For lab managers and safety professionals, the COVID-19 pandemic has been a masterclass in the complexities of personal protective equipment (PPE). From supply chain breakdowns to the delicate issues of fit testing, PPE has become the front line of workplace safety. But understanding the specific, evolving challenges facing workers across the country is a massive data task.
A new study from the National Institute for Occupational Safety and Health (NIOSH) suggests the answer lies in machine learning. Researchers Nora Payne and Emily Haas have developed a way to turn thousands of case reports into actionable intelligence by applying AI to Occupational Safety and Health Administration (OSHA) complaint data.
Integrating machine learning to monitor PPE compliance
Workers turn to OSHA when faced with safety risks. At the peak of the pandemic, these complaints surged, providing a raw, unfiltered picture of workers’ struggles. But there was a problem. The amount of data was too large for manual analysis to provide timely assistance. Using machine learning to monitor these concerns and track their changes over time is a cost-effective way to learn about the challenges facing workers.
Train a model based on real-world complaints
To bridge the gap between raw data and insights, the NIOSH team manually reviewed approximately 3,000 pandemic-related PPE complaints filed between January 2020 and July 2022. This manual review served as the basis for training a machine learning model, which was then tested for accuracy in detecting specific PPE-related complaints.
The study found that nearly 40 percent of pandemic-related OSHA complaints involved at least one PPE issue. Concerns range from physical discomfort and poor fit to lack of proper training, but machine learning models have proven particularly adept at identifying three key categories:
- Availability – Lack of required PPE
- Employer Mandation – Management does not mandate the use of PPE.
- Employee Compliance – Employee not wearing provided equipment
Enhance laboratory safety with real-time risk assessment
Perhaps the most important insight was how these concerns changed over time. In the summer of 2020, as sectors began to reopen, the data showed a clear shift. Workers’ concerns shifted from shortages to fighting crackdowns. As workplaces have become more crowded, workers have become more concerned about following rules than having equipment in the building.
lab management certificate
A laboratory management certificate is more than just training; it provides professional benefits.
Earn critical skills and IACET-certified CEUs that make a measurable difference.
This methodology provides a roadmap for future outbreaks for the laboratory community and public health agencies. By identifying these patterns in near real-time, lab managers can make more informed decisions about where to allocate resources and what specific guidance or interventions their teams need.
The NIOSH team is currently exploring ways to improve this approach. Their long-term goal is to develop a system to track PPE-related concerns in worker safety complaints as they occur. In an environment where the next infectious disease challenge may be on the horizon, the ability to listen and respond to workers through the lens of machine learning could be a life-saving innovation.
Data-driven strategies for laboratory leadership and safety culture
This study provides a framework for lab managers to move from reactive troubleshooting to proactive leadership. Analyzing OSHA complaint patterns can help managers decide where to prioritize their time and budget.
This data suggests that the high frequency of enforcement-related concerns often requires a shift in focus from procurement to leadership training and behavioral safety audits. These AI-driven insights can help executives differentiate between supply chain failures and organizational cultural issues. Armed with this data, managers can implement targeted interventions, such as sophisticated fit-testing protocols and peer-to-peer safety coaching, that address the root causes of employee risk and anxiety.
This article was created with the assistance of Generative AI and underwent editorial review before publication.
