88 - 98: Cancer and Mental Illness

It is well known that people with mental illness have higher mortality rates and reduced life expectancy compared to people without mental illness. Our own work using record linkage systems in Western Australia has shown that the gap in life expectancy for people with and without mental illness has increased since 1985, to 12 years for women and 16 years for men, principally because improvements in life expectancy for the general population have not extended to people with mental illness [1].

While mental illness is associated with an increased risk of suicide, much of the excess mortality associated with mental illness is actually due to common physical health conditions, in particular cardiovascular disease, respiratory diseases and cancer [1,2]. With a particular focus on cancer - the topic of this chapter - we have previously reported that cancer diagnosis rates are comparable between people with and without mental illness; at the same time, however, not only is case fatality after a cancer is diagnosed substantially worse in people with a history of mental illness, but survival times are also much shorter (fig. 1) [2,3,4]. In order to reduce the excess mortality attributable to cancer, it is important to understand the underlying factors as to why people with mental illness have worse case fatality associated with cancer. There are obvious and immediate avenues to examine, in particular whether lifestyle factors place people with mental illness at greater risk of developing cancer. For instance, are they less likely to participate in cancer screening activities, leading to differences in how soon diagnoses are made and therefore presenting with more advanced disease? Do people with mental illness receive differential treatment once they are diagnosed? Are they more likely to reject lifesaving treatments or not comply with aftercare requirements? All of these factors may go some way to developing a more rounded understanding of why people with mental illness experience greater cancer mortality. In this chapter we address each of these questions, reviewing the evidence as to where the greatest discrepancies lie and which aspects of cancer diagnosis and treatment deserve further attention.

Study of the relationship between poor physical health and high mortality in people with mental illness has a long history. In 1841, William Farr [5 ]reported to the Royal Statistical Society on mortality within the major asylums in England. At that time, high mortality was principally ascribed to infectious diseases and the generally poor conditions within asylums. Around the same time, the Commissioners in Lunacy were established to oversee the conditions and care provided in asylums and workhouses in England. The commissioners were among the first to note that in spite of the general poor health of people with mental illness, cancers were not often seen, and they suggested in their 1909 report that people with mental illness may have a certain immunity to cancer [6]. With the dawn of the computer age, new methods of research in mental illness became viable. The Oxford Record Linkage Study, established in the 1960s, was the first study to undertake a systematic analysis of the association between physical health problems and mental illness, using administrative data from mental health services in the Oxford region [7]. At that time, the main question of interest was whether an association between a physical illness and a mental illness could reveal something about the aetiology of both conditions. The relationships between cancer and schizophrenia and between rheumatoid arthritis and schizophrenia have received particular attention [8]. The strong negative correlation between these conditions led to hypotheses that there might be specific genes involved in the aetiology of these diseases.

As more data were accumulated over time and the pattern of association between schizophrenia and cancer was observed to be more complex, more nuanced hypotheses were advanced. These included the idea that certain neuroleptic drugs may have a tumour-suppressing effect, that this effect may vary depending on the stage of cancer during neuroleptic treatment, that the effect may vary depending on cancer morphology, or there may be an interaction between neuroleptics, levels of specific hormones and genetic susceptibility to certain types of tumours [9,10,11,12]. However, to date this line of investigation has not led to any significant advance in the understanding of the aetiology or treatment of either schizophrenia or cancer.

While much of the early research focussed on the possibility of people with mental illness being at reduced risk for developing cancer, the possibility that cancer may be a cause of excess mortality was first reported by Katz et al. [13] in 1967, who found that cancer mortality amongst psychiatric patients in New York was higher than in the general population, though only for patients with shorter periods of hospitalisation. Patients with psychiatric hospital stays of more than 10 years had lower cancer mortality rates. Since then, there have been numerous studies on cancer mortality and mental illness, and most studies published in the last decade have found higher standardised mortality ratios associated with cancer [3,4,14,15,16,17]. However, these findings have not always been consistent, and have varied according to the years of study, the mental illness and type of cancer examined, and the study populations and methodologies that each study used. For example, the results of earlier studies typically found no excess mortality attributable to cancer [18,19,20,21], and in some cases lower standardised mortality ratios were observed [22,23].

There is considerable research demonstrating that people with mental illness engage more frequently in behaviours that are associated with cancer risk. For example, people with mental illness have higher rates of smoking and smoke more cigarettes on average [24], and despite wanting to quit smoking find it much harder than people without mental illness to do so [25]. Drug and alcohol addiction is also more common in people with mental disorders [26], and many psychiatric disorders are treated with medications which, in addition to other lifestyle and diet factors, result in increased incidence of obesity [27]. Additionally, a common side effect of antipsychotic medications is hyperprolactinaemia, an excess of the hormone prolactin. Hyperprolactinaemia is also associated with an increased risk of breast cancer in women [28], and possibly prostate cancer in men [29], though others have noted that hyperprolactinaemia is also associated with hypogonadism, which may be protective against breast and prostate cancer [30].

Because of these lifestyle and medication side-effect risk factors, it may seem reasonable to hypothesise that increased exposure to these cancer risk factors would cause more cancers in people with mental illness. If this were the sole reason for worse cancer mortality, then we would expect that people with mental illness would be at greater risk of developing cancer, and as such we would see an increased incidence of cancer amongst people with mental illness. Despite the high rates of exposure to demonstrated carcinogens in people with mental illness, there is in fact no clear evidence that people with mental illness are more likely to develop cancer, or at least be any more likely to be diagnosed with cancer. Studies that have examined both cancer incidence and mortality in populations of people with mental illness indicate that cancer incidence rates for people with mental illness are generally comparable with the rest of the population, though this pattern appears to depend on the type of cancer and the severity of mental illness. Many studies have found that cancer rates are lower amongst people with schizophrenia in particular [31,32,33,34]. However, not all studies have found this pattern amongst people with schizophrenia [15], and as such, this pattern for people with schizophrenia may reflect methodological or cohort differences [35]. While some studies have found an increased incidence of smoking-related cancer, such as lung cancer, in people with mental illness [36,37], the overwhelming majority of studies that examine psychiatric disorders, other than schizophrenia, have found no increased risk of cancer incidence in psychiatric populations [3,4,15,36,38,39,40,41]. In addition, people with severe mental illness do not have a greater incidence of melanoma, but have worse case fatality even though melanoma is a cancer that cannot be attributed to concurrent lifestyle factors such as diet, drugs, alcohol or tobacco, as it is mainly related to childhood sun exposure [3].

In the case of cancers that are associated with risk factors such as smoking, alcohol and other drugs, comparable rates of cancer incidence in people with and without mental illness may imply some other, as yet unknown, factor that reduces incidence of cancer in people with mental illness. Some exceptions are apparent, for example a systematic review of the incidence of breast cancer amongst women with schizophrenia found that while the incidence rates of breast cancer varied across the studies, those with larger samples and better quality indicators found higher rates of breast cancer in women with schizophrenia [42]. A possible explanation for this finding, if true, may relate to the increased likelihood of hyperprolactinaemia for people who are administered antipsychotic medication. However, we have not observed any increase in breast cancer in women with severe mental illness in our own work in Western Australia and Nova Scotia [2,3,4].

Brain tumours are a notable exception to the finding that the incidence of cancer is comparable in psychiatric and general populations. A number of studies have identified higher rates of brain cancer diagnosis following diagnosis of mental illness. A likely explanation for this finding is the general difficulty of diagnosing brain cancers, and the common early presentation of brain cancer sufferers with symptoms of mental distress including depression and memory impairment. This may lead to people with brain tumours being referred for mental health treatment prior to the diagnosis of their tumour [2,43].

Less studied than cancer incidence, but more consistent in its findings, is the higher case fatality following cancer diagnosis in people with comorbid mental illness. This has been consistently reported from a number of large record linkage and cohort studies including our own record linkage work using the populations of Western Australia and Nova Scotia [3,4]. This worse case fatality from time of diagnosis deserves additional focus and attention. Here we examine the cancer screening rates for people with mental illness and further treatment discrepancies that occur following diagnosis.

Another possible explanation for the comparable rate of cancer diagnosis in people with mental illness despite high exposure to known carcinogens is the hypothesis that cancers do occur more frequently in people with mental illness, but are less likely to be detected because of lower rates of screening and access to general healthcare. As life expectancy is substantially reduced in people with mental illness, and mortality from other causes of death is substantially higher, more people with mental illness who have cancer may die before that cancer is diagnosed. To date, the results about the extent to which people with mental illness receive routine cancer screening have been mixed. For example, in a study of privately insured women with, and without, claims for mental illness, women with a mental disorder were significantly less likely to receive mammography, particularly women with more severe disorders [44]. Similar patterns have been observed in other studies with different populations of patients [45,46,47,48], though others have found that women with mental illness were more likely [49], or at least equally likely, to receive mammography [50]. In a narrative review of 17 studies examining the uptake of screening services by people with mental illness, Happell et al. [51] showed a 20-30% reduced likelihood of breast, cervical and colorectal cancer screening for patients with severe mental illness in the majority of studies. The pattern was most commonly observed for people with more severe disorders, with up to a 60% reduced likelihood [52]. Taken together, the evidence suggests that people with mental illness are less likely to receive preventive cancer screening services than those without [53]. As a result, cancers may be more advanced by the time the diagnosis is made, with poorer outcomes associated with the delay in diagnosis.

The question of why people with mental illness may receive less cancer screening remains unresolved. For instance, some research points to the possibility that people with mental illness are less likely to be offered screening for cervical, breast, prostate or colorectal cancer by their primary physicians [54], or if screening is recommended, are less likely to adhere to recommended screening [55]. In their systematic review, Lord et al. [56] identified the relationship with the primary care physician as perhaps the most important factor to explain deficits in receipt of cancer screening and other preventive health measures. They note that proactive contact from the primary care physician, and the use of peer support workers, are two interventions where there is some evidence to support their effectiveness in boosting screening rates.

Differences in screening rates aside, there is a growing body of literature documenting other disparities in cancer treatment for people with mental illness, including the length of time it takes to be diagnosed with cancer after presenting with symptoms. In a study of patients with oesophageal cancer, O'Rourke et al. [57] found that cancer patients who had a DSM-IV diagnosed psychiatric comorbidity waited a median of 90 days between reporting symptoms and receiving a diagnosis, compared to 35 days for patients without mental illness. They were also more likely to have advanced disease at the time of diagnosis (37 vs. 18%) and also had a decreased likelihood of receiving surgical therapy (38 vs. 59%). In other research, the proportion of people with cancer who had metastases at presentation was significantly higher in psychiatric patients (7.1 vs. 6.1%), though larger gaps were observed for specific types of cancer [3]. For example, for people with breast cancer the difference was 6.3 vs. 4.5%, and lung cancer 0.6 vs. 0.2%. No differences in rates of cancer with metastases were found for other sites. Though statistically different, these difference rates are small in absolute terms, and cannot account for the significantly larger differences in case fatality for cancer patients. These patterns, reinforced by findings from another recent study [58], suggest that the elevated mortality associated with cancer is unlikely to be primarily the result of the cancer being more advanced by the time it is diagnosed.

There are apparent disparities in other ongoing treatment options, including rates of surgical intervention and the time it takes to get to surgery. In our recent work using the Western Australian Data Linkage System [3], psychiatric patients had a reduced likelihood of surgery after diagnosis for all types of cancer, and for those who did have surgery, the length of time between cancer diagnosis and surgery was longer (see fig. 2 for trends). Men were less likely to have a colorectal resection, and women were less likely to have surgery for colorectal, breast and cervical cancer. The psychiatric diagnosis also mattered - colorectal resections occurred less frequently for patients who were diagnosed with dementia, affective psychoses, other psychoses and depression. Psychiatric patients also received 10.3 sessions of chemotherapy on average, compared to 12.1 for the general population, and were less likely to receive radiotherapy for breast cancer (2.6 vs. 4.1%), colorectal cancer (1.6 vs. 3.9%) and uterine cancer (13.0 vs. 21.1%). To summarise, although the cancer incidence in psychiatric patients is no higher than in the general population, psychiatric patients are more likely to have metastases at diagnosis and less likely to receive specialised interventions. Together, these may explain the greater case fatality found in people with a psychiatric disorder.

Though our focus here has been on cancer outcomes for people with a prior history of mental illness, for many people, the onset of mental illness can occur following cancer diagnosis and treatment [59], particularly post-traumatic stress disorder and depressive and anxiety disorders. Research indicates that people with comorbid diagnoses of cancer and mental illness have poorer outcomes and greater difficulty adhering to treatment regimens [60,61]. In this respect, issues surrounding treatment disparities apply to all people with mental illness, irrespective of their psychiatric history prior to a cancer diagnosis. However, different types of mental illness offer different challenges to people with mental illness and to the practitioners who treat them, and the types of mental illness that arise following the development of cancer are somewhat narrower in scope than the range of illnesses that develop over a lifetime. Research that compares treatment outcomes for people whose onset of mental illness was before or after cancer may offer further insights as to why disparities occur.

People with mental illness who also have physical health problems are often more likely to be among the more complex cases with multiple presenting symptoms and problems. For example, in a small-scale review of 29 people with schizophrenia and lung cancer [62], potentially curable cancers were identified in 17 individuals, with 5 of the 17 receiving less than best practice care. Issues that resulted in disparities in care included other physical comorbidities that complicated the treatment and ethical issues with patients not consenting to invasive treatments. Irrespective of the presence of mental illness, patients with complex needs who have multiple comorbidities have worse outcomes and are more poorly served by standard medical care [63]. With the high level of specialisation now required to deliver high-quality healthcare for any condition, it is clear that few healthcare providers are well equipped to deal effectively with diverse and challenging health problems simultaneously [64].

Equitable access to healthcare has become an important focus of research in the gaps in physical health outcomes associated with mental illness, and a number of contributing issues have been noted in the literature [65,66]. First, there is the long established separation of mental and physical healthcare, with many mental health service users not receiving ongoing physical healthcare, and many psychiatrists reluctant to diagnose and treat physical health conditions. Additionally, stigma is seen as a major issue in the equitable delivery and access to general healthcare. Unfortunately, mental illness can rob people of aspects of personality that others find most endearing, and even among health professionals there are many who find it difficult to provide the same level of concern and support to people with mental illness. Also, the way that healthcare is delivered in developed countries generally results in greater benefit to those who have the best understanding of health conditions and the healthcare system. People with mental illness may also have limited means to pay for healthcare or might not be covered by health insurance. Many mental illnesses are associated with cognitive impairments and disadvantages that can impact people's ability to maximise their use of healthcare options. These include smaller support networks and impacts on motivation, concentration, assertiveness and communications ability. In some cases these deficits have been used as justification not to provide health interventions. For instance, some cancer surgeries are highly invasive and debilitating, and pose substantial demands during recuperation. Similarly, chemotherapies and other more experimental drug treatments can have significant side effects and exact a toll on the patient.

The issue of obtaining informed ethical consent to risky treatments or treatments with substantial side effects also requires sensitive consideration. Some people with mental illness require a higher level of support to understand the implications of treatments, which may require more time commitment from the practitioner. Compliance with aftercare may also require a higher level of support. However, in the same way that physical disability is not a valid reason for not providing access to healthcare, any cognitive or behavioural impairment associated with mental illness should be recognised as a component of the illness that deserves support.

Finally, another aspect of medical care for people with both mental illness and cancer is end-of-life palliative care, which shares some similarities with the recovery model of mental healthcare that is based on respect, dignity, and valuing both life and the individual [67]. Results from limited research on this issue have been contradictory. Some work suggests that people with severe mental illness may be less likely to receive palliative care [67]. In contrast, our own preliminary work, using data from the Western Australian Data Linkage System for the period 1988-2007, indicates that people with a history of mental illness who were dying of cancer received the same level and duration of palliative care prior to death as the general population, with an average of 19 days of palliative care [unpubl. data]. In spite of the challenges that people with mental illness may pose to palliative care environments, they may especially benefit from such care when making difficult end-of-life decisions or deal with significant pain, as they are less likely to have strong personal support networks [67]. Again, the issue of stigma and the cognitive and behavioural consequences of mental illness can affect the way people are received and treated in the palliative care environment [68,69].

With significant advances in the treatment and understanding of mental illness, and large-scale change in mental health service delivery to community-based care, there have been significant shifts in treatment goals and expectations. There are now effective treatments available for most mental health conditions [70], and a major focus of mental healthcare is on recovery from mental illness and valuing the contribution people with a history of mental illness make to their communities [71,72]. This has led to a much greater focus on the human cost of the high physical comorbidity associated with mental illness, the wasted life seen in the substantial gaps in life expectancy [73].

Mental illnesses are common, and people with mental illness suffer a disproportionate burden associated with physical illness, including higher mortality from cancer. Bringmann et al. [74 ]estimate that up to one third of cancer patients have a psychiatric history. As such, dealing with comorbid mental illness is a significant component of cancer care. Though people with mental illness may not be at any higher risk than anyone else for developing cancers, the clearly worse outcomes once cancer is diagnosed highlight the role that preventive and treatment services are likely to play in reducing these disparities. There is insufficient research at present to definitively identify the contribution that individual issues make to the overall gaps in outcomes, but it is highly likely that a range of factors contribute. Further research into ways to improve health service delivery to people with comorbid mental health problems is required. It appears that reduced access to screening, delays in diagnosis, differential access to and uptake of treatment options, and impacts of other comorbidities may all play a role. Clearly, these issues are not straightforward to resolve as shown by the lack of major progress anywhere in the world. There is a small and developing evidence base supporting some directions for improving physical healthcare for people with mental illness, although none of this literature is specific to cancer. These include variations on the theme of more holistic care models including multidisciplinary teams and co-location, use of peer supporters and addressing stigma in the healthcare profession [66.]