Possible reasons for low screening rates

One plausible hypothesis is that the low uptake in those with mental illness is explained by current distress. Stress, distress and anxiety have been linked to short-term risk-averse behaviours resulting in avoidance of screening invitations. ^{Reference Ng and Jeffery72} Indeed, distress has been linked with low uptake of mass screening such as mammography and colorectal cancer screening. ^{Reference O'Donnell, Goldstein, DiMatteo, Fox, John and Obrzut73,Reference Kotwal, Schumm, Mohile and Dale74} However, there may also be a bimodal, inverted U-shaped relationship between screening and anxiety, with both high concern and high distress linked with lower screening rates. ^{Reference Consedine, Morgenstern, Kudadjie-Gyamfi, Magai and Neugut75} In this study in a large sample of 21 491 women we found no significant link between distress and mammography. Unfortunately there were insufficient data to analyse the effect of anxiety alone, but when anxiety and distress were pooled we still were unable to demonstrate any consistent link. This suggests that current distress (or anxiety) is probably not the explanation for low receipt of mammography.

Another possible reason for the low screening rates among women with mental illness is low rates of presentation for medical help, largely a patient-determined factor (help-seeking). For instance, Hardy & Gray found that only 66% of those with SMI attended an appointment in primary care when specifically invited for a physical health check, compared with 81% of those with diabetes in the same centre. ^{Reference Hardy and Gray76} However, other studies have found that attendance is not the main variable in determining medical care inequalities. ^{Reference Mitchell, Malone and Doebbeling8,Reference Lord, Malone and Mitchell9,Reference Mitchell and Hardy70} Kahn et al stipulated that patient education is key to facilitating breast cancer screening among women with mental illness. ^{Reference Kahn, Fox, Krause-Kelly, Berdine and Cadzow77} For example, patients commonly fear pain or discomfort when attending for mammography, and education and reassurance might ameliorate these concerns and improve attendance. ^{Reference Kahn, Fox, Krause-Kelly, Berdine and Cadzow77} If the low receipt of screening is explained by low help-seeking rates then more attention is needed during the invitation process aimed at those with known psychiatric diagnoses and other low-attendance groups. A follow-up telephone call can improve attendance in those with mental illness. ^{Reference Mitchell and Selmes78} A second explanation might be the influence of cognitive impairment on decision-making. Most people with mental illness do not have substantial enduring cognitive impairment, but several past studies have found that low uptake occurs in those with learning disabilities. ^{Reference Sullivan, Glasson, Hussain, Petterson, Slack-Smith and Montgomery33,Reference Legg, Clement and White79} This requires further study.

Fig. 4

Mammography screening in women with severe mental illness. Pooled odds ratios: random effects meta-analysis.

Not all mammography occurs as a result of routine screening invitations. Indeed, in many low-income countries mammography is by clinician request. However, in such countries most clinicians do not send asymptomatic women for screening, unless the patient makes a specific request. ^{Reference Temitope, Daniel and Ademola80} Hence a number of authors have stipulated the importance of good communication between primary care providers and mental health services to maximise cancer screening among women with mental illness. ^{Reference Kahn, Fox, Krause-Kelly, Berdine and Cadzow77,Reference Miller, Lasser and Becker81,Reference Friedman, Puryear, Moore and Green82} Research in the general population has also stipulated the importance of having a primary care provider to improve screening rates, ^{Reference Starfield, Leiyu and Macinko83} but it may be that trust and social support are additionally important to facilitate screening attendance in women with mental illness. ^{Reference Aggarwal, Pandurangi and Smith11} The importance of good communication and continuity of care has previously been emphasised in a qualitative study involving people with SMI and healthcare professionals. ^{Reference Lester, Soroham and Tritter84} However, the low screening rates may be partly ascribed to a failure of primary healthcare providers to take the physical healthcare complaints of people with mental illness seriously. ⁴ For instance, healthcare providers may fail to screen people with mental illness for cancer owing to a preoccupation with other comorbidities and confusion around symptom attributes. ^{Reference Aggarwal, Pandurangi and Smith11} One author stated that providers attributed many of the patients’ physical complaints to their psychiatric symptoms, which resulted in an underestimation of the post-test probability of other medical conditions. ^{Reference Masterson, Hopenhayn and Christian54}

Study limitations

It is important that a number of limitations are considered when considering the results of this review. First, most of the analyses that we conducted had considerable heterogeneity. We attempted to negate this and improve clinical relevance by stratifying the results according to clinical diagnosis and also by reporting only random effects meta-analysis. Second, most studies were conducted in North America, and may not be generalisable to other areas of the world with different healthcare systems. Third, it may be that shared risk factors such as social deprivation may account for the low mammography uptake, but it was not possible to investigate this clearly within the analysis. Fourth, the time frame over which mammograms were recorded varied across studies. Lastly, there was heterogeneity in the classification of mental illness, with a range of classification criteria used, and some authors used retrospective history. However, despite these factors, the results from our large meta-analysis were consistent: people with mental illness, particularly those with serious disorders, are substantially less likely to receive mammography screening compared with members of the general population.

Clinical implications

People with mental illness are at significant risk of not attending breast cancer screening although reasons for this disparity remain to be confirmed. Most research shows that people with mental illness receive inferior medical care. ^{Reference Thornicroft1,Reference Mitchell, Malone and Doebbeling8} Cancer is particularly burdensome in people with mental illness, and clearly such people should receive care that is at least comparable with care given to the general population. Although there has been some debate in the general medical literature regarding the use of mammography (see, for example, Jorgensen & Kotzsche), ^{Reference Jorgensen and Gotzsche85} the inequality demonstrated within this review is clearly not proportionate to this group’s healthcare needs. This is particularly so when research has demonstrated that among people with mental illness cancer is often detected later, and when it is detected more metastases are often found. ^{Reference Kisely, Crowe and Lawrence27} However, care is needed when employing large-scale mammography screening because false positives may cause psychological harm. ^{Reference Jorgensen and Gotzsche85} A recent longitudinal study found that in the general population annual mammography screening had no effect on breast cancer mortality beyond that of physical breast examinations. ^{Reference Miller, Wall, Baines, Sun, To and Narod86} However, it remains unclear if this finding applies to those with mental illness. Indeed, people with severe mental illness may not receive regular physical health checks and are probably less likely to carry out self-examination. In order to enhance breast cancer screening, it is important that there is better communication between NHS England’s screening services, primary care providers and mental health services. ^{Reference Kahn, Fox, Krause-Kelly, Berdine and Cadzow77,Reference Miller, Lasser and Becker81,Reference Friedman, Puryear, Moore and Green82} Efforts should also be made to educate and support women with mental illness to engage in breast cancer screening and social support may be particularly important in achieving this. ^{Reference Aggarwal, Pandurangi and Smith11}

Future research

Future prospective research should investigate the barriers to as well as the facilitators of mammography screening among women with mental illness. This should use both qualitative and quantitative research methods, drawing upon the experiences of the patients and of mental healthcare and primary care providers. Additionally, it is essential that prospective longitudinal studies are conducted to investigate the influence of mammography screening upon the diagnosis (including false positives), treatment and ultimately the mortality of women with mental illness.