Timeliness, simulated hemorrhage-control success, and technical errors across one-handed self-aid/restricted-hand and two-handed buddy-aid/responder-assisted windlass tourniquet training pathways: a two-year registry study

Rapid control of life-threatening extremity hemorrhage is central to tactical and civilian prehospital care. Operational one-handed self-aid/restricted-hand and two-handed buddy-aid/responder-assisted training pathways impose different task demands, but training studies often emphasize speed without separating technical errors and measurement quality. To compare primary timing and recorded simulated hemorrhage-control success between one-handed self-aid/restricted-hand and two-handed buddy-aid/responder-assisted operational training pathways and to characterize actionable technical-error patterns. This retrospective training-registry study included 200 unique operators assessed from 15 January 2024 to 19 December 2025 (98 one-handed and 102 two-handed records). Primary outcomes were total application time and recorded simulated hemorrhage-control success. Unadjusted effects were reported with 95% confidence intervals (CIs). Adjusted timing was analyzed with median quantile regression, and binary success with Firth penalized logistic regression. An objective-only sensitivity analysis excluded legacy checkbox-only records. Median total application time was 51.0 seconds for the one-handed pathway and 42.5 seconds for the two-handed pathway (median difference, -8.5 seconds; 95% bootstrap CI, -13.0 to -5.5; p<0.001). Recorded simulated hemorrhage-control success was 82/98 (83.7%) versus 96/102 (94.1%), respectively (risk difference, 10.4 percentage points; 95% CI, 1.7 to 19.6; p=0.018). The adjusted median difference was -9.0 seconds (95% CI, -11.6 to -6.3; p<0.001). Firth regression estimated higher odds of recorded success with the two-handed pathway (adjusted OR, 4.75; 95% profile-likelihood CI, 1.59 to 16.64; p=0.004), with a similar estimate after additional adjustment for sex, glove use, and limb location (OR, 5.03; 95% profile-likelihood CI, 1.65 to 18.37; p=0.004). In the 184 records with objective fields, success was 75/90 (83.3%) versus 90/94 (95.7%), a risk difference of 12.4 percentage points (95% CI, 3.6 to 21.8). Incomplete rod securing and second-tourniquet use showed the clearest between-pathway technical-error differences. The two-handed buddy-aid/responder-assisted pathway was associated with faster application and higher recorded simulated hemorrhage-control success than the one-handed self-aid/restricted-hand pathway. The association remained similar after additional adjustment for sex, glove use, and limb location. Because pathway classification incorporated assistance context and other operational circumstances in addition to the number of hands used, the findings support pathway-specific training and competency assessment rather than a causal conclusion that two-handed application is mechanically or clinically superior.