Characteristics of the testing set
| Women | Men | |
| No of participants | 43 170 | 28 980 |
| Baseline characteristics | ||
| Age, years | 50.6 (42.5–58.3) | 51.9 (43.2–60.3) |
| Rural areas | 22 449 (52.0) | 15 772 (54.4) |
| Array 1 | 5948 (13.8) | 4503 (15.5) |
| Primary school and below | 23 605 (54.7) | 11 882 (41.0) |
| Daily smokers | 915 (2.1) | 16 317 (56.3) |
| Body mass index, kg/m2 | 23.6 (21.4–26.0) | 23.3 (21.1–25.7) |
| Waist circumference, cm | 78.0 (72.0–84.5) | 81.5 (74.5–88.5) |
| Hypertension | 14 062 (32.6) | 10 653 (36.8) |
| Diabetes | 2477 (5.7) | 1553 (5.4) |
| Family history of stroke | 7619 (17.6) | 5075 (17.5) |
| Follow-up | ||
| Follow-up time, years | 12.6 (11.7–13.4) | 12.4 (11.4–13.3) |
| Total person-years* | 529 498 | 343 421 |
| Incident events† | ||
| Any stroke | 4763 (11.0) | 3751 (12.9) |
| Ischaemic stroke | 4254 (9.9) | 3253 (11.2) |
| Intracerebral haemorrhage | 600 (1.4) | 593 (2.0) |
| Subarachnoid haemorrhage | 87 (0.2) | 45 (0.2) |
Data are presented as n (%) or median (25th–75th percentile) unless otherwise specified.
*Person-years were calculated as the time from the baseline date to the first of the following: death, lost to follow-up or the global censoring date (31 December 2018).
†Only the first event was counted.