Risk factors for recurrent injuries in victims of suspected non-accidental trauma: a retrospective cohort study

BMC Pediatrics

Table 2 Multivariable Cox Proportional Hazards model for recurrent events

Variable	Hazard ratio	95% CI		P
Lives in a rural area (Non-MSA vs. MSA)*	1.37	1.14	1.67	0.001
Age ≤30 months vs. >30 months	1.67	1.20	2.33	0.002
Dislocation	1.77	1.15	2.72	0.01
Open wound	1.54	1.22	1.94	<0.001
Superficial injury	1.50	1.17	1.92	0.002
Number of injuries
1-2 vs. 0 injuries	1.85	1.42	2.40	<0.001
≥ 3 vs. 0 injuries	1.03	0.74	1.43	0.86

Results are from a Cox proportional hazards Prentice, Williams and Peterson gap time (PWP-GT) model for time between events. Values of the risk factors at the previous event were the independent variables, with the exception of the zip code based variables, which were determined only at the first event. The global p-values for differences among age groups (0–6, 6–12, 12–18, 18–24, 24–30, and >30 months) and number of injuries (0, 1, 2, 3, 4, ≥5) were significant at p < 0.05, so these categories were collapsed into the smallest number of categories showing statistically significantly different associations with the outcome after adjustment for multiple comparisons using Fisher’s least significant difference test. *Based on the child's zip code at their first event. MSA = metropolitan statistical area (urban area).

ISSN: 1471-2431