---
title: "Regional Variation in Life Expectancy"
format:
  html:
    code-fold: true
    code-summary: "Show code"
vignette: >
  %\VignetteIndexEntry{Regional Variation in Life Expectancy}
  %\VignetteEngine{quarto::html}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = ""
)

library(targets)
library(DT)

# Shared safe_tar_read with RDS fallback (see inst/vignette_utils.R)
source(file.path(tryCatch(rprojroot::find_root(rprojroot::is_r_package),
  error = function(e) "."), "inst", "vignette_utils.R"))
```

This vignette explores **regional variation in life expectancy** across Western Europe, based on research by Bonnet et al. (2026) published in *Nature Communications*.

## Understanding the Data Structure

**Each row represents aggregated population statistics for one region-year-sex combination**, NOT individual survey responses.

| region_code | year | sex | life_expectancy | What this means |
|-------------|------|-----|-----------------|-----------------|
| FR10 | 2019 | Male | 82.5 | Average LE for **all males** in Île-de-France in 2019 |
| FR10 | 2019 | Female | 87.1 | Average LE for **all females** in Île-de-France in 2019 |
| FR10 | 2019 | Total | 84.8 | Average LE for **entire population** of Île-de-France in 2019 |

The underlying Eurostat data represents **~400 million people** across Western Europe. Life expectancy is calculated from official death registrations and census population counts—not a sample survey.

**Row count formula:** `regions × years × 3 sex categories`

- Sample data: 11 regions × 28 years × 3 = **924 rows**
- Full dataset: 450 regions × 28 years × 3 = **37,800 rows**

## Key Finding: A Two-Tiered Europe

Since the mid-2000s, Western Europe has fragmented into:
- **Vanguard regions**: Continued progress (~2.5 months/year gain for men)
- **Laggard regions**: Stalled improvement (<0.5 months/year gain)

This divergence reversed decades of convergence observed in the 1990s.

```{r regional_variation-chunk-1}
#| echo: false
show_target("vig_regional_classification_summary")
```

## The Microlives Gap

The ~7 year gap between vanguard and laggard regions translates to a substantial lifetime difference in microlives:

```{r regional_variation-chunk-2}
#| echo: false
#| results: asis
g <- safe_tar_read("vig_regional_le_gap")
if (!is.null(g)) cat(sprintf("**%s years** LE gap = **%s lifetime microlives** = **%s microlives/day**", round(g$le_gap_years, 1), g$lifetime_microlives, round(g$daily_microlives, 1)))
```

**Interpretation:** Living in a vanguard region vs a laggard region corresponds to ~`r { gap <- safe_tar_read("vig_regional_le_gap"); if (!is.null(gap)) gap$daily_microlives else "8.4" }` microlives per day—roughly equivalent to the benefit of 30 minutes of daily exercise.

## Regional Data Explorer

**Data period:** 2019 (pre-COVID baseline year, last year before pandemic distortions)

**Column definitions:**

| Column | Definition | Units |
|--------|------------|-------|
| `region_name` | NUTS2 administrative region | — |
| `country_code` | ISO 2-letter country code | — |
| `life_expectancy` | Period life expectancy at birth | Years |
| `microlives_vs_eu_avg` | Daily microlives gained/lost vs EU average | Microlives/day |
| `classification` | Vanguard (top 20% + growing), Laggard (bottom 20% or stagnant), Average | — |

**Key findings:**

- **Vanguard-laggard gap:** ~7 years LE difference = **~8.4 microlives/day** (equivalent to 30 min daily exercise)
- **Gap trend:** Widened from ~5 years (1992) to ~7 years (2019) as laggard regions stagnated post-2005
- **Top region:** Comunidad de Madrid (ES) at 86.1 years
- **Bottom region:** Mayotte (FR overseas) at 74.9 years
- **Microlives interpretation:** +1.0 microlives/day ≈ +30 min life expectancy/day ≈ +7.6 days/year

```{r regional_variation-chunk-3}
#| echo: false
show_target("vig_regional_explorer_data")
```

## Trends Over Time

The divergence became pronounced after 2005:

```{r regional_variation-chunk-4, fig.width=8, fig.height=5}
#| echo: false
#| fig-cap: "Life expectancy trends in Western Europe diverged after 2005, with some regions stalling while others continued to improve."
#| fig-alt: "Line chart of life expectancy over time for multiple Western European regions. Lines converge until ~2005, then diverge, with UK regions showing the slowest improvement."
show_target("vig_regional_trends_plot")
```

## Mortality Risk Multiplier

Use `regional_mortality_multiplier()` to adjust baseline micromort estimates by location:

```{r regional_variation-chunk-5}
#| echo: false
show_target("vig_regional_paris_multiplier")
```

**Application:** If the baseline risk for an activity is 10 micromorts, the location-adjusted risk in Paris would be approximately `10 × 0.93 = 9.3 micromorts` (7% lower due to favorable regional factors).

## Ecological Fallacy Warning

**IMPORTANT:** These regional statistics reflect population averages, not individual-level causation.

High life expectancy in "vanguard" regions results from multiple interacting factors:

| Factor | Mechanism |
|--------|-----------|
| Healthcare access | Better hospitals, preventive care |
| Socioeconomic composition | Higher income, education levels |
| Selection effects | Healthy/wealthy people move to desirable regions |
| Historical factors | Long-term infrastructure investments |
| Cultural factors | Diet, social cohesion, lifestyle norms |

**Moving to Switzerland will NOT automatically extend your life.** The regional advantage reflects the aggregate characteristics of people who already live there.

## Data Source

The regional classification methodology follows Bonnet et al. (2026):

> Bonnet F, et al. "Potential and challenges for sustainable progress in human longevity."
> *Nature Communications* 17, 996 (2026). [doi:10.1038/s41467-026-68828-z](https://doi.org/10.1038/s41467-026-68828-z)

Raw data from Eurostat `demo_r_mlifexp` dataset. Interactive exploration available at the [ReLoG_Europe tool](https://histdemo.shinyapps.io/ReLoG_Europe/).

## Functions Reference

| Function | Purpose |
|----------|---------|
| `regional_life_expectancy()` | Full dataset with filters |
| `vanguard_regions()` | Top-performing regions only |
| `laggard_regions()` | Stagnating regions only |
| `regional_mortality_multiplier()` | Location-based risk adjustment |

## Reproducibility

```{r regional_variation-session-info, eval=TRUE}
sessionInfo()
```

```{r build-info}
#| echo: false
#| results: asis
cat(safe_tar_read("vig_build_info") %||% "")
```