August 8, 2017
We’ve been talking quite a bit about predictive analytics and Big Data at Cority over the past several months, and it’s not without good reason. More and more organisations are investing more time and money into understanding their data better, in many cases in order to employ predictive analytics. In the UK, this is no different. The UK Health and Safety Executive (HSE) are using predictive data analytics to improve their effectiveness in identifying which organisations they will audit or inspect. They are using their comprehensive historical data to build a model of future outcomes in their HSE Mind-IT tool. The tool was introduced in 2012 in response to two main drivers:
- The need to reduce numbers while increasing the effectiveness of proactive inspection
- To target those inspections to high risk areas
Dr Helen Balmforth, Head of Data Analytics at the HSE’s Health & Safety Laboratory, explained the functionality and rationale of the approach at the 2017 Health & Wellbeing at Work Expo in Birmingham. The tool accesses various databases, including the HSE Record of Prosecutions and Notices, the HSE Accident Database, and the Adverse Insurance Reports. (See the HSE rolled-up annual statistics on work related health and safety incidents – which estimates the resultant cost annual at over £14 billion) They are also currently exploring interfaces with other government departments such as HMRC to gain access to further information to make the most effective use of their inspectors’ time.
Predictive analytics is a hot topic in the world of EHSQ and the organisations with which we work in this area get great benefit from it. A feature of many organisations is that they have relatively mature EHSQ programs. Although this is true, most of these organisations do not have “clean” data; that is, they have lots of data, but it’s not organised or gathered in a useful way. In this case, some organisations are putting this data in “data lakes” where they can stay until it can be cleaned up and moved into a data warehouse. For a more in depth look at cleaning up “dirty” data, please refer to my colleague Jessica’s post from earlier this month.
For many other organisations there are simpler – and perhaps even more essential – first steps to take. Much effort is put into accurate tracking of safety incidents, (and for anyone still using Excel – we can help!) but that recording is only the first step. Have corrective actions been identified? Have they been assigned? Have they been closed out?
Preventing a recurrence of what has already gone wrong is the first step in making the workplace safer. With accurate information management, like Cority offers, you can confirm that these corrective actions have been carried out. If the corrective actions are overdue, automatic reminders can be sent to the assigned employee, and escalations automatically sent to their manager. Reports can be generated to show how effectively this is being done by site, business unit, country, or region.
One simple leading indicator could be to report on how effectively your team closes out corrective actions. This can be measured by metrics like how quickly and comprehensively they're closed out, especially the most serious ones. You can take this information and compare different sites or geographies to see how they're dealt with at each and make improvements from there. One of the biggest hurdles in this process is the standardization, or “cleanliness”, of your data. In order to benefit from any predictive analytics offering – including Cority’s – your data must be standardized so queries can be properly run in the system. If your data isn’t clean, you can’t make accurate forecasts.
Once you are doing this, we can move onto the more exciting world of predictive analytics which I'll write about in my next post in September.
About the AuthorMore Content by Howard Pullman