Abstract
Using the Vital Statistics three-character tables (by age in 2024, and by year 1995–2024), "falls" were divided into four items — falls on the same level (W01), stairs (W10), buildings (W13), and ladders and other falls (W11 + W17) — and their age patterns and yearly movements were lined up. It claims no new mathematical theorem and no new law.
First, the tables are consistent. (W00-W17) is the sum of 18 three-character items (at every age and sex), and the 2024 numbers match between the age table and the yearly table (Section 3).
Second, of 11,935 fall deaths in 2024, 0.843 (10,066) are falls on the same level; stairs are 632 and buildings 364 (Section 4).
Third, this is the core of the paper. Over the 20 years from 65–69 to 85–89, the rate of falls on the same level rises 32.22-fold (doubling every 3.99 years) and rises at every step from 45–49 to 100+. Stairs, buildings, and ladders and other falls rise only 3.42–4.80-fold over the same 20 years. The share under 65 is 0.033 for falls on the same level, against 0.316 for buildings, 0.263 for ladders and other falls, and 0.163 for stairs (Section 5).
Fourth, from 1995 to 2024, falls on the same level rose from 2,692 to 10,066 (3.739-fold), while falls from buildings went from 940 to 364 (0.387-fold), stairs 0.878-fold, and other falls 0.641-fold (Section 6).
Fifth, the 2016-to-2017 ratio for falls, 1.205, divided by the ratio from recoding the same death records under the new and old classifications (M041's new/old ratio 1.171), is 1.029, within the range of the other years (0.993–1.134). Falls on the same level also rose 1.347-fold from 2017 to 2024 (Section 6).
Sixth, what separates them is where the person fell from (the same level or a height) and age. "Fall deaths increased" is almost entirely about falls on the same level; stairs, buildings, and other falls have decreased.
Scope note: This paper claims no new mathematical theorem and no new law — the death counts are in public tables, and this paper only lined them up by item, age, and year. This paper is a structural interpretation of the shape of numbers in public aggregate data; it is not medical advice, not a diagnosis, and not a treatment recommendation. The author is not a physician. Decisions about individual health should be discussed with a physician or a medical institution. This is a preprint that has not been peer reviewed. It does not go into how to prevent falls, how homes are built, or how care is given. The cause of death is the underlying cause on the death record, and how people who died of another illness after a fall were counted cannot be known from these tables. Why the rates rise with age cannot be separated with these tables. Rates at ages with small counts are unstable, so ages were also combined into larger groups. It does not blame those who lost family members or those who give care. Prefectures are not compared. It is not written as an outlook for any one person. About the data: the Vital Statistics (e-Stat) are used under the Government Standard Terms of Use (Version 2.0), and the ratios in this paper were created by processing them. What this paper adds: dividing falls into the same level versus a height, and showing that only falls on the same level rise with age by an order of magnitude more, and that the yearly increase is almost entirely theirs while stairs, buildings, and other falls have decreased.
1. Notation and question
- Four items: falls on the same level (W01 fall on same level from slipping, tripping and stumbling), stairs (W10 fall on and from stairs and steps), buildings (W13 fall from, out of or through building or structure), and ladders and other falls (W11 fall on and from ladder + W17 other fall from one level to another).
- Rate: per 100,000 population. The population is back-calculated from the 2024 deaths ÷ death rate × 100,000 in Volume 1 Table 5.15 (5-year groups).
- Ratio and doubling years: the 85–89 rate ÷ the 65–69 rate, and 20 × log 2 ÷ log (ratio).
- New/old ratio: the ratio M041 obtained from the material in which the same 2016 death records were counted under the new and the old classification.
Question: Do falls on the same level and falls from a height rise with age in the same way? Which item does the yearly increase belong to?
Observation 1.1: M068 showed that the largest type of unintentional injury changed from transport accidents to falls. M063 showed that the rate of open reduction of femur fractures keeps rising up to 90+. This paper divides falls into their items and lines up their age patterns.
2. Survey and design
To be honest: the design was decided after seeing the tables.
| Content | |
|---|---|
| Tables | Vital Statistics 2024, Volume 2 Table 1-1 (three-character × 5-year age), and Volume 1 Table 5.30 (three-character × year) and Table 5.15 (back-calculating the population) |
| Noticed in the survey | One item, falls on the same level, was 0.843 of all falls. The remaining items are small, so they were grouped into four |
| Measures | Age-specific rates, the ratio from 65–69 to 85–89, the share under 65, yearly ratios, and the 2017 step ÷ new/old ratio |
3. Data and checks
Calculation 3.1 The 3 retrieved files were fixed by SHA-256, and the following was checked.
- Table 1-1: age headings (total, single years 0–4, 0–4, 20 five-year groups, unknown). (W00-W17) = the sum of 18 three-character items (at every age and sex); every row has total = sum of ages + unknown, and men + women = total.
- The 2024 W00-W17, W01, W10, W13, and W17 in Table 5.30 match the same items in Table 1-1.
- The 2016 and 2017 deaths in the falls row of M041's checker output match Table 5.30.
4. Counts by item
Calculation 4.1 Three-character items of falls in 2024 (items with 10 or more)
| Item | Deaths | Share of falls |
|---|---|---|
| W01 same level | 10,066 | 0.843 |
| W10 stairs and steps | 632 | 0.053 |
| W17 other falls | 491 | 0.041 |
| W13 building or structure | 364 | 0.030 |
| W11 ladder | 142 | 0.012 |
| W06 bed | 84 | 0.007 |
| W05 wheelchair | 50 | 0.004 |
Observation 4.2: The largest item is W01, at 0.843. There are also W15 (cliff) 43, W12 (scaffolding) 17, and W07 (chair) 14; the 8 items under 10 total 32.
5. Core — only falls on the same level rise with age by an order of magnitude more
Calculation 5.1 Age-specific rates for the four items (per 100,000, 2024)
| Age | Same level | Stairs | Buildings | Ladders and other |
|---|---|---|---|---|
| 45–49 | 0.29 | 0.11 | 0.16 | 0.09 |
| 65–69 | 1.84 | 0.56 | 0.31 | 0.64 |
| 75–79 | 9.48 | 1.43 | 0.66 | 1.22 |
| 85–89 | 59.13 | 2.67 | 1.25 | 2.19 |
| 95–99 | 263.98 | 3.16 | 1.58 | 1.26 |
| Ratio 65–69 → 85–89 | 32.22 | 4.80 | 4.08 | 3.42 |
| Doubling years (over 20 years) | 3.99 | 8.83 | 9.87 | 11.27 |
Observation 5.2: The rate of falls on the same level rises 32.22-fold over the 20 years from 65–69 to 85–89, and rises at every step from 45–49 to 100+ (465.52). Stairs, buildings, and ladders and other falls rise only 3.42–4.80-fold over the same 20 years. The rate for ladders and other falls peaks at 85–89 (2.19), and buildings is highest at 95–99 (1.58). To be honest: I expected the smallest ratio to be for buildings, but it was ladders and other falls (3.42).
Calculation 5.3 Deaths by age group and the share under 65
| Item | 0–44 | 45–64 | 65–74 | 75–84 | 85+ | Share under 65 |
|---|---|---|---|---|---|---|
| Same level | 45 | 284 | 403 | 2,115 | 7,219 | 0.033 |
| Stairs | 14 | 89 | 120 | 223 | 185 | 0.163 |
| Buildings | 58 | 57 | 65 | 96 | 88 | 0.316 |
| Ladders and other | 43 | 123 | 134 | 201 | 130 | 0.263 |
Observation 5.4: The share under 65 is largest for buildings (0.316) and smallest for falls on the same level (0.033). Of falls on the same level, 7,219 are at 85+.

Figure 1: Age-specific rates by type of fall in 2024 (per 100,000, log scale). Orange is falls on the same level, navy stairs, purple buildings, and green ladders and other falls. Buildings at 100+ is 0, so there is no point.
6. Yearly movement and the 2017 step
Calculation 6.1 Items by year (Table 5.30)
| Item | 1995 | 2005 | 2016 | 2017 | 2024 | 1995 → 2024 |
|---|---|---|---|---|---|---|
| All falls | 5,911 | 6,702 | 8,030 | 9,673 | 11,935 | 2.019 |
| W01 same level | 2,692 | 3,879 | 5,788 | 7,475 | 10,066 | 3.739 |
| W10 stairs | 720 | 687 | 695 | 621 | 632 | 0.878 |
| W13 buildings | 940 | 747 | 509 | 442 | 364 | 0.387 |
| W17 other falls | 766 | 724 | 541 | 607 | 491 | 0.641 |
Observation 6.2: From 1995 to 2024, falls on the same level rose 3.739-fold, while falls from buildings went to 0.387, stairs 0.878, and other falls 0.641. The increase in all falls (6,024) is smaller than the increase in falls on the same level (7,374); the other items together decreased.
Observation 6.3: Dividing the 2016-to-2017 ratio for falls, 1.205, by M041's new/old ratio, 1.171, gives 1.029, within the range of the other year-to-year ratios (0.993–1.134). The 2017 step is nearly accounted for by recoding. Falls on the same level also rose 1.347-fold from 2017 to 2024, when both years are under the new classification.

Figure 2: Deaths by type of fall (1995–2024, table years, log scale). Orange is W01 same level, navy W10 stairs, purple W13 buildings, and green W17 other falls. The vertical dotted line is 2017.
7. Limitations
- The cause of death is the underlying cause, and how people who died of another illness after a fall were counted is unknown. The level of the rates depends on how deaths are counted as well as on how often falls happen.
- Rates at ages with small counts are unstable. Ages were grouped, and rates were read from 45 on.
- The yearly series is counts by item only; it is not divided into age composition and rates (yearly tables by age were not used).
- The population was back-calculated from deaths and death rates. It does not go into fall prevention. Prefectures are not compared. It is not an outlook for any one person.
Conclusion
In 2024, 0.843 of fall deaths were falls on the same level. Their rate rose 32.22-fold from 65–69 to 85–89 and kept rising up to 100+, while stairs, buildings, and ladders and other falls rose only 3.42–4.80-fold. From 1995 to 2024, falls on the same level rose 3.739-fold and falls from buildings fell to 0.387. The 2017 step, divided by the recoding ratio, fell within the range of the other years.
Two things separate them — where the person fell from (the same level or a height) and age.
On hearing "fall deaths increased," check together whether it is falls on the same level or from a height, which ages are meant, and whether the way of counting moved that year.
Data and reproduction
| Data | Vital Statistics 2024 (final figures), Volume 2 Table 1-1, and Volume 1 Tables 5.30 and 5.15 (Ministry of Health, Labour and Welfare; e-Stat) (retrieved 26 September 2026) |
| Terms of use | e-Stat follows the Government Standard Terms of Use (Version 2.0). The ratios in this paper were created by processing it and are not figures produced by the government |
| Scope | Japan, 2024 (yearly 1995–2024) |
| Checker | 検算/M071.py (--zu also draws the figures). Every number in the text is taken from 検算/M071_output.txt. The new/old ratio is read from the output of M041's checker |
| Raw data | The retrieved files and SOURCES.md in data/ (URL, retrieval date, SHA-256, terms of use, definitions and how calculations are set up) |
| Figures | 図/M071_zu1, 図/M071_zu2 (png, svg, English versions) |
References
- Ministry of Health, Labour and Welfare. Vital Statistics 2024 (final figures), Volume 2 Table 1, and Volume 1 Tables 5.30 and 5.15 (in Japanese). e-Stat (accessed 26 September 2026). https://www.e-stat.go.jp/
- Yamagishi, Y. (2026). The Fall in Pneumonia Deaths in 2017 Is Nearly Predicted by the Ratio from Recoding the Same Death Records under the New and Old Classifications — for Pneumonia, an Actual Step of 0.826 Against a New/Old Ratio of 0.810, a Correlation of 0.969 Across 65 Causes, and the Lost Deaths Moved to Dementia, Alzheimer's Disease, Heart Failure, COPD, and Others [M041]. Zenodo. https://doi.org/10.5281/zenodo.22974374
- Yamagishi, Y. (2026). Rates of Open Reduction of Fracture Peak at Different Ages by Site — For the Femur Both Sexes Peak at 90+ (1966.2 per 100,000 for Women, 833.9 for Men) and Double Every 4.25–5.84 Years from 75–79 On, While Women's Forearm Peaks at 80–84 and Lower Leg at 75–79 [M063]. Zenodo. https://doi.org/10.5281/zenodo.22976616
- Yamagishi, Y. (2026). In 2024, Unintentional-Injury Deaths Recorded as Occurring at Home Were 17,320, 5.59 Times Road Traffic Deaths (3,100), but 17,625 Had an Unspecified Place — Home Accidents Are Fewer Than Road Traffic Deaths Only at 15–44, and 10.39 Times at 80+ [M068]. Zenodo. https://doi.org/10.5281/zenodo.22977842
Reference
Yamagishi, Y. (2026). The Death Rate from Falls on the Same Level Rises 32.22-Fold from 65–69 to 85–89 and Keeps Rising up to 100+, While Falls from Stairs, Buildings, Ladders, and Others Rise Only 3.42–4.80-Fold [M071]. Zenodo.
DOI: 10.5281/zenodo.22978072
YOI: https://www.yamagishiyuuki.com/yoi/M071
ORCID: 0009-0006-8587-3450
PDF (English) · PDF (Japanese) · Date proof of the English PDF (OpenTimestamps .ots) · SHA256SUMS
The record page for this paper (in Japanese, with the PDFs in each language and the date proof): YOI M071. This paper is a preprint and has not been peer reviewed.