How can Rich Data from Wearables Improve our Understanding of Fatigue?
The impacts of fatigue and the case for wearables
The consequences of fatigue during work are not to be underestimated. This is particularly true in safety-critical industries, where workers often have irregular shift patterns or are required to focus for long periods of time. The Incident Factor Classification System developed by RSSB as part of the Safety Management Information System (SMIS) identified fatigue as a key factor in 21% of high risk incidents between 2011-2013. Between 2001 and 2009, fatigue was considered a possible causal or contributory factor in at least 74 RAIB investigation reports into railway accidents and incidents (RAIB, East Somerset Junction report 2009).
Furthermore, fatigue can be costly to employers beyond the risk of accidents, incidents and ill-health, in that it makes expensive mistakes more likely, reduces productivity and morale, and increases absenteeism. It has been estimated that the cost of insufficient sleep to the UK economy is £40bn per year (RAND, 2016).
But wearables can provide a wealth of real-time data and provide insights that may have been more difficult for workers, managers and Safety, Health and Wellbeing (SH&WB) specialists to arrive at otherwise.
To reduce the risk of accidents, employers are increasingly considering resorting to the use of dedicated wearables that can detect heart rate, eye movement, temperature, and blood pressure in real-time and send alerts.
Types of devices and trials
There are different types of monitoring devices available on the market that can contribute to the understanding of workers’ fatigue. Devices such as smart watches and fitness trackers enable workers to be proactive in managing their fatigue and prioritise their Health and Wellbeing. Some wearables that monitor activity and sleep can be worn both on and off-shift, while other devices monitor subtle cues in worker behaviour and compare them to machine learning algorithms to identify a drop in alertness, or drowsiness, encouraging workers to take a break from their task prior to experiencing a microsleep during work.
Devices and technologies of this kind vary in their maturity, and manufacturers may be tempted to overstate the risk control benefits when marketing equipment, so it is important for potential users to establish what evidence suppliers are relying on when making claims about the effectiveness of their systems. Several trials have been undertaken that illustrate how the technology can be implemented and the benefits that could be gained.
Drowsiness can be monitored through the use of eyewear. For example, Optalert (pictured left) provides lightweight drowsiness detection glasses, which work by using an LED built into the frame to measure the operator’s eyelid movement. The data is analysed and translated into a score measured on the Johns Drowsiness Scale (JDS), which is presented to the driver on their indicator or processor situated within the cab.
SmartCap (pictured right) offers a non-invasive piece of headwear fitted with a removable sensor to monitor drowsiness, catering for drivers and operators of heavy vehicles. The device is either worn on its own or is attached to the driver’s choice of headwear and measures electrical brain activity (EEG) in real-time, which is then processed to understand how alert the person is likely to be. As well as presenting an indicative “fatigue level” on an app for workers, centralised monitoring allows supervisors to remotely monitor their fleet online.
A Fujitsu collar (pictured left) with an ear clip that identifies episodes of drowsiness by measuring vital signs such as pulse rate has been piloted for eight weeks by Amey. The technology was worn by each member of staff working on Highway England’s North East Regional Maintenance Contract, who were required to drive vehicles during their shift. The collar integrates a vibration mechanism to inform the driver when they are losing alertness and also uses an automatic calibration and learning function to adapt and compensate for differences between individuals’ baseline metrics.
Two Crossrail projects trialled wearable technology in an attempt to evaluate sleep-related fatigue risk within the workforce. Around 120 workers received wristbands which continuously collected their sleep/wake data. This information was then analysed using a biomathematical model to calculate an “Alertness Score” – a prediction of a worker’s cognitive effectiveness. The score was plotted on a scale ranging from 0-100, which was put in context mapping scores onto equivalent blood alcohol impairment and change in reaction time. The scale was made accessible to shift workers, managers and SH&WB specialists alike. Overall, the trial established when alertness would be likely to peak during a typical shift and assessed the average sleep duration of the workers off-shift. This data-gathering exercise allowed schedule patterns to be adjusted where necessary to maintain a good level of alertness throughout shifts and offset the risk of fatigue related to sleep. The project recommended to encourage workers to use the mobile app for effective self-management, to consider fatigue risk management in the earliest phases of schedule design, and to consider equipping the entire workforce with the technology for detection of chronic sleep issues and possible sleep disorders that could affect performance in the work environment.
Wearable devices are a welcome addition, not substitutes
Although the above examples are early attempts at implementations and vary in maturity, wearable-based systems bring data where none existed before, and some welcome opportunities to improve our understanding of fatigue risk. Of particular interest is the value of wearables for learning about organisations’ exposure to work activities that can lead to fatigue, the relationship between work and sleep, the impact over time of sleep patterns, job design etc.
However, wearables should not be considered as substitutes for fully integrated Fatigue Risk Management Systems (FRMS). It is important to understand that wearables do not provide a direct measurement of fatigue, but instead make assumptions and calculations to infer drowsiness, alertness or other related states. They do not necessarily pick up on the poor decisions that people can make and the delayed reactions that people can display well before they start to fall asleep.
There are also possibilities of false positives and false negatives, which means that when making decisions that could affect safety, we’d rather not rely solely on one device, but assess multiple metrics in parallel to increase the validity and reliability of the results. Careful implementation of new technologies, taking proper account of their limitations, is always vital.
The devices should be regarded as what they are, i.e. a very useful source of additional indicators to inform safe decisions, and good complements to existing fatigue performance indicators and risk controls. Wearables do not replace the need for risk assessment, education, good scheduling, well-functioning feedback and reporting systems or any other components of a FRMS but could provide data to inform and improve these measures.
There are likely to be issues around perceived intrusion and user acceptance which will need thoughtful, sensitive handling.
Additionally, it should not be forgotten or underestimated that wearables may lead to changes in behaviour, which could have their own safety consequences. If the industry becomes over-reliant on them, especially in safety critical contexts, situations may arise where workers go to work when they shouldn’t, e.g. when they are extremely tired but trust the system to stop or alert them in time. Equally managers under pressure might ask workers to work longer hours under the (potentially) false assumption that wearables would eliminate any additional risk.
To conclude, we believe that wearable technology has a key part to play in the future of understanding and managing fatigue. Recent trials have highlighted benefits of some of the technologies available. Wearables that are currently on the market, with the abundance of data that they can bring, can inform how we manage fatigue today. Robust analysis of this data can provide SH&WB specialists with insight into workers’ exposure to fatigue or fatiguing conditions, and their SHW in general. Workers can track their sleep, alertness or drowsiness and SH&WB specialists can provide specific and actionable advice where necessary. Moreover, aggregating the data across the workforce can help minimise risk in a broader sense and improve productivity.
As technology develops, we are optimistic that there will be an even stronger case for the rail industry to adopt devices that can bring new data and greater insight, and that wearables, if intelligently implemented, should be welcomed by the industry. From headwear to wristbands, a variety of technologies have been trialled in safety-critical contexts, and sharing key findings can help to direct the development wearables in a way that helps to deliver a safer work environment within the rail industry, where safety, Health and Wellbeing are top priorities. To take forward the use of fatigue monitoring wearables across the railway, the joint expertise of those developing and supplying wearables working together with stakeholders across rail will help the industry to fully reap the benefits of this latest technology.