Zero-inflated Poisson regression with random effects to evaluate an occupational injury prevention programme

Area of Science:

• Occupational Health
• Biostatistics
• Public Health

Background:

• Hospitals experience high rates of compensable injuries, with over 60% from manual handling.
• Manual handling injuries pose a significant risk in healthcare settings, particularly for support staff.
• Previous injury rates necessitated the development of targeted prevention strategies.

Purpose of the Study:

• To evaluate the effectiveness of a manual handling injury prevention strategy.
• To analyze longitudinal injury count data with an excess of zero occurrences.
• To assess the impact of the Workplace Risk Assessment Teams (WRATS) program on injury rates.

Main Methods:

• Utilized a zero-inflated random effects Poisson regression model.
• Applied a workplace risk identification, assessment, and control approach.
• Incorporated WRATS intervention and individual cleaner data as fixed effects.

Main Results:

• The WRATS program provided statistical evidence of its value in reducing injuries.
• The zero-inflated model effectively analyzed longitudinal count data with excess zeros.
• The intervention demonstrated a positive impact on manual handling injury reduction.

Conclusions:

• The WRATS program is an effective strategy for reducing manual handling injuries in hospital settings.
• The developed statistical methods can be applied to evaluate other occupational injury prevention trials.
• Findings support the implementation of similar risk management strategies in high-risk populations.