Occupational Mortality in Spain: An Age-Standardised INE Reanalysis, 2022–2023

Source-verified INE analysis of occupational mortality differences in Spain, 2022–2023. Using published national death registrations and annual employed-person counts, we compared age-standardised all-cause mortality indicators among adults aged 35–64 in nine major occupational groups. This is an independent reanalysis and source audit, not a peer-reviewed causal study.

Research report and verification record · Ilya Nikitin, Ph.D. · The Null Institute · 14 October 2026.

Main result

For elementary occupations versus scientific and intellectual professionals, the ratio of age-standardised indicators was 1.9856 for men (approximate 95% confidence interval 1.8810–2.0960) and 1.3004 for women (1.2188–1.3875). The ratio of those two ratios was 1.5269 (1.4033–1.6614) under the stated model. These compare two specified occupation categories, not the extremes of every group.

Using the same pooled standard weights, the respective men’s ratios were 2.0658 in 2022 and 1.9008 in 2023; women’s ratios were 1.4418 and 1.1757. The comparison remains above one in both years but varies substantially in magnitude.

Data and methodology

Official Spanish National Statistics Institute (INE) inputs: table 32458, registered deaths by single-year age, sex and occupation; and table 69961, annual numbers employed by five-year age band, sex and occupation. We selected national totals, 2022–2023, ages 35–64, nine standard occupational groups, and both sexes. Single-year deaths were aggregated to the six five-year age groups.

For each age × sex × occupation cell, the indicator equals the sum of deaths during 2022–2023 divided by the sum of the two annual employed-person counts, multiplied by 100,000. Rates were directly standardised using the common pooled age distribution of employees across the nine included occupations and both sexes. The confidence intervals use a Poisson-count approximation with fixed weights; the between-sex comparison uses a delta approximation on the log ratio.

Independent source-level verification (14 October 2026)

We retrieved filtered death and employment series directly from the official INE API, independently of the stored Mike’s Balance workbook, and matched them against the original source-level research archive. All 108 age-sex-occupation cells matched exactly, with zero discrepancies in death counts or employed denominators. Recomputed standardised rates, ratios, approximate uncertainty intervals and year-specific sensitivity results matched the published calculations.

Open the source-level verification, 108-cell CSV and runnable reproduction code. The supporting page includes instructions for refreshing values directly from the INE API. This verifies extraction and arithmetic for the chosen tables; it does not establish that all modelling or causal assumptions are valid.

Interpretation and limitations

• These are descriptive annual mortality indicators using registered deaths and annual employed stock counts, not observed person-time. They are not life expectancy estimates.
• No occupational effect is causally identified. Prior health, education, income, selection into employment, occupation changes, migration and register coding may explain part of the differences.
• The population at risk and death counts arise from different administrative processes and may not match perfectly for every individual over the year.
• The reported intervals reflect only the stated count-based statistical model, not systematic errors or unmeasured confounding.
• Unknown occupation, military and people outside the reference employed population are excluded. These results cannot automatically be generalised to all workers or Spanish residents.

Publication and reuse

For the accessible narrative, charts and editorial-ready graphics, see Does Your Job Affect How Long You Live? (Mike’s Balance). Original graphics prepared by Mike’s Balance / The Null Institute may be reused under their indicated CC BY 4.0 licence with credit; INE source data remain subject to their own terms.

Research status: source data and numerical calculations independently checked against INE; not externally peer-reviewed. Correspondence: office@nullinstitute.org.