years survived.

A concrete example: take a 68-year-old female non-smoker with well-controlled type 2 diabetes. Step 1 pulls her baseline mortality curve. Step 2 raises each year's death probability to reflect the diabetes — modestly, because it is well-controlled — while her never-smoker status and active lifestyle pull the other way. Steps 3 and 4 turn the adjusted curve into an estimate: perhaps a median around 87 with a 90% confidence range of roughly 78 to 95, versus the low-to-mid 80s a population table would assign her. The point of the example is not the specific numbers — it is that every input moved the whole curve, not just a single output number.

The Role of Monte Carlo Simulation

A deterministic model (one that takes inputs and produces a single output) can generate a point estimate of life expectancy. But a point estimate is insufficient for financial planning, because it hides the uncertainty that is inherent in any longevity projection.

Even with perfect health information, the exact date of death is unknowable. Two people with identical health profiles will not die on the same day. Mortality is a stochastic process. There is inherent randomness that no model can eliminate.

Monte Carlo simulation addresses this by generating a large number of possible mortality outcomes, each drawn from the probability distributions implied by the mortality model. Rather than producing a single life expectancy number, the simulation produces a full distribution of possible ages at death.

From this distribution, the model can extract:

• Mean life expectancy: The average across all simulated outcomes.
• Median life expectancy: The age at which exactly half of simulated outcomes result in death before and half after, often more intuitive than the mean for clients.
• Confidence intervals: For example, the 5th and 95th percentiles, which define the range within which the actual age at death is expected to fall with 90% probability.
• Survival probabilities: The probability of surviving to age 80, 85, 90, 95, etc.

This distributional information is exactly what a financial plan needs. A withdrawal rate strategy, for instance, should not be designed around the median outcome alone. It should be stress-tested against the tails of the distribution. What happens if the client lives to the 90th percentile? What if they only live to the 10th?

Practical Applications for Financial Advisors

Health-adjusted life expectancy with Monte Carlo confidence bands has direct applications across the advisory practice:

Retirement Income Planning

The central question of retirement income planning ("how long does the money need to last?") is a longevity question. Health-adjusted estimates with confidence intervals allow the advisor to set the planning horizon at an appropriate percentile of the longevity distribution, rather than using an arbitrary age like 90 or 95.

For a healthy client, the 90th percentile age might be 97. For a client with significant health issues, it might be 85. Both clients deserve a plan calibrated to their actual risk, not a one-size-fits-all assumption. See our complete guide to longevity planning for a step-by-step workflow.

Insurance Needs Analysis

Life insurance needs depend on how long the insured is expected to live, which determines the premium burden, the expected timing of the death benefit, and the policy's internal rate of return. Health-adjusted longevity directly informs whether to keep, reduce, exchange, or settle an existing policy, a decision framework that requires individualized mortality inputs to execute properly.

Social Security Optimization

The optimal claiming age depends heavily on life expectancy. Health-adjusted estimates allow advisors to give genuinely personalized claiming recommendations rather than relying on the generic advice that defaults to "delay if you can afford to."

Estate and Legacy Planning

For clients with estate planning goals, the expected remaining lifespan affects gifting strategies, trust funding timelines, charitable giving plans, and the projected growth of the estate. A client expected to live another 10 years has a very different estate planning profile than one expected to live another 25 years.

Long-Term Care Planning

The probability and expected duration of long-term care needs are directly related to health-adjusted longevity. Clients with shorter life expectancies may face lower cumulative LTC costs (but potentially higher near-term needs), while clients with longer life expectancies face greater cumulative exposure.

Client Communication

Beyond the technical applications, health-adjusted longevity estimates are powerful communication tools. Showing a client their personalized longevity distribution, rather than a single number from a population table, makes the planning process more tangible and more credible. Clients understand that their health matters. They appreciate when their advisor uses tools that reflect that reality.

Try our calculator to generate a health-adjusted longevity estimate and see how it compares to population averages.

What Makes a Good Model

Not all health-adjusted longevity tools are created equal. When evaluating a model, whether for your own practice or for a client-facing application, consider:

Foundation tables. Is the model built on recognized actuarial tables (such as the SOA 2015 VBT) with appropriate mortality improvement projections (such as MP-2021)? Or is it using ad hoc assumptions?

Condition coverage. Does the model handle the full range of health conditions that affect mortality in the age groups you serve? A model that only handles five or six conditions will miss important factors for many clients.

Severity differentiation. Does the model distinguish between mild and severe manifestations of the same condition? A model that treats all diabetes the same way is missing critical information.

Comorbidity interactions. Does the model account for the non-additive interaction between multiple conditions? This is where many simplified models fall short.

Uncertainty quantification. Does the model produce confidence intervals and survival probabilities, or just a point estimate? A point estimate without uncertainty bounds is less useful for financial planning.

Transparency. Can you understand what the model is doing at a conceptual level, even if the specific parameters are proprietary? You should be able to explain to a client and to a regulator why the model produces the estimates it does.

Validation. Has the model been validated against observed mortality data? Actuarial models should be calibrated and tested, not just built on theoretical assumptions.

The Limits of Any Model

Intellectual honesty requires acknowledging what health-adjusted longevity models cannot do.

They cannot predict the date of death. They produce probability distributions, not certainties. A client with a median life expectancy of 82 might die at 74 or live to 96. The model tells you the distribution of possibilities. It does not eliminate the uncertainty.

They depend on the quality of health information provided. A model can only adjust for conditions that are disclosed. If a client omits a significant diagnosis, the estimate will be too optimistic.

They reflect current medical knowledge and treatment patterns. A breakthrough treatment for a client's condition could change the prognosis. Conversely, a new diagnosis not present at the time of assessment would shift the estimate.

They are not diagnoses or medical advice. A longevity estimate is an actuarial projection, not a clinical assessment. It should inform financial decisions, not replace the client's relationship with their physician.

These limitations are real, but they do not undermine the value of health-adjusted modeling. A probability distribution based on individual health data is always more informative than a population average that ignores individual health entirely. The goal is not perfection. It is improvement over the status quo.

Conclusion

Health-adjusted life expectancy represents a fundamental shift in how financial advisors approach longevity. Instead of treating all clients of the same age as interchangeable, it recognizes that individual health (the conditions they have, the severity of those conditions, and the way those conditions interact) is the primary driver of how long they are likely to live.

Built on actuarial foundations like the SOA 2015 VBT and MP-2021 mortality improvement scales, enhanced with comprehensive condition modeling and comorbidity interactions, and expressed through Monte Carlo simulation as a full probability distribution rather than a single point estimate, health-adjusted life expectancy gives advisors the longevity input that every financial plan needs but few currently have.

The practical applications span the entire advisory practice, from retirement income planning to insurance analysis to Social Security optimization to estate planning. And the communication value is just as important: clients trust advisors who use tools that reflect their individual reality, not a one-size-fits-all average.

Get a longevity report and see how health-adjusted life expectancy changes the conversation with your clients.

Not an advisor? Get your personal longevity report. Health-adjusted life expectancy in seconds.