Severe mental illness
Definition of severe mental illness
Severe mental illness (SMI) describes a group of mental disorders that are persistent, cause serious functional impairment and substantially interfere with, or limit, major life activities. A key feature of SMI is the presence of psychosis, which is a loss of contact with external reality. Both the American Psychiatric Association1 and the World Health Organization2 (WHO) definitions of psychosis require the presence of hallucinations (i.e. perceptions occurring in the absence of corresponding external or somatic stimuli) and/or delusions (i.e. fixed false beliefs). These symptoms may be accompanied by cognitive impairment, disordered thinking and problems with motivation, energy and mood.
Although there is some debate about which conditions come under the SMI umbrella, there is consensus that the term includes schizophrenia, bipolar disorder and depression with psychotic symptoms.3–5 Schizophrenia is characterised by diverse psychopathology including delusions, hallucinations and ‘negative symptoms’, such as impaired motivation, reduction in interest, limited spontaneous speech, social withdrawal and cognitive impairment. Psychotic symptoms tend to relapse and remit, although around 20% of people with schizophrenia may experience chronic unremitting residual symptoms.6 Negative symptoms tend to be persistent and are associated with long-term effects on social functioning.
By contrast, bipolar disorder typically has an ‘episodic’ course, with recovery between relapses. It causes extreme mood swings ranging from mania or intense happiness, grandiosity, euphoria or irritability or decreased need for sleep, to low mood and depressive symptoms. Typically, a person with bipolar disorder cycles from one extreme to the other, while experiencing periods with few or no symptoms in between. Major depression is similarly a relapsing–remitting disorder, but it can have a chronic presentation. It is characterised by a persistent feeling of sadness or lack of interest in outside stimuli. When accompanied by psychotic symptoms, it is categorised as SMI.
Severe mental illness includes several other conditions, such as schizoaffective disorder and persistent delusional disorder, with symptoms that overlap with schizophrenia, bipolar disorder or severe depression. Apart from the presence of psychosis, these illnesses typically have a severe and enduring impact over the life course, affecting education, employment, relationships and wealth.7–9
Global burden of disease due to severe mental illness
Although lifetime prevalence of SMI is < 5% (consistent across the world), these conditions contribute a large and increasing global burden of illness. The global age-standardised point prevalence of schizophrenia is estimated to be 0.28% [95% confidence interval (CI) 0.24% to 0.31%], but it contributes 13.4 (95% CI 9.9 to 16.7) million years lived with disability (YLD), equivalent to 1.7% of total YLD.10 Similarly, the age-standardised prevalence rate of bipolar disorder is estimated to be 0.7% (95% CI 0.6% to 0.8%), contributing 1.3% of total YLD.11
Treatment of severe mental illness
Treatment of SMI includes a combination of psychotropic medication, psychological therapies and psychosocial support.
In recognition of the high risk of comorbid physical conditions and the socioeconomic impacts of SMI, treatment guidelines also include provision of annual general practitioner (GP) health checks to monitor physical health; smoking cessation support; and supported employment, exercise and diet programmes.12,13
The majority of people with SMI are likely to be prescribed antipsychotics, either typical (sometimes called ‘first generation’, developed in the 1950s) or atypical (sometimes called ‘second generation’, developed in the 1990s).14 Although all antipsychotics have been associated with side effects such as drowsiness, tremors, muscle spasms and weight gain, atypical antipsychotics are thought to have fewer extrapyramidal (motor control) side effects, but some may be associated with greater metabolic side effects and weight gain.14
Comorbid physical conditions and excess mortality and morbidity in severe mental illness
Burden-of-disease estimates do not account for the premature mortality and morbidity due to physical illnesses in SMI. People with SMI die, on average, 15–20 years earlier,15 with a death rate 3.7 times higher16 than the general population. Recent evidence indicates that this ‘mortality gap’ is widening for people with schizophrenia and bipolar disorder.17,18 Suicide accounts for around 15% of the increased mortality, but the majority of premature deaths are due to physical disorders, including non-communicable diseases, such as diabetes. Rates of such disorders are increased in people with SMI, typically being two to three times higher.4,19,20 It is estimated that as many as two out of three premature deaths among people with SMI are due to preventable physical illnesses.15
Most of the risk factors for developing conditions such as diabetes are likely to be the same in people with SMI and the general population, but there are multiple additional reasons that could explain the high prevalence seen in SMI populations and the excess mortality from physical disorders. Mental and physical disorders have a complex bidirectional relationship, sharing individual and socioenvironmental risk factors (e.g. childhood adversity and social and economic disadvantage).21,22 Medications for mental illness are often associated with adverse metabolic side effects, increasing the risk of cardiometabolic disorders. Sedentary behaviour is common because of the motivational deficits associated with these disorders, or the side effects of treatment such as tiredness and sedation. Additional health risk behaviours such as smoking, alcohol use and unhealthy diets are also elevated. Problems with cognition, energy and motivation also pose challenges for accessing and adhering to medical treatment for physical conditions. Moreover, people with SMI may be less likely to receive adequate treatment because of ‘diagnostic overshadowing’, whereby health-care services attribute reported physical symptoms to mental illness, and, consequently, fail to fully investigate and treat these symptoms.23
Comorbid physical disorders such as diabetes can therefore both drive, and be a consequence of, the significant health and socioeconomic inequalities faced by people with SMI. A better understanding of the determinants, and improved prevention and treatment of physical illness, could help to tackle the inequalities for this group.
Diabetes mellitus
Diabetes mellitus is a complex metabolic disorder characterised by chronic hyperglycaemia as a result of relative insulin deficiency, resistance or both. In 2019, the International Diabetes Federation estimated that 463 million people (1 in 11 of the global population) had diabetes and estimated an increase to 700 million by 2045. Approximately 4 million people in the UK are currently living with diabetes.24
Diabetes is associated with several short- and long-term complications that cause considerable morbidity, reduced quality of life and shortened life expectancy. These include acute metabolic perturbations (hypoglycaemia, diabetic ketoacidosis and hyperosmolar hyperglycaemic state), macrovascular disease (coronary artery disease, peripheral vascular disease and stroke) and microvascular disease (retinopathy, nephropathy and neuropathy). Diabetes was responsible for approximately 4.2 million deaths worldwide, or 11.3% of all deaths, in 2017, outnumbering the combined number of global deaths from human immunodeficiency virus/acquired immunodeficiency syndrome, tuberculosis and malaria.24
Type 2 diabetes mellitus (T2DM), which results from a combination of insulin resistance and less severe insulin deficiency, is the commonest form of diabetes, accounting for 90–95% of all cases. The rapid rise in T2DM largely explains the global epidemic, which is driven by an ageing population, longer survival with T2DM, earlier age at onset and better diagnosis. The prevalence of T2DM has increased as rates of obesity have risen.
Several risk factors have been identified for T2DM. The major modifiable risk factors are poor-quality diet and reduced physical activity. Smoking, mental illness (and psychotropic medication) and decreased sleep have also been implicated. Environmental risk factors include urbanisation, poverty and toxins. Non-modifiable risks include family history, ethnicity, birthweight and fetal under- and over nutrition, and history of gestational diabetes.25
It is estimated that, worldwide, approximately half of people with diabetes are unaware of their condition; however, this is lower in high-income countries, in part because of the introduction of screening programmes to systematically identify those with undiagnosed diabetes.
Once diabetes is considered, it is relatively easy to diagnose by the laboratory measurement of fasting plasma glucose (≥ 7.0 mmol/l), random glucose (> 11.0 mmol/l) or a 2-hour plasma glucose after a 75-g oral glucose tolerance test (> 11.0 mmol/l). The use of glycated haemoglobin (HbA1c) (≥ 48 mmol/mol) was introduced by the WHO as an alternative method in 2011.
Prevention of diabetes
Clinical trials have shown that it is possible to prevent or delay the development of T2DM through lifestyle and/or pharmacological interventions.26–30 Lifestyle interventions aim to reduce body weight and dietary fat intake, in particular saturated fat, while increasing dietary fibre and moderate physical activity, to ≈ 30 minutes a day. The UK has recently launched a National Diabetes Prevention Programme to support at-risk individuals to implement these lifestyle changes through referral to a behaviour change programme.31
Management of diabetes
For most people, diabetes is a lifelong condition. People will manage diabetes themselves (self-management) for most of the time, with only a few hours per year spent in contact with health-care professionals. Consequently, people must develop the skills to manage their condition effectively. Structured self-management education programmes have been developed and are recommended by the National Institute for Health and Care Excellence (NICE) as an integral component of diabetes management.32,33
Diabetes is managed through a combination of lifestyle changes and, when needed, treatment with antidiabetes drugs or insulin. The latest guidance highlights the importance of certain foods and dietary patterns.32 In common with the general population, people with diabetes should be encouraged to eat a healthy diet. For individuals who are overweight, losing weight is important; loss of 10–15 kg of body weight may trigger remission.32,34 Increased physical activity has profound benefits, including improved fitness, reduced insulin requirement, better glycaemic control, lower cardiovascular risk and greater life expectancy.
Diabetes in people with severe mental illness
The prevalence of diabetes is two to three times higher in people with SMI, and around 10% of people taking antipsychotic medications live with diabetes.35,36 A meta-analysis of 41 studies comprising 161,886 participants reported 9.0% (95% CI 7.3% to 11.1%) diabetes prevalence in SMI.36 In people with multiple episodes of psychosis, the prevalence of diabetes was double that of the general population, with earlier age at onset.37 The incidence rates of diabetes in SMI, however, are usually < 1%, meaning that the absolute risk for any one individual is small.38,39 Although diabetes is rare in adolescents and young adults, the relative risk of diabetes in young people exposed to antipsychotic drugs appears to be increased, compared with healthy controls and people with psychiatric disorders but unexposed to antipsychotic drugs.40 These increased risks are for T2DM; there is no evidence that the incidence of type 1 diabetes is increased in people with SMI.41
There are multiple potential reasons for the higher incidence of T2DM in people with SMI, including living conditions, lifestyle and disease-specific factors, as well as the effect of antipsychotic medications on insulin secretion and action.42–47 People with SMI tend to have a dietary pattern that includes increased intake of energy-dense foods that are rich in fat and refined sugars, with low levels of fruit and vegetable intake, and tend to have low levels of physical activity.42,43 Smoking and social deprivation are also important risk factors for diabetes, and the risks of both are higher in people with SMI.
In the early 2000s, as the use of atypical antipsychotics rose, reports of substantial weight gain, diabetes and dyslipidaemia began to emerge.48,49 It has now become apparent that rates of diabetes are higher among people taking typical or atypical antipsychotics than among the general population. Among people with SMI who have not yet started antipsychotic treatment, the prevalence of diabetes is low, but it rises rapidly after treatment initiation, suggesting that antipsychotics are involved in the aetiology of T2DM, with women and those with multiple episodes of psychosis being at higher risk.44
Studies, including randomised controlled trials comparing different antipsychotics, indicate that the risk of developing diabetes differs between medications.35,44,49–51 A consistently higher risk of diabetes is reported in people taking clozapine or olanzapine, with the lowest risks associated with aripiprazole, although all antipsychotics have been associated with increased risk of diabetes.44 Clinical trial data are supported by real-world comparisons between atypical antipsychotics.45,46 Diabetes risk also increases with the number and dose of antipsychotics prescribed.47 Potential mechanisms mediating this increased risk include weight gain, which increases insulin resistance, and the direct effects of antipsychotics, which decrease insulin sensitivity of cells and also impair insulin secretory capacity. There is debate about the magnitude of increased diabetes risk, with estimates varying widely up to a 33-fold increase; typically, however, the relative risks are < 2.52
Prevention of diabetes in people with severe mental illness
No diabetes prevention studies have been undertaken in people taking antipsychotics, but lifestyle interventions have been used to prevent weight gain or manage obesity. Short-term studies with follow-ups of < 6 months report weight loss of around 3.1 kg over a period of 8–24 weeks,53 but the results of longer-term studies are less consistent.54 Similarly, there have been no studies of pharmacological interventions, but the use of metformin leads to a modest 3.3-kg reduction in body weight over 3–6 months, in association with improved insulin sensitivity.53,55
Screening and diagnosis of diabetes in people with severe mental illness
Although regular screening for diabetes is recommended for people with SMI,56–58 many people taking antipsychotics are not screened regularly; further work is needed to understand why this simple measure has not been embedded into routine clinical practice.59,60 In 2011, targets for monitoring blood glucose were included in the national Quality and Outcomes Framework (QOF) to incentivise general practices to screen patients with SMI for diabetes. However, these targets were removed in 2014.61
Management of diabetes in people with severe mental illness
Diabetes appears to have a greater impact on people with SMI than on the general population through an increased incidence of acute metabolic emergencies and diabetes complications.62,63 Current NICE guidance12 suggests that diabetes in people with SMI should be managed in a similar way to diabetes in the general population.64 However, the guidelines do not take into account the implications of antipsychotic use or the unique challenges people with SMI may face in managing their diabetes.
Additional challenges faced by people with severe mental illness
Although the principles of diabetes management are the same as for the rest of the population, people with SMI face additional challenges in achieving optimal diabetes outcomes. These include, but are not limited to:
- health-care systems that separate diabetes and mental health
- overshadowing when physical health problems are considered to be caused by the mental illness
- excessive weight gain caused by antipsychotics and effects on insulin secretion
- psychotic symptoms
- poor cognition that interferes with decisions about self-management
- lack of social support.
Given the burden of illness, and health and health-care inequalities for this population, improving diabetes care for people with SMI is a high priority for the NHS.65 Little is known, however, about how SMI and other risk factors and challenges combine to generate high diabetes prevalence and poor diabetes outcomes, and how the quality and quantity of health-care services and interventions can influence these risk factors and outcomes in people with SMI.
Understanding this is a first step in developing health-care interventions to improve outcomes for people with diabetes and SMI. Better prevention and management of diabetes have the potential to significantly reduce the risk of diabetes complications, deliver large cost savings to the NHS and help reduce health inequalities (including life expectancy and morbidity) experienced by people with SMI.
In this report, we have focused on T2DM, as it is the most common type of diabetes in people with SMI. When the term ‘diabetes’ is used without specifying type, this refers to T2DM.
Social inequalities
Socioeconomic factors are powerful determinants of health outcomes, and a theoretical framework based on socioeconomic conditions is, therefore, appropriate for this study, in which social inequalities are likely to play a key role.66,67 Under this framework, health gaps in morbidity and mortality arise for people living in disadvantaged circumstances, from reduced access to the material, financial, social or structural resources that the advantaged population leverage to maintain or improve health. People with SMI are more likely to be socially disadvantaged than people without SMI.20,68–73 For example, only 6% of people with SMI in England and Wales are employed, whereas 58% are on long-term sick leave or are disabled and receiving benefits.73
There are three related explanations for this. First, there is evidence that social disparities, including racism and the multifactorial stresses of urban living, are among the causative factors for incident SMI.74–77 Second, the presence of severe and enduring illness patterns, and the physical and social consequences of these, thwarts efforts to maximise and improve social conditions. For example, SMI is diagnosed most often in early adulthood, and prodromal symptoms that interfere with functioning can be present months or years prior to onset.78 This may disrupt secondary education and limit educational progression, thus constraining the greater economic opportunities afforded by further and higher education. Third, if poor social conditions are experienced, these are likely to compound over time, for example due to individuals not having the resources to invest in social contact, high rates of housing precarity that places individuals into increasingly deprived neighbourhoods, and further restriction in income due to potential discrimination against mental illness in the benefits system.79–81 Deprived neighbourhoods are more likely to have environments that are obesogenic, where access to safe physical activity is limited and accessibility of poor-quality, calorie-dense foods is greater, increasing the risk of poor health.82,83 Stigma and discrimination, represented through labelling, stereotyping, separating, emotional reactions and status loss, and enacted through behaviours, in public attitudes towards people with SMI, internally held beliefs about how one is viewed by others, and structural barriers can thwart efforts to obtain health care and prevent illness, and can perpetuate social disadvantage.84–87
Given these details, it is not surprising that there is evidence that social inequalities play a role in driving poor mental and physical health outcomes for people with SMI.88–92 A complicating, and further disadvantaging, factor is an increased risk of developing diabetes, which is itself socially patterned.93 These greater health-care needs, combined with socioeconomic disadvantage, create a vulnerability that demands increased levels of care while simultaneously decreasing the likelihood of successful navigation of the health-care system.94 Social disadvantage also reduces one’s capacity to take advantage of health promotion opportunities, and creates more barriers to taking up interventions designed to prevent or treat illness. Further details about how we applied this framework are presented in Chapter 3, Theoretical framework.
Overview of routine health-care databases
This section describes the routine health-care databases used in this study; specific information about how they were used is provided in Chapter 5.
Clinical Practice Research Datalink
The Clinical Practice Research Datalink (CPRD) GOLD is the world’s largest computerised database of anonymised longitudinal medical records from primary care. The CPRD data set (originally the General Practice Research Database) was established in 1987. CPRD data include records of clinical events (medical diagnoses), referrals to specialists and secondary care settings, prescriptions issued in primary care, records of immunisations/vaccinations, diagnostic testing, lifestyle information (such as smoking and alcohol status), and all other types of care provided as part of routine general practice. Data have been collected on > 17 million patients from a network of 718 general practices throughout the UK, representing around 8% of the UK population registered with a GP.95 A cross-sectional study of the regional distribution of clinical computer systems in primary care found that 9% (636/7526) of general practices in England used Vision® (In Practice Systems Ltd, London, UK) software, from which the CPRD primarily draws its data.95
Clinical information is captured in the CPRD database using ‘Read codes’, which are recorded by primary care practice staff as part of routine data entry. Read codes96 are a comprehensive hierarchical coding system of clinical terms used in primary care to classify diseases, history and symptoms, patient characteristics, procedures and tests.97 The purpose of this coding system is to provide a standard language to enable accurate recording of patient information and more efficient retrieval of information for clinical or research purposes.97
The CPRD data set benefits from being a large, retrospective and prospective longitudinal primary care record. Patient characteristics in the database are broadly representative of the general UK population in terms of age, sex and ethnicity.98 Although other clinical computer reporting systems are used more commonly across the UK than the Vision system, patterns of area deprivation (based on the locations of general practices) have not been found to differ between these different systems.95 Practices using the Vision software system are, however, geographically concentrated in the south and north-west of England – particularly around London and Manchester – with relatively few practices based in Yorkshire and the Humber, and the north-east.95,98
Individual patient data in the CPRD can be linked electronically to external data sources, including Hospital Episode Statistics (HES) data for hospital admissions, Office for National Statistics (ONS) data for dates and causes of death, and the Index of Multiple Deprivation (IMD) to assess area deprivation in the locality. The CPRD provides a single extract of linked HES/ONS/IMD anonymised data.
Hospital Episode Statistics
The HES database contains patient records for inpatient admissions, outpatient appointments and accident and emergency (A&E) attendances at NHS hospitals in England.99 Admissions recorded in HES include those for physical and mental health problems where the patient was admitted to an acute hospital (but admissions recorded do not include specialist mental health facility admissions). HES Admitted Patient Care data contain information on patients’ demographics, admission sources and methods, discharge methods and destination, and primary and secondary diagnoses, as well as procedures conducted during the stay.
Diagnoses from hospital admissions are classified using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), codes. Procedures are classified using the Office of Population Censuses and Surveys Classification of Interventions and Procedures version 4 (OPCS-4) codes.
Office for National Statistics mortality data
The ONS mortality data include information from a person’s death certificate, such as cause of death, and date and place of death, for all registered deaths in England and Wales.100 These data are the most complete source of information on deaths based on information from medical practitioners and/or coroners.
Index of Multiple Deprivation data
The English IMDs provide a relative measure of deprivation at the lower-layer super output area (LSOA) level. There are around 35,000 LSOAs in England, each containing an average of 672 households.101 The IMD rank used for the linkage is from 2010, and is a combined index of seven domains: income, employment, health and disability, education, crime, barriers to housing and services, and living environment deprivation.102