Article Text


Mental ill health and fitness for work
  1. Nick Glozier
  1. Correspondence to:
 Dr Nick Glozier, Epidemiology/WHO Collaborating Centre, Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK;

Statistics from

Mental ill health at work seems to be rising inexorably both in terms of self report (as shown by comparing surveys of work related illnesses throughout the 1990s), and as a cause of absenteeism, long term sickness, and early retirement. It is a burgeoning field for many professionals; human resource, medical, psychology, and, with the advent of anti-disability discrimination legislation, lawyers. “Fitness for work” is considered in two main areas: recruitment of new staff, and return to work of those who have been off sick. In many cases the issues are the same for employers, human resource, and occupational health staff: Is the person able to perform the job adequately? Do they pose health and safety risks? How likely are they to require future sick leave? The introduction of anti-discrimination legislation in several countries has also introduced the requirement to consider workplace adjustments that “enable” the disabled. Mental ill health poses particular problems when addressing these issues, as the impairments associated with such illness are difficult to distinguish from the cognitive and behavioural performance intrinsic to many jobs.

Summarising the effects of mental ill health on work (and vice versa) is complicated by the wide spectrum of “mental ill health” which tends to be covered by three different streams of research. Firstly, there is a literature, drawing predominantly from the fields of occupational psychology and health, examining the differing and interacting effects of workplace stressors and “stress” (or what will be termed here “common mental disorder”), performance, health, and absenteeism. Secondly, there are epidemiological studies demonstrating associations between functioning, often as “disability”, and psychiatric disorders. Finally, stemming from psychiatric and rehabilitation research is a body of work investigating the area of supported employment, training, and education for those people with serious mental illnesses, generally psychoses. There are enormous differences in the prevalence in the working age population of these different levels of “mental ill health” as shown in fig 1.

Figure 1

Relative prevalence of self reported stress, and mental illness in the UK working age population. Source: Office of National Statistics. Psychiatric morbidity amongst adults living in private households, 2000; summary report. London: ONS, 2000)

Overall, mental ill health is the second largest cause of work related problems after musculoskeletal disorders. In the UK it accounts for one third of all work related illness,1 is the second major cause of long term absences in the Whitehall studies,2 and is responsible for 20% of all early retirements from the National Health Service.3 Given the stigma attached to mental ill health this is probably under reported. Despite this, there seems a lack of ideas over what to do to ameliorate these problems.


Common mental disorder

This encompasses those who are considered as cases by such measures as the General Health Questionnaire,4 the most commonly used measure of “stress” or “mental ill health” in occupational studies, and those with the minor, and usually mixed, anxiety and depression often seen in primary care. At any one time some 20–35% of the working population fall into the former category5 and approximately 10% the latter. Even this level of mental ill health has adverse effects upon externally rated work performance,6 disability days,7 and absenteeism.2 Sickness absence has a multifactorial nature. Within the workplace those with common mental disorders are approximately 1.5 times more likely to later suffer from musculoskeletal problems, such as low back pain,8 radiating neck pain (with dose response at increasing levels of stress),9 and work related upper limb disorders,10 and associated absences.

Depression and anxiety disorders

These are the common, but more discrete, conditions that are associated with more obvious individual impairment and symptoms. The UK survey of psychiatric morbidity estimated that at any one time some 5% of the population would have an anxiety disorder, and 2% be suffering a depressive episode.11 Lifetime prevalence is far higher, with US estimates of 17% for a major depressive episode and around 20% for an anxiety disorder not including phobias.12 These people are far less likely to be employed than the general population, and even those that are report twice the amount of sick leave than the general population,13 an effect that is possibly more pronounced in men.14 When at work, those with depression in particular report a much greater effort required to function.15

Studies following clinical samples of those affected paint a more detailed picture. A review by Simon et al16 of the different study types showed that, when well, these people had no occupational impairment. This is backed up by a long term follow up of those with major depression17 which also highlighted a gradient of psychosocial impairment with illness severity—that is, there is no “all or nothing” effect. After the depressive episode has been treated there is often some temporary residual work impairment, presumably reflecting subthreshold depressive symptoms and a shift into the common mental disorder group. These disorders should be treated as chronic conditions, like diabetes or asthma, as approximately 8 out of 10 will experience some level of recurrence of a depressive episode.

Substance abuse

Community surveys from both the USA and Canada have shown that people with substance abuse problems experience many more disability days than the rest of the population.12,15

The range of ill health associated with alcohol ranges from coming to work hungover, through drinking at work, to alcoholism. There is an ongoing controversy about the degree to which alcohol problems impact employment. Clinical samples of alcohol abusers seemed to suggest that alcohol misuse does affect employment, but community samples show minimal effect of alcohol problems over and above any associated psychiatric illness. Sophisticated economic analyses18 allowing for labour supply relationships, also suggest that any relationship found may not be as strong as originally thought.19


The uncommon conditions of schizophrenia and manic depression affect approximately 0.5–1% of the population at any one time.12 These are among the most discriminated against of all disabled people. In the UK only 60% of people with psychoses are “economically active”, yet only 20% consider themselves unfit for work.13 The level of unemployment varies considerably across developed countries, reflecting labour conditions, occupational differences, and stigma. Their rarity in employment precludes their inclusion in community surveys of occupational dysfunction. Symptoms of schizophrenia are hallucinations, delusions, disorganised thinking, and a general lack of social interaction (negative symptoms). When ill it is the last two groups of symptoms that affect employment rather than the more obvious one of hearing voices, etc. Manic depression can be particularly debilitating as the individual is usually very well in between episodes yet most suffer an episode, which can last several months every couple of years.20 Manic symptoms of irritability and grandiosity can be especially disruptive.

In summary virtually all levels of current mental ill health have deleterious effects upon functioning. Comorbidity is a serious issue. A recent US study has shown that combinations of mental ill health, substance misuse, and chronic physical illness produce more disability days than would be predicted simply by adding their component effects.21 However, when well those with a history of mental ill health could be expected to perform similarly to the rest of the workforce and others who might be disabled.



One employer response to this obvious occupational impairment is to screen at interview. However, in some countries this is currently outlawed, such as in the USA by the Americans with Disabilities Act (ADA). It is still allowable in the UK, and Poole22 has suggested a hierarchy for assessing risk of future absenteeism for both physical and mental ill health. This emphasises hospitalisations and recurrence for mental disorders. Although understandable, if used to assess the potential for individual support and workplace adjustment, then it might be sustained. If used to discriminate in the hiring process then this strategy may backfire.

Screening for common mental disorder is probably pointless. Half of those who are positive will change their status over the next year, the sheer numbers involved would overwhelm any occupational health service, and the predictive value is low. Screening for alcohol misuse may also prove a false hope. Firstly, in assessing problem drinking, occupational health workers need to be aware of the relative performances and limitations of different screening instruments—for example, CAGE and BMAST, which give a different prevalence.23 Secondly, recent data have shown that it is not average daily intake that is associated with work performance problems, but drinking at work and alcohol dependency.24

For those with a history of depression or anxiety, there is no significant research on the variables that might predict successful employment25 beyond comorbidity. Effective maintenance treatment, both psychotherapeutic and pharmacological, is now thought to be required to reduce the relapse rate in those with a previous episode of depression, but predicting which individuals will relapse is unreliable. Furthermore, issues of confidentiality may prevent an employer from finding this information out.

When looking at reintegrating those who have had psychoses into the workforce, a number of factors are considered to be associated with good occupational outcomes. The question facing most employers would be the fitness to work of someone after a first episode of illness. From naturalistic studies of patients admitted for schizophrenia, comes evidence that those with predominantly negative symptoms are less likely to be employed.26 This is not surprising as these symptoms undermine many of the attributes required to function in the workplace and can be present after the more obvious symptoms have responded to treatment. After a first episode of mania, those with later onsets and better premorbid employment and education recover their occupational functioning better. Symptoms of the illness have little bearing.27

The success of individuals moving from vocational rehabilitation schemes to competitive employment is dependent upon better overall functioning, and good employment histories and interest in employment, which can be enhanced with minimal vocational training.28 Individual supported placements with job coaches in the workplace seems to be the way forward to successful competitive employment rather than prevocational training.29 All decent studies have been done in the USA. The applicability to other countries with very different economic and benefit systems is debatable, although a similar programme at an NHS trust in south west London30 has had some promising results.

For people with these psychoses the research is still one of chicken or egg: do those with better premorbid functioning have more chance to develop the skills that make them employable, or do employers look more favourably on better work histories?

Certain occupations may be amenable to screening. The recent claims for psychiatric illness arising from exposure to “traumatic stress” in the emergency services, and knowledge of vulnerability factors may potentially influence selection and early training.31 The legal ramifications of this are unclear at present

Health and safety issues

One of the primary reasons behind the (often covert) discrimination of those with mental ill health is concern about health and safety. The public generally view those with mental illness as unpredictable and potentially dangerous.32 While those with current psychoses may have a slightly elevated risk of violent behaviour compared to the general public, particularly if associated with substance abuse, this is less than that posed by young men and substance abusers in general.33 Those with anxiety, depression, and common mental disorder have no greater risk of violent behaviour. More subtle issues such as conduct at work, which is explicitly dealt with in the ADA, and cooperating with co-workers, are more likely to be affected by an individual’s personality than their illness, and can be handled as usual. Personality disorders, which can seriously disrupt occupational functioning, may potentially be covered by workplace legislation but there are as yet no cases, certainly in the UK.

The effects of drugs and alcohol can be severe. Firstly, most medication taken for mental ill health can induce drowsiness and poor concentration, particularly at the beginning of treatment. Machinery and computer operation and driving can all be affected. However the newer antipsychotics, and antidepressants such as the selective serotonin reuptake inhibitors (SSRIs), have less of a sedating effect.

All street drugs affect cognitive and day to day function at some level. However, gathering data on non-clinical and occupational groups for many of these is difficult because of disclosure. Marijuana use has a greater association with work performance problems including punctuality and attendance, poor work quality and quantity, and arguments with co-workers than screening positive for alcohol problems.24

It will come as no surprise that alcohol is thought to be associated with safety issues. In an Australian study34 problem drinkers were 2.7 times more likely to take injury related absences, although overall high alcohol consumption was not found to be related to work injuries. In Spain rising drinking levels were associated with increased odds of having an industrial accident,35 although this may be confounded by an association of heavy drinking in hazardous occupations. Recent studies have examined random samples of workers and emerged with more surprising results. Coming to work hungover and drinking at, or just before, work were greater predictors of such work problems as sleeping on the job or problems with co-workers and supervisors than overall heavy drinking.36 The Amtrak corporation in the USA realised the safety implications of alcohol and took an interesting approach by involving the unions as a co-worker safety issue. This has lead to a dramatic fall in alcohol related incidents through the RedBlock early intervention programme (fig 2).

Figure 2

Poster proclaiming the RedBlock programme devised by the Amtrack corporation in the USA to foster an alcohol-free work environment.

Legal issues

Legal issues obviously vary from country to country but those in the UK, and lessons to be learned from the USA, will be concentrated upon here

Anti-discrimination legislation

Discrimination in the workplace exists, as was experienced by nearly half of respondents to a UK survey,37 and revealed by employer attitudes.38 Employers are wary of the disabled but this is particularly acute for those with mental ill health. In the US 1994–95 National Health Interview Survey, 1.1% of respondents reported a functional disability associated with a mental condition and, of these, one fifth reported examples of discrimination at work.39 Anti-discrimination legislation is fundamentally rights legislation, so it is understandable that the USA was the first to introduce it for disability. In the UK the Disability Discrimination Act was introduced in 1995. In assessing mental ill health, a major difference between the two Acts is in the specific exclusion in the UK of alcohol dependence. Both exclude conditions associated with “badness” rather than “sickness” (boxes 1 and 2). A fundamental point for employers and those associated with them is that it is not the diagnosis but the associated impairment that must be examined. Assuming that everyone with a psychiatric diagnosis will be covered is probably fallacious.

Box 1: UK Disability Discrimination Act 1995 definition of disability

“a physical or mental impairment which has a substantial and long term adverse effect on a person’s ability to carry out normal day to day activities”.

All five criteria within this definition must be met. People who have had a disability for a period of time but no longer have one are included

Normal day to day activities can include sensory abilities of concentration, learning, understanding and perceiving danger A long term adverse effect is one which is detrimental and has lasted or is expected to last 12 months


  • ▸ detrimental effect of medication

  • ▸ addiction to alcohol, nicotine, or any other substance (unless resulting from medical prescription)

  • ▸ a tendency to set fires, steal, physically or sexually abuse others, and exhibitionism or voyeurism

Box 2: US Americans with Disabilities Act 1992 definition of disability

  • A physical or mental impairment that substantially limits one or more of the major life activities of such individual

  • A record of such an impairment or

  • Being regarded as having such an impairment

  • An impairment means any mental or psychological disorder, such as mental retardation, organic brain syndrome, emotional or mental illness, and specific learning disabilities

    Major life activities include such things as caring for one’s self, performing manual tasks, walking, seeing, hearing, speaking, breathing, learning, and working


    • ▸ sexual behaviour disorders: transvestism, transsexualism, paedophilia, exhibitionism, voyeurism, gender identity disorders not resulting from physical impairments,

    • ▸ compulsive gambling, kleptomania, pyromania

    • ▸ psychoactive substance use disorders resulting from current illegal use of drugs

    In the UK it is unlawful for an employer to treat a disabled person less favourably than someone else because of their disability unless there is a good reason. This includes matters such as recruitment, training, promotion and dismissal, for temporary, contract, and permanent staff. It does not prevent employers from enquiring into the health conditions of a prospective employee. Employers have a duty to make reasonable adjustments in the workplace to overcome the effects of the disability. If it is felt that there has been unfair discrimination or a failure to make reasonable adjustments the (potential) employee makes an application, within three months, to an industrial tribunal. Employers with fewer than 15 employees and certain uniformed occupations are currently exempt from the Act, although this is under review.

    In the USA the situation for mental ill health has been summarised by Mechanic.40 He has highlighted the tensions for employers and individuals. For employers and occupational health staff the lack of observable limitations can lead to some suspicion of gaming the system. Knowing what constitutes a “reasonable” adjustment can be very difficult. For employees, disclosure becomes an important issue. Adjustments can only be made if the employer is aware of a disability yet discrimination is very hard to prove. The formation of the Equal Employment Opportunities Commission in the USA has enabled accurate data to be derived on the effect of the ADA. It makes fairly heartening reading for employers. Although mental ill health has overtaken back impairments as the most cited disability, of 175 000 charges filed up to March 1998 only 15.7 % resulted in some benefit to the claimant. Only 1.7% resulted in reinstatement or new hiring. Determination of no cause for complaint was found in over 50% of cases, and a third of cases were closed before resolution, possibly because of litigation. The median actual monetary benefit to those with psychiatric disabilities was $5000.41 The prioritisation system seems to work against psychiatric claimants as higher priority cases are more likely to result in benefit and those claiming for psychiatric disabilities less likely to be assigned to this category.42

    There is little comparable data in the UK and there is a reliance on those cases that make it to the courts. These give a biased view but illustrate the potential pitfalls. Employment appeals tribunals have upheld that an occupational physician acts as an agent of the employer and knowledge given to them about a possible disability is imputed to the employer when they make their report.43 The most high profile case was that of Watkiss v John Laing plc. Here the claimant revealed a history of schizophrenia during a routine medical examination, having been offered the post of company Secretary designate. The subsequent withdrawal of the offer “on medical grounds” led to a case where the employers admitted to unlawful discrimination and settlement damages.

    There are several useful websites in the UK offering information and help for employers—for example, the Department of Health sponsored MindOut campaign at (fig 3).

    Figure 3

    Poster for the UK MindOut campaign, sponsored by the Department of Health.

    Health and safety law

    The turning point under the Health and Safety Act was the case of Walker v Northumberland county council 1995. Mr Walker was a social worker who had previously had a nervous breakdown in 1986 which was attributed to an increased strain at work. A year later when fit for work he returned under the assumption that extra assistance be provided on his return. This was withdrawn after one month and in September 1987 he went on long term sick leave for a second “nervous breakdown” (the actual diagnosis seems to change from report to report). He was dismissed on the grounds of ill health in February 1988 and later sued his employers for breaching their “duty of care”. The high court found in his favour, extending the employer’s duty of care from reasonable foreseeable risks from physical to psychiatric illness. It was the second episode of illness, presumed to have been caused by the work load that was considered foreseeable and thus potentially avoidable.

    This principle was recently upheld in a case which should give heart to concerned employers (Cross v Highlands and Islands Enterprise 2001). A senior training officer was absent for two months with “stress”. On his return, having been certified fit for work, his employer took steps to reduce his workload and offered increase support. He tragically committed suicide two months later. His family sued for breach of duty of care but in this case the court found that his employers had exercised reasonable care for his foreseeable mental ill health by moderating his work load.

    The ruling in the recent court of appeal case (Hatton v Sutherland, 5 February 2002) has provided some important guidelines on foreseeability of breach of duty. The court held that an employee returning to work implies to their employer that they believe themselves fit to return to their previous work unless coupled with an explanation. The type of occupation cannot arguably be inherently stressful.


    The issue of adjustment is fundamental to “fitness for work”. Unfortunately there is no uniformly applicable psychiatric wheelchair ramp. Adjustments might include altering working hours, accommodating early morning sedation, allowing absences for treatment during working hours, and providing supervision for those who lack confidence. A reasonable level of absenteeism is “little more than what the employer accepts as sick leave for other employees”.44 The “reasonableness” is decided on a case by case basis and takes into account the effectiveness of the adjustment, its costs and disruption, and the resources of the employer. Again stemming from the Sutherland case, the court gave examples of positive steps that could be taken to avert the risk of relapse in someone returning to work: redeployment, offloading work, giving a sabbatical, bringing in temporary work or buddying.

    The multifactorial and individual nature of mental ill health has hampered research into what adjustments might actually work. Results from cohort studies using the Karasek45 model identify some prognostic indicators for subsequent absenteeism. High levels of demands, low levels of decision latitude and support at work predicted later depressive symptoms46 and common mental disorder.47 However, of all the workplace stressors examined by Mino et al48 only an alleviation of perceived responsibility was associated with a recovery from mental ill health in an occupational cohort. A further difficulty is that the causal relationship between work environment and mental ill health is not as well established as often thought. Comparative analyses suggest that while individual interventions may be beneficial, organisational interventions aimed at that holy grail of public health, prevention, may not be.49 For those with serious mental illness, or who have been off for some time, an access to work scheme will pay for some adjustment costs. The individual placement and support schemes with job coaches may potentially be backed by the government’s “New deal for disabled people”, with a network of job brokers and workstep scheme through the employment agency. The Royal College of Psychiatrists is shortly to publish a set of recommendations for those with psychiatric disability.

    A final novel idea might be to ask the individual what would help maintain their fitness for work. Bullying is often identified as a risk factor.50 Tensions can arise. How much can the adjustment interfere with the schedule of others and add to their workload? What about not getting on with specified individuals at work? How can these be applied in a manner that does not lead to further discrimination by co-workers? Good case studies are required on which to base future research.


    The relationship between mental ill health and fitness for work is still very unclear. Much of the research has looked at specific cohorts and often ignored factors external to the workplace. The advocates of the effects of “workplace stress” should not lose site of the fact that being unemployed is associated with twice the level of psychiatric morbidity of any employed group (teachers and nurses being possible exceptions). However, the working environment can be an important determinant of both mental ill health and, for many, wellbeing. Employers respond to economic arguments and legislation. For larger employers the most important determinants of workplace economic problems might not be those with clinical problems—for example, dependent drinkers or those with clinical depression—but the effect of common mental disorder and common drinking problems. However, from an occupational health view individuals cause problems in determining fitness for work. Anecdotally general practitioners and psychiatrists often act as patient’s advocates, which may lead the individual down the path of long term sickness and disability. Greater cooperation and understanding between different professional groups can only help, but confidentiality is often cited as a barrier, presumably through fear of discrimination. Challenging stigma and discrimination may aid this process.

    Legislation shows that, in recruiting and retaining individuals, employers and occupational health professionals need to be aware of the relapsing and remitting nature of many of these conditions, the potential for adjustments, and the foreseeability of future ill health. The courts and employment tribunals seem to be taking a reasonable line in assessing the other element: causality. However, the very nature of psychiatric illness makes prediction difficult in a chaotic world.

    The area is ripe for exploration. What seems to be lacking is research into the area in between stress and psychosis that so often flummoxes clinicians, occupational and human resource personnel, and managers: people with neurotic psychiatric illnesses such as depression. Addressing the fitness for work of those with mental illness and piecing together the effects of the work environment and individual is not helped by the lack of good data. Very little exists to aid accurate prediction. There is even less information on workplace adjustments under anti-discrimination legislation, and policies and practices that may prevent these illnesses developing or recurring, yet the commonplace nature of these conditions has far reaching medical, social, and economic implications.


    For each question please indicate which answers are true or false.

    1. Which of these are the approximate prevalences (currently ill) of different types of mental ill health in UK working age population?

      1. self reported stress: 30%

      2. psychosis: 5%

      3. depressive episode: 10%

      4. alcohol dependence: 14%

      5. mixed anxiety and depression: 10%

    2. The following factors have been shown to be predictive of occupational problems:

      1. average daily levels of alcohol intake in non-dependent workers

      2. psychoses with an early onset

      3. coming to work hungover

      4. having fully recovered from a single episode of depression

      5. being bullied at work

    3. Which of these factors may be associated with health and safety problems at work?

      1. common mental disorder

      2. smoking marijuana

      3. psychotropic medication

      4. comorbid psychosis and substance abuse

      5. a history of a psychotic episode

    4. Which of these conditions are probably covered by the UK Disability Discrimination Act?

      1. manic depression

      2. workplace stress

      3. a single episode of depression lasting four months

      4. alcohol dependence

      5. two episodes of depression each lasting seven months


    View Abstract

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles