I spy with my little eye ... a knee about to go 'pop'? Can coaches and sports medicine professionals predict who is at greater risk of ACL rupture?

Key points

1. 01237 assessors rated 102 video clips of elite female handball and football players doing a vertical drop jump test, 20 of whom later tore their ACL.
2. 02Individual accuracy (AUC) ranged from 0.36 to 0.60 with a mean of 0.47, essentially random guessing.
3. 03No occupational group (physicians, coaches, athletic trainers, researchers, physiotherapists) did meaningfully better than any other.
4. 04Assessors agreed with each other (high inter-rater reliability) and felt moderately confident, yet were still wrong, a sign of poor judgement calibration.
5. 05The authors conclude the drop jump test should not be used to identify elite female athletes at risk of ACL injury.

How it was conducted

Design

Diagnostic accuracy study using video clips drawn from a 738-player prospective cohort

Participants

237 assessors (physicians, coaches, athletic trainers, strength coaches, physiotherapists, researchers, students) rating 102 player video clips

Player sample

102 elite female handball and football players, 20 of whom sustained an ACL injury during follow-up

Task

Watch a 2-3 second frontal-plane vertical drop jump clip and rate ACL injury risk from 1 (low) to 10 (high)

Primary outcome

Classification accuracy measured by area under the ROC curve (AUC)

Analysis

ROC/AUC with 95% CI, one-way ANOVA with Bonferroni correction, independent t-tests, latent variable modelling for reliability

What they found

• Individual AUC values ranged from 0.36 to 0.60, mean 0.47, similar to random guessing.
• Group mean AUC ranged from 0.45 to 0.47 with no significant difference between occupational groups (p=0.67).
• Mean risk rating of injured players was 4.5/10 (SD 1.7) versus 4.8/10 (SD 1.7) for non-injured players (p=0.52, d=0.16).
• Sensitivity analysis using the natural injured-to-non-injured ratio gave mean AUC 0.43 (95% CI 0.40 to 0.47), no different from the main result.
• Inter-rater reliability was high (r=0.75), and assessors reported a mean confidence of 6.0/10 (SD 1.9).
• Most-used cues were knee position in landing (99.6% of assessors, importance 9.0), inward/outward knee motion (99.2%, importance 8.8) and landing symmetry (98.7%, importance 8.1).

Limitations

• Players were assessed from a frontal-plane view only, so other planes of movement could not be judged.
• Picture quality and zoom factor differed slightly between test sessions.
• The drop jump is preplanned and loads both legs evenly, unlike the cutting and single-leg landings where ACL injuries typically occur, so it may not challenge elite players enough.
• Some players counted as uninjured may have torn their ACL after the observation period ended.

Why it matters

For patients

A clinician eyeballing your landing technique cannot reliably tell whether you will tear your ACL, so do not read a clean or sloppy drop jump as a verdict on your risk.

For clinicians

Do not use visual assessment of a vertical drop jump to screen elite female football and handball players for ACL injury risk, and be wary of feeling confident in such judgements.

For readers

Expert agreement and confidence do not guarantee accuracy, this is a clear example of a widely used screening test failing to predict the outcome it is meant to flag.