The Sickening Mind: Brain, Behaviour, Immunity and Disease

Trouble, strife and sickness

Intense emotion usually falls short of causing people to drop down dead; it may simply make them more vulnerable to illness. And here again, at least some of the folklore has withstood scientific scrutiny. Research has confirmed the existence of systematic links between psychological factors such as anxiety, stress, depression and hostility, and a wide range of physical disorders including minor infections, gut disorders, herpes, allergies, asthma, arthritis, coronary heart disease and cancer. Indeed, according to some characteristically controversial research by the London University psychologist Hans Eysenck, certain psychological measures of personality and behavioural style have a greater bearing on which individuals will die from cancer or heart disease over the following ten to fifteen years than whether or not they are smokers.

Anxiety and stress have frequently been linked with vulnerability to illness. Numerous long-term studies have found that people who experience pronounced feelings of tension or anxiety are substantially more likely to develop coronary artery disease, or die from it, over the following years. For example, an American study which tracked several hundred people over a twelve-year period found that individuals who exhibited high levels of psychological distress were roughly twice as likely to die as those with only average levels of distress. This connection between distress and death held up even when other medical risk factors such as old age, obesity, smoking, high blood cholesterol and high blood pressure were taken into account, so it was not merely a question of distressed subjects also being old, fat or smokers. Psychological distress was related to subsequent mortality in its own right.

Similar conclusions emerged from a Harvard University project. This investigated the health of former Harvard students whose psychological and biological profiles had been assessed thirty-five years earlier, as part of a series of laboratory experiments on stress. The way subjects reacted during the laboratory tests predicted their physical health years later. Individuals who displayed signs of severe anxiety during the original stress tests subsequently suffered from significantly more physical illnesses, including coronary heart disease, over the following decades. Responding anxiously to a stressful situation when a young adult proved to be a reliable marker for ill-health of all types in middle age. Another investigation by scientists at Harvard Medical School found that very shy children, who suffered from severe anxiety when in social situations, were more prone to allergic disorders such as hay fever.

It may help to look in greater detail at one specific example of a fairly subtle connection between psychological factors and subsequent disease. An American research project conducted in the 1970s investigated the psychological characteristics associated with infectious mononucleosis, otherwise known as glandular fever. This unpleasant and debilitating disease is prevalent among teenagers and young adults. The symptoms include a general malaise, fever, sore throat, loss of appetite, headaches, together with swelling of the lymph nodes or ‘glands’ in the neck, groin and armpits. Recovery can take many weeks. Occasionally, serious complications arise, such as damage to the liver or spleen.

The disease is caused by a type of herpes virus known as the Epstein-Barr virus (EBV), which we encountered in chapter 1 as a once-favoured cause for chronic fatigue syndrome. In common with other herpes viruses like herpes simplex (which causes cold sores and genital herpes), EBV can remain dormant in the body for years without causing any symptoms. Dormant viruses are normally held in check by the individual’s immune system, but anything that weakens immunological control over the latent viruses can trigger the emergence of disease symptoms.

The subjects of this investigation comprised over 1300 young men entering the West Point military academy. On arrival at West Point each student was screened to see whether he was already infected with EBV. About two-thirds of the students carried the virus, which is typical for a normal population. The remaining third had not yet been infected. These potentially susceptible students were then tracked to see who would become infected with EBV. And here lies an important general point: not everyone who is exposed to disease-causing bacteria or viruses becomes infected. In fact, only about one in five of the originally virus-free students went on to be infected with EBV during their four years at West Point. Of those who did become infected, a quarter developed obvious clinical symptoms of disease. And here lies a second general point: not everyone who gets infected with disease-causing viruses or bacteria develops a clinical disease.

Psychological assessments revealed that those men who went on to be afflicted with infectious mononucleosis shared certain distinctive psychological characteristics. In particular, they tended to be the ones who had suffered most from academic pressure. Students who had the dispiriting combination of a strong motivation to do well, but a poor actual performance, had a greater likelihood of contracting infectious mononucleosis. And once they became ill these highly motivated but poorly performing students spent longer on average in hospital. They were more susceptible to the disease and when they got it, they got it worse.

Relatively minor traumatic events can also push up the odds of becoming ill. For example, Australian scientists found a marked increase in high blood pressure, gut disorders and diabetes among people who had been indirectly affected by a bushfire that occurred in southern Australia in 1983.

Long-term observations of normal families have shown that there is often an increase in family-related stress, or disruptive changes in family circumstances, in the period immediately before one or more family members develop infections. A number of studies of families in their home environments have unearthed associations between stressful conflicts and minor infectious illnesses such as coughs, colds, ’flu and sore throats. These stressful episodes tend to precede infections rather than follow them, implying that the stress contributes to the illness and not vice versa. In other words, it is not simply a matter of arguments arising because everyone is feeling ill and crotchety.

Research in the States has uncovered comparable links between stress and illness among children in rural Dominica. In the week following a high-stress event such as a big family upheaval, the risk of the children acquiring an infection of the upper respiratory tract increased by a factor of three.

Life events

For over thirty years scientists have been systematically exploring the idea that the risk of falling ill increases when we are exposed to a lot of disruptive changes or emotional turmoil. This research stemmed from the informal observations of certain perceptive doctors, who noticed that their patients often seemed to have experienced unusually large amounts of change and upset in the period before they fell ill. Further impetus came from a pioneering investigation of illness and absenteeism among the employees of the Bell Telephone Corporation in the 1950s. This indicated that employees with unsettled personal lives tended to suffer frequent bouts of illness and take more sick leave from work.

Suggestive observations such as these led psychologists to formulate the concept of life events. A life event is defined as any significant change in a person’s circumstances which requires them to make psychological and practical readjustments. The disruptive event can be either desirable or undesirable; the prime criterion is that it causes a degree of upheaval.

Examples of life events include the death of a partner or family member, divorce, marriage, starting a new job, moving house or financial problems. At the other end of the scale, minor upheavals such as family holidays and Christmas are also classified as life events. The basic hypothesis underlying this work is that any disruptive changes, whether desirable or undesirable, are potentially stressful and can increase our chances of falling ill.

Thousands of research projects have investigated the relationships between life events and health. The majority of these studies have used a standardized method for assessing life-event stress called the Social Readjustment Rating Scale. In its simplest form this involves asking each individual to record which of forty-three types of life event they have experienced over a specified period, usually between six months and two years.

Each type of life event is assigned a standard score according to its supposed severity, rated on a scale from o (least severe) to 100 (most severe). The maximum rating of 100 is awarded to the death of a spouse; divorce is rated 73; marriage, 50; changing to a different line of work, 36; moving house, 20; Christmas, 12; and so on. (Personally, I would rate Christmas at around 60, and anyone who has recently experienced the horrors of moving house may be excused for wondering at its modest rating.) A composite score is then calculated for each individual, taking account of both the total number of life events they have experienced and the relative awfulness of those life events. A high score can denote a few serious life events or a multitude of minor ones.2

If it is true that life events act as risk factors for illness then people who register high life-event scores should, on average, have more illnesses than those whose lives have been undisturbed by change. Simple. By and large, this is what the research has found.

A seminal early investigation looked at the effects of life-event stress on US Navy personnel during the Vietnam War. The results showed that individuals with the highest life-event scores suffered almost twice the number of illnesses over the following months as those with low scores. In another study scientists asked young men in a navy submarine training establishment to record the life events they had experienced over the previous twelve months; again, the incidence of life events correlated with subsequent illness.

The general conclusion from several thousand such studies is that people who have been exposed to lots of life-event stress have a slightly greater risk of illness. This increased risk applies across the board and seems to encompass virtually every form of ailment and disease under the sun, ranging from headaches, common colds, allergies and inflammation of the gums to mental illness, coronary heart disease, leukaemia, diabetes, tuberculosis and multiple sclerosis. Life-event stress also has an impact on childbirth; women who register high stress ratings during the year or so before pregnancy tend to give birth to babies with slightly lower birth weights and a slightly poorer overall state. Life events are even associated with an increased risk of minor accidents and sports injuries.

As well as suffering more episodes of illness, people with high life-event scores also tend to be ill for longer, have more severe symptoms and take longer to recover.

Not surprisingly, the adverse effects of life events are generally worse when the life events are severe, undesirable and clustered together in time. In the early days of life event research it was widely assumed that ‘good’ life events, such as getting married or starting a new job, were potentially just as damaging to health as ‘bad’ life events of comparable disruptiveness. However, more recent research has tended to support the common-sense assumption that, other things being equal, undesirable life events are inherently more damaging than desirable ones.

It has to be said that the link between life events and later illness is not as neat and simple as it sometimes appears. Some of the research on life events has been justifiably criticized for a variety of reasons. This is not the right place to debate the abstruse technicalities of research methodology. Nonetheless, the difficulties inherent in life event research are of broader relevance and therefore merit our attention.

First of all, the statistical correlation between life events and illness is highly consistent but it is also fairly weak. Life events do have a bearing on health, but not a very major bearing. Typically, life events account for only about 10–15 per cent of the total variation in the incidence of illness. A number of those who are exposed to stressful life events become ill, but most do not. Conversely, it is possible to fall ill despite living a life of unruffled stability. A phenomenon that is highly significant in a strictly statistical sense – meaning that the patterns in the data are more than just chance variations – may not necessarily be highly significant in a clinical or scientific sense.

A second fundamental point is that correlation is not the same as causation. The existence of a statistical association between two things is not proof that one of them causes the other. The population of the world and the age of the current pope are correlated, but there is no causal connection between the two. They both happen to be independently related to a third variable – time. So the correlation between life-event scores and illness does not by itself prove that life events are a direct cause of illness. The causation might even work the opposite way round; that is, chronic illness might conceivably precipitate life events. For instance, someone’s marriage or career might run into problems because they are ill. And it may be the case that things which are classified as life events, such as sexual problems or changes in sleep patterns, could in fact be symptoms of an existing but undiagnosed illness.

In order to disentangle cause and effect in this type of research it is vital to establish which came first, the life events or the illness. There is plenty of evidence that life events do indeed tend to precede illness, which suggests that they may genuinely contribute to ill health.

A third pitfall with life event research, especially in its early days, has been its retrospective nature. When investigators ask subjects to recall their life events during, say, the previous year, great reliance is placed on frail and faulty memories. And therein lies a weakness. It is an awkward fact of life that most of us grossly over-estimate our ability to recall the past accurately and objectively. Ask any policeman, lawyer or judge about the reliability of witnesses to crimes. Psychologists have found that after a period of ten months people are typically able to recall life events with an accuracy of only 25 per cent. Conclusions that depend on people’s memories of what happened to them one or two years ago are therefore bound to be suspect.

As well as the inherent difficulty of recalling past events accurately there is also a danger of systematic bias. People who are unwell may focus on a particular trauma in their past and assume it must have been responsible for their illness. We all have a basic need to find explanations for our illnesses and some people understandably attribute their poor health to traumatic experiences. But in doing so they inadvertently undermine the objectivity of the research data.

Fortunately, not all research on life events has had to rely on faulty memories. Instead, scientists have monitored groups of initially healthy subjects over a period of time, recording their life events and illnesses as and when they happen. This style of research is referred to as prospective, in contrast to the backward-looking retrospective method. And plenty of these prospective studies have borne out the link between life events and subsequent illness.

Another potential pitfall lies in failing to distinguish between an interviewee’s actual health, as measured according to objective, clinical criteria, and what they say or think about their health. The problem here is not that people deliberately lie; the majority of those who volunteer to take part in scientific research try hard to be truthful. The real problem is that few of us are capable of being entirely objective about our own health. We all perceive and interpret our physical symptoms in different ways; something that would constitute a distressing malady for one person might not even be noticed by another.

Problems also arise if we attempt to measure health in terms of what is called sickness-related behaviour. This means behaviour like going to the doctor or taking sick leave from work. Sickness-related behaviour is obviously not the same thing as actual sickness.

The way humans respond when they think they are ill depends on other factors besides their state of health, including such mundane considerations as whether expert medical advice is freely and conveniently available. Sickness-related behaviour is often more a reflection of psychological factors than physical health.

To complicate matters further, people’s perception of their own state of health varies according to their mental and emotional state. Anxious or stressed individuals, for example, are more apt to notice and worry about minor symptoms, interpret them as evidence of disease and seek expert help. Someone who has been experiencing lots of stressful life events is more likely to feel unwell and visit their doctor, but this does not necessarily mean that they are actually ill.

We shall be looking at this issue in more depth in the next chapter. Suffice it here to say that there is a world of difference between believing yourself to be ill, or going to the doctor, and having a clinically verifiable disease. For this reason, research that relies wholly on self-assessments of health or on sickness-related behaviour can be misleading. Such measures often say more about people’s mental state than they do about their true physical health. I should add, however, that the dubious practice of using sickness-related behaviour as an ersatz measure of health is a pervasive problem in medical research and is certainly not unique to work on life events.

Despite these caveats there is consistent evidence, garnered from thousands of scientific studies, for a connection between life events and subsequent illness. It is now clear that even the mundane hassles of everyday life have an impact on physical health. Indeed, some scientists have argued that because these hassles are such a frequent occurrence their cumulative influence on health may be more pervasive than the effects of rarer, but more traumatic life events.

The general idea that psychological factors can affect susceptibility to physical illness is amply supported by research on other species. As in so many other respects there is nothing biologically unique about humans. Several decades of experimental work on other species have confirmed that various forms of psychological stress can increase (or, occasionally, decrease) animals’ susceptibility to a wide spectrum of diseases, including bacterial and viral infections, heart disease and cancer.

For instance, when mice or rats are exposed to stressful situations, such as being physically restrained or subjected to unpleasant electric shocks, they become less resistant to infection with a whole range of bacteria, viruses and parasites including mycobacteria (the type of bacteria responsible for tuberculosis), herpes viruses, influenza viruses, polio viruses and the protozoa which cause toxoplasmosis. In one experiment, for example, frightening mice by exposing them to a cat significantly increased their vulnerability to infection with a parasitic tapeworm. (The cat was prevented from attacking the mice; the sight of it alone was enough to affect them.) Likewise, the social stress of being introduced into an unfamiliar flock makes chickens more susceptible to bacterial infections, while the stress of being transported renders cattle vulnerable to a form of viral pneumonia caused by the reactivation of latent herpes viruses.