Last year in Australia, 104,770 people were the subject of major injuries at work. One hundred and eighty-two people died in a work-related fatality¹. Workplace health and safety is serious business, yet so often it is treated as a reporting requirement: necessary, fulfilled, but ultimately reactive.
Reports on incidents that have happened tell a business how successful they’ve been at protecting their workers, but rarely much more.
What if it were possible to gain a better understanding of why incidents are occurring in the first place and actively employ strategies to prevent them? What if safety goals could be aligned to business goals? This is already possible with the help of a wider range of organisational datasets and advanced predictive analytics.
Traditionally, the reporting of workplace incidents has fallen to the Health, Safety and Environment (HSE) or HR functions. While organisations are getting better at prevention, for the most part, indicators of safety remain after the fact.
Anyone who has filled out an incident log knows the superficiality of the process. From this, a business can see what happened, to whom, and at what time. Viewed in totality, they may be able to see some patterns, but not always.
Lag indicators such as how many accidents have happened in the last month do little for keeping people alive in the future. Instead, assumptions and hypotheses are applied to suspected problems. Sometimes they’re correct, often they aren’t.
On the whole, businesses are not good at using the data sources they have available, and even worse when it comes to using sophisticated analytics methods, to gain insight. An evolution from simple to sophisticated data analytics use can be viewed as a maturity curve.
Analytics sophistication maturity
Businesses at the simplest end of sophistication merely report on safety incidents. The descriptive analytics applied to this data is simplistic but may reveal areas, teams, staff or equipment that are problematic.
At the next level, companies doing diagnostic analytics are merging multiple datasets. Typically, this is the initial incident data combined with other HR or payroll data. This can highlight correlations between employee behaviour (say absenteeism) and workplace accidents.
More sophisticated businesses use predictive analytics. Here, HR and incident data is augmented with operational, equipment or external data. This level of information, based on multiple large datasets, provides unique insights and actionable findings. Not only can these businesses see what happened during an accident, they can further understand the factors that contributed to the incident occurring in the first place.
Finally, there are organisations who are engaged in prescriptive, or optimisation, analytics. These data-savvy businesses benefit from predictive analytics insights and use them to optimise safety functions, making the best possible decisions based on resourcing, operational constraints and organisational goals.
In Australia, most companies are still at the first step of maturity. Imagine what a business could accomplish if it was using predictive and prescriptive analytics to action its safety efforts?
Origin Energy provides a great example of what the power of data analytics can do when it came to safety on its oil rigs. It engaged with PwC to understand how to leverage and analyse business and safety data to predict, prevent and report on potential increases and decreases in safety risk.
By using machine learning tools to analyse the sensor data from rigs and combining that with contractor data and weather information, we were able to highlight a previously unknown link between the depth of the wells and the chances of incident.
In fact, accidents were almost twice as likely in deeper wells. Being able to analyse this information further, which would not have been possible from simple incident reports, Origin could understand the root causes and improve safety to reduce the number of incidents occurring.
Similarly, our work with a another company into the causes of vehicle-related safety focused on more than just the number of car accidents. The organisation, which has thousands of cars on the road as part of its fleet, viewed safety incidents involving an employee or member of the public as one of its biggest enterprise-wide risks.
By analysing vehicle GPS data along with technicians’ in-cab ticketing dispatch system and other field workforce data, we were able to identify the risk factors for serious incidents. For instance, technicians interacting with their in-cab dispatch system while driving.
The benefit of such granular information is that the business can target its response. Rather than paying for a generic safety campaign or intensive training, costs could be significantly reduced by rolling out specific interventions to target specific high risk cohorts of employees.
For organisations that are looking to improve in this area, where to start can be confusing. Here are a few tips:
By taking a proactive approach to safety, companies can drastically reduce the number of accidents that occur or prevent incidents altogether.
Using data – not just health and safety data, but all the other diverse data sets that are often created across an organisation – combined with sophisticated machine learning analytics techniques, businesses can better understand where the risks are and roll out targeted interventions.
This approach will move safety efforts away from tick-box legislative requirements towards a business practice that drives real benefits and protects people’s lives.
For more information on improving safety performance and decision making visit PwC’s Predictive Safety Analytics.
© 2017 - 2021 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see www.pwc.com/structure for further details. Liability limited by a scheme approved under Professional Standards Legislation.