Exercise & Training

From Signal to Biological Meaning: A Systematic Review of Construct-Validity Evidence for External Training Load Metrics in Sport

TL;DR

Current evidence supports most external training load metrics as measures of external movement, work, or exposure rather than direct measures of biological load, and direct construct-matched evidence was uncommon.

Key Findings

Forty-six studies met inclusion criteria and each contributed one primary construct-evidence chain, but samples were small and mostly male.

  • Databases searched: MEDLINE, Embase, Web of Science, Scopus, and Academic Search Ultimate, from inception to March 2026 (final searches 1 March, Embase 7 March).
  • Main study samples ranged from 7 to 71 participants, with smaller analysis-specific subsamples.
  • Most studies included only men.
  • Eligible studies were peer-reviewed English reports evaluating an ETL indicator and its interpreted construct in athletes or physically active participants.

Most studies explicitly attributed a construct to the ETL metric, but a minority did so without qualification.

  • Construct attribution was explicit in 14 studies and explicit but qualified in 26.
  • Four studies were implicit but qualified and two were reverse-target.
  • General exercise intensity was the most frequent construct family (15 studies, 33%).

Direct construct-matched criterion evidence was uncommon; most studies relied on indirect routes.

  • Thirteen studies (28%) used a direct construct-matched criterion.
  • Alternative proxies supported 16 studies (35%).
  • Condition or group contrasts supported 11 studies (24%).
  • Mixed routes were used in six studies (13%).

Studies infrequently distinguished technical performance from construct interpretation or justified their choice of comparator.

  • Nineteen studies (41%) explicitly distinguished technical performance from construct interpretation.
  • Twelve studies (26%) explicitly justified comparator selection.

Critical appraisal identified common methodological limitations, particularly in hypothesis testing and analysis.

  • Measurement-property analyses were appraised with review-adapted COSMIN; prediction models with PROBAST+AI; other designs with JBI tools.
  • Heterogeneity and methodological limitations reduce confidence in biological interpretations.

Conceptual and methodological heterogeneity precluded meta-analysis, so synthesis was narrative with evidence maps.

  • Heterogeneity spanned metrics, target constructs, comparators, and study designs.
  • One page-anchored construct-evidence chain was adjudicated per study.

Most ETL metrics are supported as descriptions of external movement, work, or exposure, not as direct indicators of biological demand or response.

  • Many claims about metabolic, neuromuscular, skeletal, fatigue, or adaptation constructs relied on proxy comparisons, condition contrasts, or technical reliability evidence.
  • Biological interpretations require an explicit construct, plausible mechanism, directional hypothesis, and construct-matched reference.

The review recommends that metric selection begin with the monitoring decision, followed by explicit construct-validity support.

  • The proposed interpretation should be supported by an explicit construct, mechanism, validity hypothesis, and appropriately matched reference evidence.

What This Means

Sports teams use wearable devices and tracking systems to measure things like distance run, speed, accelerations, and impacts. These are called external training load metrics. They are often interpreted as telling us how hard the body is working, how tired an athlete is, or how they are adapting. This review asked whether research actually backs up those biological interpretations. The authors reviewed 46 studies, mostly small and mostly in men. Only about 28% (13 studies) compared the metric against a measure that directly matched the biological thing it was claimed to reflect. Most others relied on stand-in measures, comparisons between conditions or groups, or evidence that the device was simply reliable. Fewer than half clearly separated whether a device measures accurately from whether it means what people say it means, and about a quarter explained why they chose their comparison measure. Many studies also had weaknesses in how hypotheses were tested and analyzed. This research suggests that external load metrics are best viewed as descriptions of movement, work, or exposure, not direct measures of metabolic, muscular, bone, fatigue, or adaptation responses. The authors argue that anyone using these metrics should start with the decision they want to inform, then check that there is a clear construct, a plausible mechanism, and a well-matched reference supporting the interpretation. Because the studies were small and varied, confidence in biological interpretations is limited.

Check Your Own Numbers

Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.

Upload Your Labs

Have a question about this study?

Citation

Hou-Chen Li, Qing-Qiong Yang, Bin Wang. (2026). From Signal to Biological Meaning: A Systematic Review of Construct-Validity Evidence for External Training Load Metrics in Sport. Sports Medicine - Open. https://doi.org/10.1186/s40798-026-01133-7