As the arsenal of antidepressant drugs increases with time so, concurrently, does the list of caveats that must be considered when using these agents in patients with other physical illnesses and using other medications. MacHale's (2002, this issue) overview of the management of depression in physical illness serves as a crucial update for clinicians providing psychiatric services for patients with comorbid physical illnesses. Appropriate emphasis is placed on the usefulness of non-pharmacological treatments in such patients, although the reality of modern practice is that drug treatment is most often considered first-line owing to limited psychological service resources.

MacHale draws attention to the raised prevalence of depression in physically ill populations, though it is worth emphasising the complexity of this issue. First, estimating the prevalence of depression among subjects with physical illness from the research evidence available is not straightforward owing to the wide ranging methodologies used and the widely varying prevalences obtained. In general, however, a pattern emerges from this disparate literature that indicates that depression occurs in 10–15% of in-patient and out-patient populations on average, that is 2 to 3 times the rates seen in the general population. In addition to these, a similar proportion of patients have an excess of psychological symptoms. The latter group, while not fulfilling diagnostic criteria for a significant psychiatric illness, cannot be neglected because they represent an ‘at-risk population’ with a raised likelihood of developing a depressive disorder in the near future.

Prevalence rates for depression are raised, although not as dramatically, in general practice and hardly raised (if at all) in non-patient, asymptomatic subjects with physical illness (such as hypertension). Conversely, the prevalence of depression is raised further in specialist (tertiary) care settings and in subjects with highest levels of pain and disability, for example patients undergoing medical rehabilitation, with prevalence rates for depressive disorder reaching 50% in some patient groups. This pattern partly reflects an increased risk of depression as the severity of symptoms present increases. This is not the entire story, however. Patients with physical illness and depression are more likely to worry about physical symptoms, to interpret symptoms catastrophically and to present to their general practitioners. Such patients with depression are less likely to be satisfied with their medical treatment or to comply with any treatment offered. This probably results in these patients with depression being referred on to specialist services more frequently.

The causes of depression in physical illness are complex. In some conditions, the association with depression is strong and assumed to be due to a direct link between pathophysiological processes and depression, for example Parkinson's disease, stroke, hypothyroidism or medical treatments and depression. In other conditions, depression is often attributed to the burden of chronic symptoms. The link between depression and symptom load is, however, not clear. Certainly, there is a positive correlation between the prevalence of depressive symptoms and the number and severity of physical symptoms. However, this association appears to be weak at best. Convincing evidence of an association between physical illness and depressive disorder becomes apparent only in those who have the most disabling conditions, such as advanced rheumatoid arthritis. Social stresses, either associated with the medical condition or completely independent, appear to be very important aetiological factors for depression, as they are in physically fit patients with depression. Inherent vulnerability to depression and the efficacy of cognitive coping strategies mediate this complex process.

The impact of depression in physical illness is extremely variable. Depression increases the burden of misery, thus further impairing quality of life. In addition to this, however, depression has wider implications for the course and management of the physical illness itself. Depression, when present in association with a physical illness, is linked with a more negative outcome. Mortality rates may be raised: in the 6 months following myocardial infarction, there is a four-fold increase in mortality for patients with depression compared with patients with otherwise similar cardiovascular profiles but who do not have depression. These figures, generalised to the whole of the UK, suggest that depression may contribute to the early deaths of up to 200 000 patients after myocardial infarction.

Furthermore, patients with physical illness and depression are more likely to have further acute episodes of illness, require more hospitalisations and are likely to remain in hospital for a greater duration when they are admitted. As the result of these factors, health care costs for patients with physical illness and depression exceeds those for their counterparts without depression even after controlling for the severity of their illness. In a recent study of general medical in-patients, we found those with depression incurred costs 50% greater than those without depression. Hidden costs are also likely to be much greater for those with both physical illness and depression. Depression increases disability, predicting poorer functional status, more disability days and a delayed return to work following a period of sick leave.

A number of mechanisms have been proposed to explain how and why depression is associated with a more negative physical outcome in physical illness. Depression has been associated with a number of physiological changes that may contribute to this negative outcome, such as changes to the autonomic nervous system and the hypothalamic–pituitary–adrenal axis activity, and impairment in immune function. In addition, depression is associated with a number of adverse health behaviours that are also likely to contribute to an adverse outcome. Patients with depression are less likely to adopt positive health behaviours (diet, etc.), more likely to persist with negative health behaviours (e.g. smoking and excessive alcohol intake) and less likely to comply with treatment (take medication as prescribed, maintain self-monitoring regimes). Patients with physical illness and depression perceive their illness more negatively. They perceive the consequences of their illness as more serious, the likely duration of their symptoms as greater and their ability to control their disease as less adequate.

It is clear that treating depression in physical illness is important not only to optimise mental health but also to improve the outcome of the physical illness itself.