Skip to main content
. 2020 Nov 15;108(3):302–313. doi: 10.1007/s00223-020-00775-3

Table 2.

Descriptive statistics for analysis of adverse health outcomes

Characteristic [mean (SD) or N (%)] Men Women
White (n = 789) Black (n = 398) All (n = 1187) White (n = 732) Black (n = 561) All (n = 1293)
Death
 Incidence 510 (64.6%) 282 (70.9%) 792 (66.7%)* 387 (52.9%) 328 (58.5%) 715 (55.3%)*
 Follow-up time (years) 10.6 (6.5, 13.5) 9.2 (4.5, 13.3) 10.2 (5.8, 13.4)* 12.4 (8.6, 13.6) 11.4 (7.1, 13.5) 12.1 (7.9, 13.6)*
Hospital admission
 Incidence 695 (88.1%) 345 (86.7%) 1040 (87.6%)* 617 (84.3%) 476 (84.8%) 1093 (84.5%)*
 Follow-up time (years) 3.0 (1.2, 6.3) 2.9 (1.2, 5.4) 2.9 (1.2, 6.0)* 4.3 (1.8, 8.4) 3.4 (1.3, 7.5) 3.9 (1.6, 8.1)*
 Prevalence before follow-up 305 (38.7%) 154 (38.7%) 459 (38.7%)* 205 (28.0%) 204 (36.4%) 409 (31.6%)*
Low trauma fracture
 Incidence 106 (13.7%) 23 (5.8%) 129 (11.0%)* 209 (29.8%) 63 (11.6%) 272 (21.9%)*,†
 Follow-up time (years) 9.8 (5.7, 11.5) 8.7 (4.3, 11.4) 9.6 (5.2, 11.5) 10.0 (5.0, 11.6) 10.2 (5.6, 11.6) 10.1 (5.2, 11.6)
 Prevalence before follow-up 16 (2.1%) 3 (0.8%) 19 (1.6%)* 42 (6.0%) 14 (2.6%) 56 (4.5%)*,†
Recurrent falls
 Incidence 347 (45.6%) 124 (33.5%) 471 (41.6%)*,† 342 (47.8%) 239 (44.1%) 581 (46.2%)*
 Prevalence before follow-up 162 (21.3%) 61 (16.5%) 223 (19.7%) 147 (20.5%) 115 (21.2%) 262 (20.8%)

Of the entire study sample, 265/3075 (8.6%) died before Year 4. 256 (63.8%) of those who had a low trauma fracture also died during follow-up; figures for hospital admission and recurrent falls were 1369 (64.2%) and 662 (62.9%). Statistics presented for follow-up time starting at Year 4 among 2480 individuals with level and change measures for at least one of the following parameters: gait speed, grip strength, ALM, and fat mass; these statistics differ for hip BMD as the exposure where follow-up time started at Year 3. Individuals with no recurrent falls and missing responses for recurrent falls at Year 4 and after (n = 713) were regarded as not having had recurrent falls for these descriptive statistics as these individuals were included in generalized estimating equations models for recurrent falls (their non-missing responses contribute information)

*Statistically significant sex differences (p < 0.05)

Statistically significant ethnic differences within sex (p < 0.05)