Absence management: Can new technology predict when employees are going to be off?
Robert Gray, June 25, 2010
There is no shortage of statistics on rates of staff absence. And systems are now being developed that not only record absence data but actually predict when an employee will next be off. The danger is that such information can be open to misinterpretation. Robert Gray reports.
Is absence the biggest barrier to Britain's economic recovery? If the figures are to believed, the 'case for' is certainly a convincing one. Despite reaching a 20-year low at the turn of the millennium, absence figures have since been on a relentless upward curve. Between 2005 and 2006 the number of days lost to absence nearly doubled from 78 million to 164 million; in 2007 it reached 172 million and the latest figures for 2008/9 (not available at the time of writing) are expected to show yet another unwelcome rise.
In employees' defence has been the fact that sick notes, or phone calls to line managers from ill-sounding staff, need to be taken at face value. But could absence management systems - the software that has historically recorded and number-crunched the incidents of absence - be emerging as both judge and jury? Increasingly systems are now being developed that not only record absence data but actually attempt to predict when staff, or groups of people, will next be off - in essence these microprocessors think they can spot foul play. This being so, is there either a danger that what is presented to HRDs is misleading or open to misinterpretation? Should HRDs be jumping to accuse if there is a pattern of absence that looks suspicious? Or could the real cause be far from what it appears?
Without doubt, absence data is a good and useful thing. But, as the old saying goes, there are lies, damned lies and statistics. Roger Moore, general manager of Bond TeamSpirit, believes the purpose of technology is to cement the HR professional into a role where they can be advisers and implement processes that help a business become more efficient. In capturing absence information the basic data input highlights that an individual was absent on a particular day but it's down to the HR professional to analyse that information as to the reason behind the absence.
"Predictive software is about assembling the business data and declaring that a certain pattern of data means a certain outcome," he says. "The fact of the matter is that HR professionals deal with human beings. He or she almost needs to take on the role of a psychologist to read and understand what is influencing an individual's behaviour. Software systems can give indicators that there is a problem but not the reasons why. This is best tackled with the experience and skills that an HR professional has acquired because the true answer may not be the problem that the predictive software suggested."
Part of the problem is that it is human nature to suspect regular 'illness' on a Friday or Monday and, if software spots this too, suspicions are naturally aroused. This is supported by research from Mercer in late 2009 into absence by day of the week. Based on data from 11,000 employee records, it found some interesting results: Monday absence was 35%; Tuesday was 23%; Wednesday was 21%; Thursday was 18% and Friday was 3%. While Friday is not yet a popular candidate for creating a long weekend, the research would appear to support suspicions of 'Monday-itis'. But the issue for employers is to separate the malingering from the genuinely ill.
According to the experts, sickness absence always follows a similar pattern depending on the season; any deviation from this should be acted upon immediately. They argue it is also important to use a rolling 12-month average when looking at absence levels, as monthly snapshots can fluctuate depending on the number of working days in the month and on events such as the World Cup (see p38).
The actual absence level calculation is also an area where managers could be misled. Absence is traditionally reported in one of two formats - as a percentage of working time available or as a 'days lost' figure.
"Both can be misleading if the calculation isn't comprehensive enough," says FirstCare's CEO, Aaron Ross. "Take the public sector: out of the three core services - local government, police and the NHS - local government tends to have the higher absence rates when looking at 'days lost' per employee. However, both the police and the NHS operate longer shift patterns. As a result, when you look at actual hours lost the police and NHS have the higher absence figures."
In local government, East Sussex County Council has been working hard in recent years to reduce the number of days lost to sickness. In 2003/04 it had 320 staff on long-term sickness (defined as four weeks or more), but by last year that number was down to 42. Strengthening its in-house health team, ensuring greater accountability with line management, engaging with the trade unions and providing free physiotherapy referrals as part of an 'invest to save' strategy were among the factors behind this major improvement.
Having made great strides with long-term absence, the council turned its attention to short-term absence last year. In October 2009 it launched a pilot scheme across three of its departments: Adult Social Care, Children's Services and Governance & Community Services. "The whole process is tracked and monitored very supportively," says Leathram Green, assistant director of training and personnel at East Sussex County Council. "We've had good feedback from trade union colleagues about how well our absence management product has worked and we've had little, if any, complaints from staff.
"The management information we get allows us to see patterns," adds Green. "For example, if lots of people from a particular team are off, is that telling us something? Or why is one person repeatedly taking the same day of the week off?" While he says it is important not to jump to any rash conclusions based on data alone, having more sophisticated systems in place does appear to be reducing abuse. In the first five months of the scheme, short-term absence has declined by 17%, he says.
Simon Macpherson, senior director, business development and operations at Kronos Systems, says it is vital to have timely back- to-work interviews as a way to reduce future absence. Systems like Kronos will alert HR within seconds of an employee returning to work following a period of sickness, allowing HR to conduct the return-to-work interview quickly and to work with the employee to understand the reason for absence.
"Patterns of absence can also be good indicators of problems," says Macpherson. "One customer had an employee who was habitually late and regularly absent. This was flagged up by the Kronos system. When HR investigated, the problem was that the employee was caring for a sick mother. As a result, HR was able to agree different work hours. Spotting patterns of absence can highlight peer-to-peer or manager/employee problems. Again, if data collection is automated and held centrally, HR can quickly spot problems and intervene."
Macpherson observes that more and more operational staff are becoming users of systems like Kronos and he feels HRDs should be encouraged to give access to line managers and operational teams.
Blending technology with sensitive human intervention is one area of growth. Such is the case with the automated WorkPoint system, where as part of key-pad notification, employees give a verbal (and recorded) reason for their absence. Nurses listen to all recordings and update the system with this reason. Call-back triggers can be bespoke to the needs of an employer - for instance, all Monday morning absences. WorkPoint's Peter Cutler says a list of medical and non-medical reasons are used to confirm the absence, which forms part of the regular data provided to clients. Nurses are trained to discuss reasons with the employee in a tactful and non-diagnostic way.
"Often, alternative (for example, physical manifestations of stress) or more detailed (for example, 'I am going through a divorce and I cannot concentrate on my work') reasons are highlighted," says Cutler. "We know that employees are more likely to discuss these 'subjective' causes with third parties, rather than their manager in the first instance."
For motor retailer Now Vauxhall, WorkPoint helped address a problem by comprehensively recording absence, examining which teams had consistently high rates and rewriting the company's sickness and unauthorised absence policy. It's absence rate is now 1.27% compared to the industry average of 3.8%.
There's no shortage of absence data available - the aim, as ever, lies in ensuring it is correctly interpreted and acted upon.
HOW SICK IS YOUR ORGANISATION?
For HRDs keen to assess how their organisation fares against the norm, employee risk and benefits management firm Aon Consulting has launched a free index. The European Sick Leave Index (ESLI) is based on data gathered from almost 200 employers with 370,000 workers across all main sectors.
It focuses on Belgium, France, Germany and the Netherlands, but employers from any country can benchmark against the index by visiting the company's website
The 370,000 employees tracked took roughly 4.2 million sick days over the past year. On average, each day cost employers almost EUR160. "ESLI demonstrates how combining some easy accessible data extracted from payroll gives more informative ratios on average duration, frequency, prevalence and impact on productivity," says Peter Abelskamp, Aon's executive director health and benefits EMEA "Comparing the ratios with the country or industry benchmark tells employers roughly how they are doing on management style, organisational culture and costs respectively."
SHOULD HRDS MEASURE SICKNESS 'PRESENCE' RATHER THAN ABSENCE?
A recent Work Foundation report sponsored by AXA PPP Healthcare, Why do Employees Come to Work when Ill?, studied 'sickness presence', defined as attending work when ill-health justifies taking time off. Among its findings were that sickness presence is significantly related to performance. Higher levels of sickness presence were associated with lower performance.
Moreover, sickness presence appears far more prevalent than sickness absence: 45% of employees reported one or more days of sickness presence compared with 18% reporting sickness absence over the same period. Crucially, the report concluded that employees with higher levels of sickness presence had significantly lower performance scores than those with lower levels of sickness presence.
This raises the question of whether employers should be measuring sickness presence instead of absence.
"Employers definitely need to be more aware that employees coming to work when their health justifies time off may be adversely affecting their health and wellbeing and performance," says Katherine Ashby, researcher, health and wellbeing, The Work Foundation, and co-author of the report. In her research, Ashby used a version of the Stanford presenteeism scale, which requires people to think about the last time they came to work unwell and answer a number of questions about how easy or difficult it was to carry out their role in those circumstances.
"It is important to recognise that self-perception is important," she says. "Some people with chronic health problems and disabilities may not define themselves as 'sickness present' if their health does not impact on their work ability."
OWN GOALS GOAL AAGOALS
In 2006, five million absence days were attributed to the World Cup. This year's tournament - yes, it's still on - is predicted to be responsible for nine million days off, as many matches start at 1.30 pm and continue throughout the rest of the afternoon. A survey by Kellogg's found only one in 20 staff would use their own holiday entitlement to watch daytime games.