Skip to main content

Table 2 Self-harm episodes and repeat self-harm or suicide within 6 months by scale cut-off points (missing data imputed)

From: Accuracy of risk scales for predicting repeat self-harm and suicide: a multicentre, population-level cohort study using routine clinical data

Scale Thresholds Repeated self-harm (N = 1133, 28.3%) Did not repeat self-harm (N = 2867, 71.7%) Total (N = 4000) Died from suicide (N = 18, 0.5%) Did not die from suicide (N = 3944) Total (N = 3962)a
Manchester Self-Harm Rule (MSHR) Low risk (0) 23 (2.0) 435 (15.2) 458 (11.5) 2 (11.1) 450 (11.4) 452 (11.4)
Moderate/high risk (1+) 1110 (98.0) 2432 (84.8) 3542 (88.6) 16 (88.9) 3494 (88.6) 3510 (88.6)
ReACT Self-Harm Rule (ReACT) Low risk (0) 63 (5.6) 665 (23.2) 728 (18.2) 4 (22.2) 719 (18.2) 723 (18.2)
Moderate/high risk (1+) 1070 (94.4) 2202 (76.8) 3272 (81.8) 14 (77.8) 3225 (81.8) 3239 (81.8)
SAD PERSONS (SPS) Low risk (0–4) 781 (68.9) 2029 (70.1) 2810 (70.3) 12 (66.7) 2766 (70.1) 2778 (70.1)
Moderate risk (5–6) 251 (22.2) 642 (22.4) 893 (22.3) 4 (22.2) 885 (22.4) 889 (22.4)
High risk (7–10) 101 (8.9) 196 (6.8) 297 (7.4) 2 (11.1) 293 (7.4) 295 (7.4)
Modified SAD PERSONS (MSPS) Low risk (0–5) 1020 (90.0) 2561 (89.3) 3581 (89.5) 15 (83.3) 3532 (89.6) 3547 (89.5)
Moderate risk (6–8) 99 (8.7) 276 (9.6) 375 (9.4) 3 (16.7) 369 (9.4) 372 (9.4)
High risk (> 8) 14 (1.2) 30 (1.0) 44 (1.1) 0 (0) 43 (1.1) 43 (1.1)
  1. a38 individuals were lost to mortality follow-up