Related Experiment Videos
Modeling zero-inflated count series with application to occupational health
Kelvin K W Yau1, Andy H Lee, Philip J W Carrivick
1Department of Management Sciences, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China.
Computer Methods and Programs in Biomedicine
|March 3, 2004
Summary
A new zero-inflated Poisson mixed autoregression model effectively analyzes workplace injury count data with excess zeros. This statistical approach aids in evaluating the impact of participatory ergonomics interventions on reducing manual handling injuries.
Area of Science:
- Biostatistics
- Occupational Health
- Statistical Modeling
Background:
- Analyzing time series count data with excess zeros presents statistical challenges.
- Evaluating the effectiveness of workplace interventions requires robust analytical methods.
- Manual handling injuries in occupational settings often exhibit high zero-inflation in event counts.
Purpose of the Study:
- To introduce a novel statistical model for analyzing time series count data with excess zeros.
- To apply this model to evaluate a participatory ergonomics intervention aimed at reducing workplace injuries.
Main Methods:
- Development of a zero-inflated Poisson mixed autoregression model.
- Incorporation of random effects to address serial correlation in observational data.
- Parameter estimation using approximate residual maximum likelihood via log-likelihood maximization.
Main Results:
- The proposed model successfully accommodates excess zeros and serial correlation in count data.
- Demonstrated utility in evaluating population-level aggregated count data from occupational intervention studies.
Conclusions:
- The zero-inflated Poisson mixed autoregression model provides a powerful tool for analyzing injury count time series.
- This methodology enables effective evaluation of occupational health interventions using aggregated data.