Years of Life Lost

Definition

Years of life lost (YLL) is the sum, across all deaths from a given cause, of the expected remaining life expectancy of each decedent at the time of death. It measures the burden of premature mortality from the perspective of the person who died, in units of person-years. YLL is distinct from population-level life expectancy loss (LEL), which divides total YLL by population to get average impact per person (living or dead).

Key Ideas

YLL per decedent: For each COVID death, YLL = expected remaining life expectancy without COVID. A 70-year-old with 13 remaining expected years loses 13 YLL; a 40-year-old loses ~38 YLL.
Population LEL vs. YLL per decedent: Population LEL = (total YLL) / (total population). Because the fraction of the population dying of COVID is small (population fatality rate [PFR] ≈ 0.2%), even large YLL per decedent translates to modest population LEL. Parra et al. (2022): national YLL/decedent = 13.7 years but population LEL = 11 days.
YLL vs. YPLL: Years of Potential Life Lost before age 75 (YPLL-75) is a related but cruder measure that assigns no weight to deaths above age 75 — understating elderly burden. YLL is preferable because it uses individual remaining life expectancy (which varies by age, sex, race, and socioeconomic status (SES)) rather than a fixed cutoff.
YLL as cause-of-death burden measure: Used in the Global Burden of Disease (GBD) framework alongside Years Lived with Disability (YLD) to compute Disability-Adjusted Life Years (DALYs). COVID-19 drug use disorder YLL rose 878% 1990–2021 in the GBD 2021 study.
Cohort vs. period YLL: YLL calculations require estimates of remaining life expectancy without the disease. The choice between period life expectancy (LE) and cohort life expectancy for this baseline affects the estimated YLL — period LE is typically used for simplicity but understates future life expectancy under improving mortality trends.
COVID temporal shift: In the three Midwest areas, mean COVID YLL per decedent rose from ~12–13 years in 2020 to higher levels by 2021 as younger individuals accounted for a growing share of deaths. By Q3 2021, mean YLL for Hispanic male COVID decedents exceeded 26 years (Parra et al. 2022).
Age profile: YLL per decedent falls with age (older decedents have fewer remaining years), but population LEL rises with age because PFR rises sharply (more elderly are dying of COVID). These two effects work in opposite directions.

How It Works

For a population:

$\text{Total YLL} = \sum_{\text{decedents}} \hat{e}(a_i, g_i, r_i)$

where $\hat{e}(a_i, g_i, r_i)$ is estimated remaining life expectancy for decedent $i$ given their age, gender, and race/ethnicity (without the disease in question).

Population LEL is then:

$\text{LEL} = \frac{\text{Total YLL}}{N}$

where $N$ is total population.

The cohort-based variant integrates over the pandemic period: $\text{LEL} = \text{LE}_\text{without COVID} \times \text{PFR}$ , where PFR is the population fatality rate — the fraction of the population who died of COVID during the observation window.

Why It Matters

Policy relevance: YLL focuses attention on premature deaths that are most preventable and costly in terms of lost productive and family life. A disease killing mostly the elderly (high count, low YLL per decedent) differs in policy terms from one killing working-age adults (low count, high YLL per decedent).
Cohort vs. period frame for time-limited events: For a pandemic expected to end, period-based LE loss estimates (which assume pandemic mortality rates persist forever) overstate the population burden by 1–3 orders of magnitude for younger ages. The cohort approach using cumulative PFR is more appropriate. See Period vs. Cohort Life Expectancy.
Racial/ethnic equity: Higher YLL per decedent for Blacks and Hispanics (who die younger of COVID) than for Whites illustrates how racial mortality disparities operate not just through count of deaths but through age at death. Population LEL disparities reflect both higher PFR and younger age at death.
Disability insurance (DI) context: DI beneficiaries have elevated mortality; measuring YLL for this population relative to age/sex-matched general population would help quantify the mortality burden of denied vs. awarded DI claims.

Open Questions

How should comorbidities be incorporated into the remaining LE baseline? Decedents with multiple chronic conditions likely had shorter expected lives than the age/sex/race average, so unadjusted YLL overstates burden.
What is the appropriate observation window for a pandemic? Parra et al. (2022) covers through Sep 2021; subsequent waves (Delta, Omicron) continued to shift the age and racial distribution of COVID deaths.

Sources

Parra et al 2022 — The COVID-19 Pandemic Years of Life Lost and Life Expectancy Decomposition Using Individual-Level Mortality Data