# U.S.-led study of British school reform finds lower dementia incidence

> A U.S.-led PNAS study used a British schooling reform and UK Biobank records to estimate how an extra school year affected dementia diagnoses.

- Canonical URL: https://newsverum.com/en/news/uk-school-leaving-reform-linked-to-lower-dementia-incidence-decades-later
- Publisher: Newsverum
- Section: Science
- Country: United States
- Published: 2026-10-28T11:03:55+00:00
- Updated: 2026-10-28T11:03:55+00:00
- Event / data period: 2026-10-23
- Other editions: [Türkçe](https://newsverum.com/tr/news/uk-school-leaving-reform-linked-to-lower-dementia-incidence-decades-later.md)

A U.S.-led study reports that a 1972 British school-leaving reform reduced the incidence of Alzheimer’s disease and related dementias (ADRD) among people whose schooling changed because of the law. The paper by researchers at the University of Wisconsin–Madison, the University of Southern California and the University of Michigan appeared online Oct. 23 in the Proceedings of the National Academy of Sciences. It uses the UK reform as a natural experiment, combining a birth-date cutoff with British health records to examine whether schooling itself—not only the circumstances associated with education—affected later dementia diagnoses.

The reform in England, Scotland and Wales raised the minimum school-leaving age from 15 to 16. It applied to people born on or after Sept. 1, 1957; those born on Aug. 31 or earlier could still leave school at 15. The first affected cohort was born from Sept. 1, 1957, through Aug. 31, 1958. The paper estimates that members of this cohort were 14 percentage points more likely to remain in school until at least age 16 than people just before the cutoff. The study’s primary diagnosis endpoint was December 2022, when the first affected cohort had reached 65; control cohorts born earlier were older.

The researchers used UK Biobank records and a regression-discontinuity design, comparing people near the reform’s birth-date threshold. The natural-experiment analysis includes up to 107,200 participants born in England, Scotland or Wales, or who immigrated to the United Kingdom by age 14; other specifications use smaller restricted samples. Dementia outcomes came from hospital records, mortality data and participants’ self-reported medical histories. The paper defines its main outcome as whether someone had been diagnosed by December 2022, which represents cumulative incidence at that point. Administrative linkages allowed follow-up through health and death registries, reducing dependence on continued survey participation.

The preferred estimate for the sample restricted to participants who left school at age 18 or younger indicates that one additional year of education reduced ADRD cumulative incidence by about 0.88 percentage points. The paper reports a 95% confidence interval from -1.763 to -0.154 percentage points and a p-value of 0.022. The authors emphasize that the interval is wide and the study was not powered to pin down the effect’s precise magnitude, even though the evidence supports a negative causal effect. A different, broader specification estimates a different point value; the paper directs attention to the restricted model as its preferred estimate. The 0.88 figure is a percentage-point change, not a relative percentage decrease.

The researchers also tested whether schooling changed the relationship between genetic risk and ADRD. They report evidence that the reform mitigated genetic risk, but the subgroup estimates are imprecise. In the preferred restricted specification, the test comparing effects for the high- and low-genetic-risk groups did not establish a statistically clear difference. The results therefore do not show that the benefit was definitively larger for people with higher inherited risk, even though the study’s overall findings suggest genetic susceptibility is not wholly fixed by biology.

The evidence concerns a specific law, cohort, country and outcome window. Dementia diagnoses were still relatively uncommon by age 65, and the authors caution against overinterpreting the point estimate. As the cohort ages and diagnoses become more common, the measured difference could change. The findings are not a direct test of a current U.S. schooling policy and do not mean that any particular person can prevent dementia by attending school for another year. The study offers causal evidence from one historical British reform; whether a similar policy would have the same effect in a different education system or population remains unknown.

## Data

### Study-reported estimates

First-stage schooling effect and preferred restricted-sample ADRD estimate from the original PNAS paper.

Methodology: The first-stage education change is 14 percentage points for the first affected cohort. The preferred restricted-sample causal estimate is approximately -0.88 percentage points per additional school year for people who left school at age 18 or younger; its 95% confidence interval is \[-1.763, -0.154\] percentage points and the estimate is imprecise. Values are from the original PNAS article, not the NIH summary’s percent shorthand; UK study population only. The school-continuation difference is recorded at the paper’s publication date because it compares birth cohorts around September 1, 1957; it is not a 1957 annual schooling statistic. Dementia incidence uses the stated December 2022 observation cutoff.

| label | value | unit | period | basis | value_qualifier |
|---|---|---|---|---|---|
| School continuation difference at the birth cutoff for the first affected cohort | 14 | percentage points | 2026-10-23 | estimate |  |
| Preferred ADRD incidence effect per additional school year | -0.88 | percentage points | 2022-12 | estimate | approximate |

Sources: <https://www.pnas.org/doi/10.1073/pnas.2529506123>, <https://beta.nih.gov/news-and-media/news/extra-year-secondary-school-lowered-dementia-diagnoses-later-life> · Download data: [JSON](https://newsverum.com/api/v1/articles/uk-school-leaving-reform-linked-to-lower-dementia-incidence-decades-later/data/study-estimates?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/uk-school-leaving-reform-linked-to-lower-dementia-incidence-decades-later/data/study-estimates?lang=en&format=csv)

## Sources

- [Proceedings of the National Academy of Sciences](<https://www.pnas.org/doi/10.1073/pnas.2529506123>) (Source checked: 2026-10-28; Published: 2026-10-23)
- [National Institutes of Health, study release](<https://beta.nih.gov/news-and-media/news/extra-year-secondary-school-lowered-dementia-diagnoses-later-life>) (Source checked: 2026-10-28; Published: 2026-10-23)

---

Cite the canonical URL above. [Editorial standards and corrections](https://newsverum.com/en/editorial)
