Machine Learning-Guided Training for Elite Athletes (MLGT)

NCT07683091 | Completed

• 2 other identifiers: DBU-SS-2023-008IRB#DBU-SS-2023-008
• Study Type: interventional
• Target: 120
• Locations: 1 country
• Sites: 2

Brief Summary

Plaintext The purpose of this study is to evaluate whether a personalized training protocol driven by machine learning can successfully reduce time-loss sports injuries and enhance athletic performance in elite athletes. During a 9-month competitive sports season, a group of elite athletes was divided into two training

Conditions

Athletic Injuries

Keywords

Machine Learning; Injury Prevention; Athletic Performance; Elite Athletes; Sports Biomechanics; Training Load Optimization

Outcome Measures

Primary Outcomes (1)

• Changes in Sprint Performance Time

  Sprint performance will be assessed using electronic timing gates to record running times over a specific distance from a stationary start. Lower times indicate improved sprint performance. Measurements will be taken at baseline and at the conclusion of the training intervention period to evaluate the impact of the workload protocols.

  12 weeks

Study Arms (2)

Control Cohort

ACTIVE COMPARATOR

Elite adolescent sprinters who followed standard, predetermined high-performance athletic training protocols typical for competitive season preparation. This group received structured training volume and intensity matching standard athletic coaching guidelines, without any machine learning interventions or adaptive workload adjustments.

Behavioral: Adaptive Machine Learning Workload Optimization

Algorithmic Cohort

EXPERIMENTAL

Elite adolescent sprinters who received a personalized training protocol dynamically optimized by a machine learning algorithm. The framework evaluated individual biomechanical variables, morning heart rate variability (HRV), sleep quality, and physiological fatigue metrics to adjust training volume and intensity.

Behavioral: Adaptive Machine Learning Workload Optimization

Interventions

Adaptive Machine Learning Workload Optimization BEHAVIORAL

A personalized, data-driven training intervention where athletic workloads are dynamically adjusted based on predictive modeling. The protocol continuously tracks individual physiological markers, biomechanical data, and workload history to optimize training volume and intensity. This adaptive approach aims to maximize performance gains while minimizing the risk of overtraining and injury during the competitive season.

Algorithmic Cohort; Control Cohort

Eligibility Criteria

• Age: 18 Years - 35 Years
• Healthy Volunteers: Yes
• Age Groups: Adult (18-64)

You may qualify if:

• Must be a competitive, elite-level or sub-elite track and field athlete specializing in short-to-mid distance running events.
• Aged between 18 and 35 years old.
• Actively participating in structured athletic training programs for at least 2 years prior to enrollment.
• Free from any acute musculoskeletal injuries or medical conditions that prevent full participation in high-intensity training protocols.
• Capable and willing to provide written informed consent to participate in the study.

You may not qualify if:

• \. Concurrent use of performance-enhancing drugs or medications that influence metabolic or cardiovascular responses.
• \. Inability to maintain consistent participation in the designated training protocols due to scheduling conflicts or travel.
• \. Any underlying cardiovascular, respiratory, or systemic condition that creates a health risk during exhaustive exercise testing.

Sponsors & Collaborators

• Debre Berhan University: lead

Study Sites (2)

Dr. Arefayne

Debre Berhan, Shewa, 445, Ethiopia

Location

M Dessye

Debre Berhan, Shewa, 445, Ethiopia

Location

MeSH Terms

Conditions

Athletic Injuries

Condition Hierarchy (Ancestors)

Wounds and Injuries

Study Officials

• Dr. Arefayne M Dessye, PhD

  Debre Berhan Univeristy

  PRINCIPAL INVESTIGATOR

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: PREVENTION
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Assistant Professor

Model Details: A parallel-group randomized controlled trial design was utilized to compare two distinct high-performance training tracks over a 9-month competitive season. Elite athletes were randomly allocated into either the experimental arm (undergoing dynamic, machine learning-driven training load and biomechanical optimization) or the active control arm (undergoing traditional, structured high-performance athletic preparation). Both groups trained concurrently under monitored conditions to isolate the effects of the technology-driven protocol on injury incidence and performance markers.

Study Record Dates

• First Submitted: June 27, 2026
• First Posted: July 6, 2026
• Study Start: January 1, 2023
• Primary Completion: September 30, 2023
• Study Completion: September 30, 2023
• Last Updated: July 6, 2026

Record last verified: 2026-06

Data Sharing

• IPD Sharing: Will not share

Locations

ET (2)