Model Card, Method and Sources

At a glance

Model status and intended use

Intended use Educational self-reflection

Bring 46 lifestyle, health, relationship and environment signals into one structured view.

Current output Rule-based model value

A free input-based sensitivity range followed by the specific model value and explanatory features after unlock.

Validation status Internal checks only

Internal plausibility, boundary and consistency checks; no independent clinical validation.

Range status Calibration not yet claimed

No probability or coverage percentage is published until independent outcome data, calibration and coverage testing are available.

What is calculated

A rule-based educational model value

The life-expectancy value is produced by a deterministic rule set, not generated by AI. The current calculation starts from a model ceiling of 110 modelled years and applies positive or negative adjustments linked to questionnaire answers. The paid result reveals this calculated point value and the factors that shaped it.

How personal context is used

Birth year determines current and remaining years and activates a small number of age-dependent rules.
Sex selects the male or female coefficient set used by the current model.
Country selects checkout currency; it does not change the life-expectancy score.
AI is used only for optional educational summaries and Action Plan explanations after the core calculation.

Questionnaire coverage

Factors that can enter the calculation

Individual questions are grouped below so visitors can see the breadth of the model before sharing health-related information. Not every answer changes the result, and some “unknown” answers use a documented fallback or smaller adjustment.

Personal context

Birth year, sex, height, weight and BMI interpretation.

Sets age context, sex-specific assumptions and body-size categories.

Movement and recovery

Exercise frequency, leisure activity, muscle context, sleep and brain activity for older adults.

Applies activity, inactivity and recovery-related model adjustments.

Smoking and alcohol

Smoking history, regular alcohol level and occasional higher intake.

Uses sex-specific lifestyle-risk adjustments where the model provides them.

Nutrition and prevention

Eating pattern, snacks, meat, fruit and vegetables, sweets, dairy, carbohydrates, coffee, tea, dental care, sunscreen and screening.

Combines dietary and prevention-related rule adjustments.

Measurements and history

Blood pressure, cholesterol, HDL, diabetes or blood sugar, heart-attack history and regular bowel function.

Uses category-based adjustments; it does not interpret laboratory records or diagnose disease.

Stress, connection and environment

Stress level and coping, social connection, relationship status, work pattern, education, air quality and risky behaviour.

Represents psychosocial, environmental and behavioural associations as simplified model rules.

Family background

Family longevity, relatives reaching advanced ages, family cardiovascular or diabetes history, cancer history and childbirth age where applicable.

Provides inherited and family-history context; it is not genetic testing.

Evidence map

Core foundations and research context

The concrete LifeTrajectory rule set was adapted primarily from the two foundations below. The additional papers support the broader research context of selected factor groups; they do not individually validate every LifeTrajectory coefficient or convert an association into a personal causal effect.

Core reference framework

Living to 100 Life Expectancy Calculator

A questionnaire framework derived from centenarian research and the primary structural reference for many LifeTrajectory questions and category-based adjustments. It is a source framework, not an independent clinical validation of LifeTrajectory.

View the original calculator

Core published study

Li, K., Liu, C., Liao, L. et al. — BMC Medicine (2014)

Lifestyle risk factors and residual life expectancy at age 40: a German cohort study.

Read the full article

Centenarian research context

New England Centenarian Study

Research context on genetic, family and lifestyle characteristics associated with exceptional longevity.

View the study overview

Selected factor context

Large cohort reviews and meta-analyses

Free result versus full result

How the model sensitivity range is constructed

Method ID: model-sensitivity-band-v1 · Input-based model sensitivity range; internally checked, not externally calibrated.

The free result starts with the calculated model value E. For selected high-impact factors, LifeTrajectory changes one answer at a time to the nearest documented response category and recalculates the value. For an unknown health measurement, documented plausible alternatives are used. Calculation numbers play no role in the range.

Local retests dᵢ = recalculated value − E

Per-factor movement Keep at most one largest decrease and increase per factor

Directional margins m = √(2.5² + sum of squared factor movements)

Displayed bounds L = max(current age, E − m₋) · U = min(110, E + m₊)

The 2.5-year term is a disclosed model-structure allowance, not an empirically estimated error. Root-sum-square combination avoids simply adding every one-factor movement as if all factors changed to an extreme together. The final bounds are rounded to one decimal place. The range is deterministic for the same saved answers and its midpoint need not be E.

This method is a transparent internal sensitivity analysis. It does not yet express probability, statistical coverage, observed prediction error or external validation. A probability or coverage percentage will not be published unless independent longitudinal outcome data support calibration and out-of-sample coverage testing.

Unlocking replaces the preview with the specific calculated model value under the same calculation number and adds the explanation, factor maps, Action Plan and progress features.

Separately, neither the sensitivity range nor the calculated value is a medical prediction. LifeTrajectory provides educational information and does not provide diagnosis, treatment, or individual medical advice.

Important limitations

What the model cannot establish

Population-level associations and simplified rules cannot determine how long an individual will live.
The current model is not calibrated to a country-specific life table and does not provide a probability distribution.
Related factors can overlap. LifeTrajectory keeps them visible as separate model rules rather than claiming that every effect is causally independent.
Not every coefficient has an individual peer-reviewed citation mapping, and the complete rule set has not been independently clinically validated.
The output is educational information, not medical advice, diagnosis, treatment or emergency guidance.

Version history

Current documentation release

LifeTrajectory LE 2026.08.1 · August 17, 2026

Added calculation-level model versioning, a complete public factor overview, explicit disclosure of the 110-year model ceiling, country handling, input-based model-sensitivity construction, enforced output boundaries, internal consistency checks and validation limits. The sensitivity method is internally testable, but no statistical coverage or independent external validation is currently claimed.

Questions, source corrections or requests for technical review: support@lifetrajectory.live