• evidence based medicine
• workplace counselling
• workplace stress

Mental health problems at work are big news;^1–7 hardly a week passes without a new report highlighting stress at work or the proportion of absences from psychiatric causes. The numbers involved are enormous,⁸ to the point where one might be concerned that this was yet another public health issue.⁹ Both unions and employers perceive a problem, though it is by no means clear that they perceive the same problem. It is therefore perhaps unsurprising that they tend to advocate different solutions.

While occupational factors such a stress have demonstrable health outcomes,^10,11 workplace stress also has major economic implications. One report has suggested that it costs the UK £3 billion per year¹²—significant in relation to the £11 billion that the CBI calculates sickness absence as a whole costs the economy.¹³ Besides sick pay, replacement costs and lost productivity, an increasing proportion of this cost burden is comprised of compensation pay-outs. The TUC reported that there was a 12-fold increase in cases of stress at work; the value of claims is over £300 million.¹⁴

The Walker case, in which a social worker successfully sued his employers for psychiatric illness caused by his work, underlined the duty of care that employers have to their employees and confirmed that harsh financial penalties could result from failing in this duty of care.¹⁵ This issue received publicity recently with an Appeal Court ruling which, in redefining those circumstances under which an employee can claim for damages, may curb the numbers of such cases coming before the courts.¹⁶ In three out of four cases before them, the judges allowed the appeals, removing almost £200 000 from claimants who had originally been awarded compensation for stress related illnesses.

“Any alleged harm must be a demonstrable injury to health”

There is much within the guidance from the judges to applaud. They confirmed that psychiatric and physical illness should be addressed in the same way, and that there are no specific occupations which are intrinsically dangerous to mental health. Any alleged harm must be a demonstrable injury to health, not simply “occupational stress”, and must be attributable directly to work. This is especially important as stress at work is common and only infrequently damages health. They stated that employers are entitled to assume that employees can withstand normal pressures of work and to take what is said to them by employees at face value. Thus an employee returning to work after a period of sickness absence who does not disclose ongoing concerns is implying that he is fit to perform his previous duties.

More controversial was the guidance about employers’ responsibilities: “An employer who offers a confidential advice service with referral to appropriate counselling or treatment services is unlikely to be found in breach of duty”. Such a ruling may have dramatic implications, leading to an explosion of the provision of workplace counselling. Many employers and their staff currently favour such an intervention over more practical measures such as reducing hours worked or increasing staff numbers. However, the judges also stated that “an employer can only reasonably be expected to take steps which are likely to do some good: the court is likely to need expert evidence on this”. Where might one find such evidence?

Counselling in the workplace: the facts¹⁷ was published in January 2001 by The British Association for Counselling and Psychotherapy, and could be seen as a timely and convenient answer to this question. Its publication was reported in the British Medical Journal under the headline “Review confirms workplace counselling reduces stress”.¹⁸ It describes itself as “an independent critical scoping search of research into workplace counselling” and “the most comprehensive possible review of all English-language studies of counselling in the workplace”. It is certainly substantial, running to over 100 pages and has identified more than 80 pieces of work. Its review methodology is well presented and each piece of work is described and commented on using the language of evidence based medicine. The conclusions are clear and unequivocal. It claims that after counselling, work related symptoms return to normal in more than half of all clients, and sickness absence is reduced by over 25%; that workplace counselling is an effective treatment for anxiety, depression and substance misuse as well as “stress”. It is claimed that such results can be produced by as little as three sessions of any style of counselling as they all turn out to be effective. Was this the expert evidence that the judges referred to?

A more detailed reading of the report raises questions as to whether its conclusions can be justified. No studies were found that reported a negative outcome. This is surprising given the record of psychotherapy research.^19–21 There is no funnel plot and the term “publication bias” is not mentioned. Having proclaimed the value of “methodological pluralism” whereby qualitative studies were included, the report divides all those found into three categories: “best”, “supporting”, and “authenticating” evidence (there is no category for negative evidence). Serious doubts about the report’s claims emerge on examination of those 19 studies described as “best evidence”. Within the work that was published in peer reviewed journals there is a description of eight years of a “Peer Assistance Programme” for nurses with substance abuse problems and two reports each by employees of an Employee Assistance Programme simply describing their work. The most common format was a before and after study, typically comparing two different models of counselling. Three papers emerging from the Second Sheffield Psychotherapy Project describe many of the same patients. There are other examples of unrecognised duplicate publication; two pairs of papers comprising early and late reports of the same study are included. The report states that it found two randomised controlled trials (RCTs). In one,²² where a cluster randomisation was performed, it was felt to be unethical to withhold counselling from those who wanted it, and the method broke down. This is one of the few studies to show no benefit from counselling. The other paper,²³ while described as a randomised controlled trial (RCT) in the report, had no element of randomisation in it.

“The report is ambiguous about the role of randomised controlled trials”

Although apparently endorsing the principles of evidence based medicine, and frequently using the language of the genre, the report is ambiguous about the role of RCTs. It agrees that RCTs provide the most rigorous test of the efficacy of workplace counselling (page 21), but in other places questions that the RCT is indeed the gold standard of counselling research (page 11). It states that it is possible to overcome the absence of control groups by using assessment measures that have been developed for use in counselling research. Many would take issue with the assertion made in a number of the studies, and echoed in the report, that randomisation to a no-treatment arm is unethical. This is true only if the treatment has been shown to work, which has yet to be shown for workplace counselling.

That counselling or a psychotherapeutic intervention might do harm is ignored by the author, although there is ample evidence that this can be the case.^24–26 Indeed we have yet to encounter any treatments without side effects. With all treatments there is a balance to be struck between risks and benefits. Such adverse effects may include increased distress and dependency. This may occur by placing emphasis on the domain of the individual when the real problem is at the level of the organisation. Psychiatric disorders have many causes and falsely assuming problems that emerge in the workplace are due to work alone, may lead to other areas such as home and family problems being overlooked. At the milder end of the scale, distress is a normal, and often short lived, human emotion which does not necessarily benefit from being medicalised.²⁷

All evidence is useful, but any conclusions must be governed by the quality of that evidence. Counselling in the workplace: the facts suggests that for one reason or another the negative studies on workplace counselling have not come to light. We accept that good quality evidence can be difficult to obtain for some psychotherapeutic interventions,²⁸ but this does not reduce the need for RCTs in this field.

While there is much wisdom in the Appeal Court ruling, and it gives some firm guidance as to how both employers and employees should address the problem of workplace stress, its blanket approval of workplace counselling must be questioned. Simply implementing it will be seen as an insurance policy and lead to other important issues being conveniently ignored. The popularity of counselling, and the commercial muscle of many of the organisations that provide it, mean it will not disappear, but it should be refined and scrutinised closely to see any benefits it may bring. It should not be seen as a “quick fix” to a multifactorial problem. Employers and employees must work together to develop a range of responses to a complex problem.