Data-Driven Intensity Thresholds for External Load in Elite Women’s Handball: A Cluster-Based Approach Using Field-Based Data

Team handball is an intermittent sport characterized by variable kinematic and neuromuscular demands, which require precise monitoring to optimize training design. The aim of the present study was to establish data-driven intensity thresholds for external load variables and to examine their distribution across training tasks in elite women’s handball. External load was monitored in 17 professional female handball players during five training sessions within a competitive microcycle using inertial measurement units. Kinematic variables (speed, acceleration, and deceleration) and a neuromuscular variable (jump impulse) were analyzed. A k-means clustering approach was applied to classify each variable into five intensity zones. Subsequently, the distribution of these zones across different training tasks was evaluated. The results showed a predominance of low-to-moderate intensity actions across all the variables, with a progressive reduction in the frequency of higher-intensity efforts. Acceleration values were consistently higher than deceleration across all the zones. Jump impulse also followed a similar distribution, reflecting the neuromuscular demands of training. A task-based analysis revealed clear differences in intensity profiles, with tasks involving opposition eliciting higher proportions of moderate-to-high-intensity actions, while tasks without opposition showed an absence of high-intensity zones. These findings provide objective reference values for external load in elite women’s handball and highlight the importance of task design in modulating physical and neuromuscular demands. The use of data-driven thresholds and ecologically valid training tasks may contribute to more effective and individualized load prescription.