MSK injury risk score — predictive, not reactive
Built on 19,019 real Australian complaint episodes. Not modelled data, not academic estimates — actual clinical outcomes from 20+ years of occupational health treatment.
Get your organisation's risk profile97%
Full-duties rate in managed cohorts
19,019
Validation population (episodes)
20+ yr
Clinical dataset depth
50%
Recurrence rate — Poor PES workers
What does the risk score tell you?
Each worker and cohort is scored into one of four bands. The band determines the recommended action — no clinical judgment required from your WHS team.
Routine monitoring. No immediate intervention required. Flag for quarterly review.
EAP referral consideration, ergonomic workstation review, physio check-in scheduling.
Proactive outreach within 5 business days. Early intervention consultation recommended.
Immediate case management. High recurrence probability — do not wait for worker to self-report.
How is it calculated?
Six clinical inputs. Population-level validation. Weekly retraining from new encounter data.
Lumbar, cervical, shoulder, knee, wrist/hand — each has a distinct recurrence profile
Repetitive strain vs acute trauma vs postural load — mechanism drives care pathway
Early intervention, standard care, or rehabilitation — timing relative to onset matters
Visit frequency adjusted for worker age cohort — controls for age-related utilisation bias
Previous claim history is the strongest single predictor of future recurrence
Week-on-week visit frequency — rising trend is an early recurrence signal
The model retrains weekly as new clinical encounter data enters the pipeline. Scores reflect population outcomes, not individual clinical judgment.
Predictive Engagement Score (PES) validity
Workers with Poor Predictive Engagement Scores show a 50% injury recurrence rate — Atlas MSK surfaces this pattern automatically from your cohort data, without requiring clinician review of individual records.
How is this different from a subjective risk assessment?
Subjective assessments are fast and clinician-led — why change them?
Subjective assessments rely on worker self-report at a single point in time. Workers under-report when they fear job security implications. Atlas MSK's risk score is derived from 19,019 population-level clinical outcomes — it catches patterns that precede worker disclosure, not after.
Our EHR system already tracks injury data — isn't that enough?
EHR systems record what happened. Atlas MSK predicts what will happen. The difference is 20 years of longitudinal outcomes data — knowing that a 45-year-old lumbar presentation with two prior claims in the past 36 months has a 3.4× higher recurrence probability than the cohort average is not something a clinical record alone can tell you.
We don't have time for another system.
Atlas MSK surfaces risk scores as a dashboard view — no data entry required by your team. The risk engine consumes data from your existing occupational health provider and returns scored outputs. Your WHS manager sees the risk profile, not the algorithm.
What the 97% full-duties rate means
97% of workers in WHA-managed cohorts return to full unrestricted duties — not modified duties, not reduced hours, full duties. This compares against a national RTW rate that includes significant proportions on ongoing modified duties or not returned at all.
The difference is early intervention timing. Atlas MSK's risk score identifies high-risk workers before they breach the threshold where modified-duties arrangements become necessary. The 18.3-day average recovery in WHA-managed cohorts vs industry averages reflects the same intervention timing advantage.
Dr James L Murray
Founder, Work Healthy Australia
20+ years in Australian occupational health. Developer of the Atlas MSK risk engine. Author of the WHA treatment dataset methodology covering 182,000+ de-identified occupational health episodes. Research partner on the Bond University macroeconomic MSK elasticity study (April 2025).
Get your organisation's risk profile
See how your workforce cohort scores against the 19,019-episode baseline. A 30-minute demo, no setup required.
Book a demoRelated