Life expectancy in 2023 so far within normal range (Germany)

A brief, actual overview

Destatis has now delivered death figures up to calendar week 33. The following analyses are based on them. Documenting the course of the year so far could become very significant at a later date.

Compared with previous years, the overall death rate is in line with expectations.

Fig. 1: Life expectancy of calendar weeks 1 to 33 of each year with trend line.

There had been a weak dip in the spring due to influenza. After that, the curve ran along the expected values. In recent weeks, record values were even reached. Dips near the 30th calendar week due to summer heat were weaker than in previous years.

Fig. 2: Time series of life expectancy in Germany, weekly.

2023 ranked second best since 2017. Surprisingly, the corona year 2020 performed best.

Life expectancy DE

year LE rank

2017 80.76 5

2018 80.52 7

2019 81.03 3

2020 81.18 1

2021 80.92 4

2022 80.75 6

2023 81.07 2

Tab. 1: Life expectancy in years from birth since 2017.

Mortality risks by age cohorts developed differently.

Fig. 3: Risk ratios and 95%-CI compared with the reference period 2016 - 2020. Trends and population changes are taken into account.

While there was almost no excess mortality overall, it continued to rise among persons <60, similar to previous years. Recently, the curve dropped somewhat (Fig. 4).

Fig. 4: Cumulative excess deaths of persons younger than 60.

The figures for younger people may be distorted by a bias due to ongoing immigration. However, late vaccination side effects are also conveivable. Generally, late reports get stronger corrections than older ones.

If the overall numbers look good, it's mostly because of the elderly. They make up the bulk of the huge death waves between late 2020 and late 2022. Such phases are followed by a calming down of the general dying process, called harvesting effect.

You may now ask, where is the sensation and why am I writing this.

The scoop is, Germany is preparing a new vaccination campaign for the fall. In pharmacy, an effect is considered proven if it starts when the drug is administered, disappears when the drug is discontinued, and starts again when the drug is administered again.

And the latter is what is about to happen. However, here we are not talking about a beneficial drug effect, but a lethal one.

---

Methods

- Weekly recording of death rates, populations and mortality risks separately in 14 age cohorts.

- Separate determination of trend and seasonality of mortality risk (TSLM = Time Series Linear Model). Reference period is 2016 to 2020.

- Baysian approach: Forecasts of the age specific mortality risks.

- The expected value of the number of deaths in a cohort in a week is the predicted mortality risk multiplied by the number of living persons in the cohort.

- Life expectancy is calculated from the weekly age-dependent death rates. Missing values are interpolated according to current exponential risk-age function. Cohorts <30 and >90 are treated separately. See Gompertz-Makeham law.