What Counts as a Healthy Year?
The subject here is the family of healthy-life indicators that turns the broad idea of healthspan into an expected duration. The focus is the boundary each indicator draws around “healthy”: a reader should finish able to say what HALE, HLY, and Dis-FLE count, what they treat as loss, and why their numbers cannot be compared by unit alone.
HLE is a family because health has multiple boundaries
A health state is a condition in which a person or population can be observed over a period of time. “Without long-term activity limitation” is one health state. “Without a new severe clinical event” is another. “Full health,” when used in a summary such as HALE, is a state defined through a weighting of health loss rather than a single yes-or-no diagnosis.
An outcome is something that happens to, or is measured about, that state: an illness, injury, functional limitation, hospital-recorded event, or death. An expectancy is different again. It is a summary of how long people are expected to remain in a specified state, given a population, an age, and a method for combining observations. The expectancy does not describe a guaranteed path for any particular person. It describes a distribution of possible paths in a defined group.
This distinction between state, outcome, and expectancy is more than vocabulary. A stroke, for example, is an event that may be recorded clinically. The health state of the person afterward could be described through activity limitation, functional capacity, or the absence of a later severe event. A measure that counts the stroke as a transition is answering a different question from one that asks how the person experiences daily activity. The event can be important evidence without being a complete description of the resulting state.
That difference also explains why a health indicator needs a rule for time. A single observation can tell us that a condition was present at one point; an expectancy summarizes duration. The rule may ask how long people remain in a state after reaching a particular age, or how much of the remaining life is weighted as healthy. It is possible for two measures to use the same people and the same calendar period yet produce different durations because they turn observations into time in different ways.
This is why Healthy Life Expectancy is an umbrella term. The French study describes HLE as a family of indicators that calculate expected years lived in various health states, while also noting that no universally used definition of health has emerged. The disagreement is not merely about statistical technique. Researchers may agree on the question “How many years do people live in good health?” and still disagree about what “good health” means, which observations count as evidence, and how different degrees of health loss should be combined.
A boundary can be set by a threshold or by a weight. A threshold classifies a year as inside or outside a state—for example, whether a long-term activity limitation is reported or whether a specified severe event has occurred. A weight allows health loss to vary in degree, so time is not necessarily treated as wholly healthy or wholly unhealthy. Neither approach is automatically more faithful to lived experience. The threshold is easier to explain and apply consistently in some settings; the weight preserves gradation but requires judgments about how losses are valued.
One way to see the problem is to imagine three boundaries around the same person’s later life. A survey may record a long-term limitation in an activity the person considers important. A hospital database may show no new event among a specified set of severe conditions. A weighted measure may count some years as partly lost to health problems while still treating them as years lived. The three boundaries can overlap without coinciding. A person may have a diagnosed condition that does not produce a limitation captured by a survey; another may report difficulty with activity without having a severe hospital-recorded event. Neither observation automatically cancels the other.
This is why HLE should not be treated as a single agreed score waiting for a more accurate calculation. The family label tells us that the output is an expected duration in a health state. It does not settle which state is most important for a particular purpose. A national comparison, a health-services analysis, and a study of serious disease trajectories can reasonably require different definitions.
The same unit can describe different health states
HALE, HLY, and Dis-FLE make the difference concrete.
HALE: health loss is graded and combined
The WHO definition of Healthy Life Expectancy describes the average remaining years a person can expect to live in “full health” at a given age, after accounting for years lived in less than full health because of disease or injury. Its method uses a life table together with a measure of functional-health loss. Rather than asking only whether a person has crossed a clinical threshold, it weights lived time according to the fraction of healthy life considered lost, then summarizes the result across causes.
That makes HALE a graded construct. A year lived with some health loss need not be treated as identical to a year lived with a much greater loss. The exact weights and the health information behind them matter, which is why the phrase “full health” cannot be interpreted independently of the method. HALE’s breadth is also its trade-off: it can summarize many causes in one population measure, but compression into a single expectancy makes the components of the result less visible. The number is useful as a broad summary precisely because it does not preserve every individual disease or functional pathway.
HALE also has an age attached to it. “At birth” and “at age 60” are not interchangeable statements. The first summarizes expected healthy years from the beginning of life under the measure’s assumptions; the second describes remaining healthy years among people who have reached that age. The age-specific form is part of the meaning, not a formatting detail.
HLY: limitation is defined through reported ability
Eurostat’s Healthy Life Years measure, as described in the French study, uses expected years without long-term activity limitation. The underlying Global Activity Limitation Instrument asks people to evaluate the extent to which health limits their usual activities. This is a different boundary from HALE’s weighted health loss. HLY is concerned with the ability or limitation people report, without requiring the measure to identify the disease responsible.
That design has a practical advantage. A broad activity question can support wide coverage and comparisons across countries and time without requiring every respondent’s conditions to be diagnosed and classified in the same way. It can capture the lived consequence of health, including consequences that may not appear as a severe clinical event. But the same design brings a measurement dependency: responses are self-evaluations. The French study notes the general difficulty of comparing self-reported health and the ways that people’s understanding of health, age, and activity can shape an answer.
HLY therefore should not be dismissed as subjective noise, nor treated as a direct clinical inventory. Perceived limitation is itself a meaningful aspect of health when the question concerns functioning in daily life. The right interpretation is narrower: HLY estimates years without a specified kind of long-term activity limitation as reported through its survey instrument.
Dis-FLE: disease-free means free of specified severe events
The French study’s Disease-Free Life Expectancy takes a clinical route. It follows hospital trajectories and defines disease-free status as the absence of new events from a set of severe conditions considered incompatible with good health in that study. The list is broad, covering conditions associated with serious deterioration, disability, or death, rather than tracking only one disease.
This definition makes a clinical transition observable. A discharge diagnosis, its date, and subsequent records can show when a person crosses the study’s event boundary. That creates a longitudinal view unavailable from a single survey response. It also creates a firm limit around the result. “Disease-free” here does not mean free from every condition, symptom, pain, or limitation. It means free from the specified severe events as recorded in the hospital system. The study itself says that its event definition was not designed to mirror the activity-limitation measure used for HLY.
The clinical boundary can be especially informative for questions about severe population health. It is less informative for a question about whether people can shop independently, feel well, or adapt to a chronic condition without hospitalization. A diagnosis can also have different consequences for different people. The same recorded event may leave one person with substantial impairment and another with a different recovery, while the event-free indicator treats the transition according to its predefined clinical rule.
The three constructs can be held in one mental model:
The diagram is not a ranking. It shows a family relationship: each measure produces an expected duration, but each attaches that duration to a different health state and evidence basis.
What to check before comparing an expectancy
The first check is the health state. Ask whether “healthy” means graded functional loss, reported activity, or freedom from a defined event. A measure can be internally coherent while remaining unsuitable for a different state. It would be a category error to treat a survey estimate of activity-free years as if it were a count of years without severe diagnoses.
The second check is the evidence. HALE combines life-table mortality with information about disease, injury, disability, or functional health loss. HLY relies on people’s reports of long-term activity limitation. Dis-FLE relies on clinical records and the rules used to identify severe events. Each evidence stream can reveal something the others miss. Each also has its own blind spots: missing mortality or morbidity data, non-comparable self-reports, incomplete access to care, and diagnoses that do not fully describe lived function.
The third check is the target population. An estimate can describe all residents, a sex-specific group, people who have reached a particular age, or a selected group visible in a health system. Those qualifiers affect the denominator. The French clinical study is large, but its multivariable analysis concerns a hospitalized population rather than everyone in France. A number calculated in that population cannot silently become a national estimate simply because the dataset contains millions of people.
The fourth check is conditioning. An expectancy at age 65 may mean remaining years for everyone who reaches 65, while a disease-free expectancy at age 65 may mean remaining years for people who have reached 65 without the study’s event. Those groups are not automatically identical. Conditioning changes who is being averaged and therefore changes the question the number answers.
A final check is whether the comparison changes the object being measured. If a reported value rises, that could reflect a change in the underlying health state, but it could also reflect a different survey response pattern, diagnostic practice, event threshold, or population composition. The available evidence does not justify assigning every change to one cause. A useful comparison keeps the observation rule stable or states clearly where it has changed.
These checks explain how two indicators can move in the same direction yet produce different levels. Age may be associated with fewer remaining years under both measures; women may have higher values than men in both; and still the gap between them can vary because one measure captures perceived activity while the other counts severe clinical events. Similar patterns can support a broad observation that health changes with age. They do not prove that the measures are interchangeable.
A useful comparison therefore keeps the label attached to the number: years without long-term activity limitation, years in weighted full health, or years without specified severe clinical events. The label is not bureaucratic decoration. It is the shortest statement of what the number is allowed to mean.
Once this distinction is clear, HALE’s method becomes easier to understand. Its calculation is not a mysterious way to produce a healthier version of life expectancy. It is a specific mechanism for combining survival with health loss across ages. The next step is to follow that mechanism and see how a weighted population summary is built from its inputs.