Introduction
A key component of the mixed-methods concurrent triangulation study design involved the convergence of results from the quantitative and qualitative studies. In this chapter, we present the integrated themes that represent the study findings. First, the methods for integration are described, and the challenges to achieving this are discussed.
Methods for integration
The primary method for integration was ‘convergence’, which involved bringing together the results from each study to generate a more complete understanding of the relationship between SMI and diabetes. Co-design workshops formed part of this process to ensure that study findings, and the integration of these, were interpreted by service users and health-care providers, in addition to the research team. ‘Embedding’ techniques were also used to ensure that, when possible, each study informed the other throughout the research.
Embedding
Although we used a convergent design, in which the integration primarily occurs when the study findings are brought together, we also considered how the studies could be integrated at the data collection and analysis stages.
The following strategies were employed:
- The scoping of the literature and expert consultation (see Chapter 4) informed the selection of a priori variables to explore quantitatively, qualitatively or in both studies when this was deemed important and was possible. For example, sleep was identified as an important determinant of diabetes self-management, but was not possible to explore reliably in CPRD data, and instead was added as a topic for the qualitative study.
- We had anticipated that the sampling criteria and interview topic guides for the qualitative study would be informed by the quantitative analysis for study objectives 1–3. However, this was not possible because of delays in obtaining CPRD data and operationalising key variables. Instead, we used preliminary analyses of a smaller CPRD data set carried out in preparation for the study to explore population characteristics to inform the sampling criteria.115
- Emerging results and observations from each study were discussed at each project meeting to identify opportunities for exploring these further; for quantitative results, using qualitative inquiry; and for qualitative findings, using quantitative interrogation of health records.
Co-design workshops
Two multistakeholder co-design workshops were used to make sense of emerging findings, to explore the potential for additional integration during analysis (e.g. re-analysing qualitative data to explain the finding from a quantitative model) and to situate the findings in the context of policy and practice from the perspective of both service users and providers. Owing to time delays in the quantitative study, the workshops primarily focused on emerging qualitative findings.
We invited study participants who had expressed an interest in the workshops and interested stakeholders who received newsletters about the study and wider research programme. To ensure service user engagement at the workshops, we also invited DIAMONDS VOICE members, who had an existing relationship with the study team and who helped to facilitate small group discussions and offered feedback during whole-group sessions.
Outputs from the workshops were used to help interpret the qualitative findings and were discussed with the study team to ascertain whether or not we could explore key points raised in the workshops in the quantitative study (e.g. the relationship between depression and diabetes outcomes). They also formed part of the final integration of study findings (see Convergence).
Workshop 1
Workshop 1 was attended by six service users and focused on learning more about two qualitative findings: (1) that mood affects diabetes management and (2) that it can be difficult to tell whether symptoms are due to SMI or diabetes. To explore the first finding, we presented two personas (David and Barbara; see Report Supplementary Material 16), based on findings from the qualitative study. We used an empathy map (see Report Supplementary Material 17 for an example) to discuss and record what the personas might think, do, say and feel when their mental illness affected them and how this affected their diabetes. For the second finding, we presented quotations from the study about the uncertainty regarding symptoms, and asked participants to write down feelings and symptoms on sticky notes and place them on a board to indicate whether they were associated with diabetes, mental illness or both.
Workshop 2
Workshop 2 was attended by nine service users, two informal supporters and seven staff (from services that support people with SMI and diabetes). We wanted to learn more about care provision for people with SMI and diabetes, in part to make sense of the passivity we had noted from participants about diabetes care, but also because of the limited data for exploring diabetes care in the CPRD (due to the unreliability of referrals data for diabetes education and specialist services).
We presented findings about care experiences and described the challenges to exploring the complexity of care provision in the quantitative study. Using a large board that represented 11 different types of health and social care professionals highlighted in the qualitative study, we asked participants to map their expectations for care against these in relation to SMI, diabetes, and health and well-being, and then to map their experiences (for an example, see Report Supplementary Material 18). In the final session, we asked participants to draw on their discussions to identify three changes that would improve diabetes care for people with SMI. In this workshop, service users and carers/informal supporters worked separately from health-care staff to explore any differences in expectations and experiences, and to enable service users to discuss their experiences together.
Convergence
To develop our understanding of the relationship between SMI and diabetes in this final phase of the study, we used principles of both convergence, which involved comparing and contrasting the study findings, and transformation, which involved translating the study findings into descriptive summaries and merging these when possible.105 This final integration was underpinned by the study’s social inequalities theoretical framework, which considered how social and economic disadvantage can contribute to worse physical and mental health.
In practice, this process entailed a full-day workshop attended by the whole study team, and subsequent analysis of workshop outputs and discussions. Specifically, this involved (1) producing narrative and tabular summaries of findings from both studies, which were reviewed by the whole study team prior to the integration workshop; (2) mapping and synthesising study findings from the quantitative and qualitative studies to the study objectives in the integration workshop and subsequently into a table to allow for comparison (Table 27); (3) identifying key themes across the study findings and discussing these within a social inequalities framework in the workshop; and (4) developing the themes and writing integrated explanatory summaries for these to ensure that all findings, including any contradictory or unusual findings, fit within this framework. Although themes were derived through the integration process, they were not selected based on their potential for integration, but rather on their importance for understanding the relationship between SMI and diabetes, and the poor diabetes outcomes experienced in this population.
| Objective | Description | Quantitative data | Qualitative data | Conclusion |
|---|---|---|---|---|
| 1 | In people with SMI, to identify which sociodemographic, illness, family history and lifestyle factors are associated with the development of diabetes | Older age, ethnic minority, higher deprivation, presence of comorbidities, antipsychotic use and family history of diabetes were all predictors for diabetes diagnosis and time to onset Higher than normal BMI and glucose dysregulation were both identified as potential risk factors for diabetes | Many participants had several physical comorbidities, which affected their ability to manage diet and weight (theme 2) Participants discussed the devastating effect antipsychotic drugs had on their physical health, but mental health had to take priority (themes 1 and 3) Unhealthy food is cheaper, so those with less money may be more likely to have a poor diet Different dietary habits in ethnic minority communities may contribute to unhealthy diet | The risk factors for developing diabetes for people with SMI are similar to those for the general population, but a combination of increased levels of deprivation and taking antipsychotic medication increases the risk of diabetes for a SMI population People with diabetes and SMI are aware of the negative effects of their medication, but feel that this is a price they must pay for reduced SMI symptoms. They also identify issues with physical health that could contribute to poor weight management and, therefore, high BMI |
| 2 | In people with SMI and diabetes, to identify which sociodemographic, illness, family history and lifestyle factors are associated with variation in diabetes and mental health outcomes | 73.3% of eligible patients had a SMI diagnosis before a T2DM diagnosis, 26.7% had T2DM diagnosed first or had both diagnosed together Patient baseline characteristics differed by diagnosis order. When included as a risk predictor, diagnosis order had a similar effect to T2DM duration Older age, lower socioeconomic status, T2DM duration, presence of comorbidities, being a smoker and substance use predicted macrovascular complications Older age, comorbidities and high baseline HbA1c level predicted microvascular complications Older age, male sex, Asian ethnicity, higher deprivation, SMI type, SMI duration, number of comorbidities, antidiabetic and antipsychotic medication use and substance use were predictors of all-cause mortality Younger age, SMI type, comorbidities, antidiabetic and antidepressant medication use were all predictors for mental health relapse | Most participants had a diagnosis of SMI before diabetes, although not all. Some participants, especially those with bipolar disorder, discussed identifying SMI symptoms long before their diagnosis Many participants had several physical comorbidities, which affected their ability to manage diet and weight (theme 2) The effect of SMI type on diabetes management was explored. Few differences were found and both groups identified barriers to their diabetes management. People with schizophrenia discussed the effect of antipsychotic drugs on their physical health to a greater extent than those with bipolar disorder (theme 3) Links were identified between presence and severity of mental health and diabetes symptoms (themes 3 and 4) | It is unclear if the findings relating to diagnosis order are due to limitations of the data set. Qualitative feedback from people with diabetes and SMI suggests that some people are experiencing SMI long before diagnosis, which may account for some of the group diagnosed with T2DM before SMI or concurrently. This is not conclusive, and needs to be explored further People with diabetes and SMI in the qualitative study identified comorbidities that affected their diabetes management. Those with more comorbidities, therefore, may be likely to have more trouble managing diabetes and therefore have worse outcomes Although all people with diabetes and SMI discussed barriers to diabetes management, the presence or absence of which could affect diabetes outcomes, people with schizophrenia were much more likely to discuss the effects of antipsychotic medication. This may highlight a difference between SMI types Links identified between mental health and diabetes suggest that outcomes could be linked, that is those with poorer mental health may also have poorer diabetes outcomes |
| 3 | In people with SMI, to compare health-care interventions and physical and mental health outcomes in those with diabetes with those of people without diabetes | T2DM prevalence in eligible patients with SMI was 10.2% Presence of T2DM increased the risk of macrovascular and microvascular complications, SMI relapses and mortality T2DM increased the chances of people with SMI receiving health checks for glucose, cholesterol and blood pressure levels | Note that all qualitative study participants had a diagnosis of diabetes Relationships were identified between diabetes and mental health symptoms and severity (themes 3 and 4) People with diabetes and SMI describe having regular diabetes health checks in primary care (theme 6) | A diagnosis of diabetes adds complexity to a diagnosis of SMI alone, increasing the risk of death, mental health relapses and macro- and microvascular complications. Relationships identified in the qualitative study between SMI symptoms and diabetes could help explain the links between T2DM and mental health relapse Having T2DM was associated with increased physical health checks. Frequent checks were described by qualitative participants |
| 4 | In people with diabetes, to compare health-care interventions and physical and mental health outcomes in people with SMI with those of people without SMI | People with SMI had a higher risk of emergency admissions and a lower risk of elective admissions for macrovascular complications People with SMI had higher all-cause mortality, were more likely to die from chronic ischaemic heart disease and had increased chance of health checks for HbA1c, blood pressure and cholesterol | Note that all qualitative study participants had a SMI diagnosis People with diabetes and SMI described having regular diabetes health checks in primary care. Some participants suggested that these were not regular enough, which may lead to loss of diabetes control in between checks (theme 6) Staff and people with diabetes and SMI identified instances whereby somatic symptoms were attributed to mental health by health-care staff (diagnostic overshadowing), sometimes leading to much more serious health problems later | People with SMI receive more physical health checks than those without SMI, yet have higher rates of mortality and a higher risk of emergency admission and a lower risk of elective admission to hospital for macrovascular complications. This suggests that people with SMI are not being referred to cardiovascular specialist care, despite receiving health checks in which health issues would be expected to be identified Qualitative study participants discussed incidences of diagnostic overshadowing, which could help explain this finding. If symptoms are being attributed to mental rather than physical health, warning signs of illness could be missed until the point at which they become an emergency and cannot be ignored when patients are admitted to hospital |
| 5 | To understand the factors that are associated with access to, and receipt of, diabetes care for people with SMI, and to explore the experience of diabetes health care by people with SMI | This was a qualitative research question exploring personal experiences. The quantitative research was not able to address this objective | Key themes:The pervasive effect of SMI – diabetes cannot take priorityThe effect of multimorbidity on diabetes managementThe interacting and overlapping nature of diabetes and SMI symptomsHow mood affects diabetes managementThe importance and impact of informal support networksExperiences of diabetes health careDiabetes knowledge and education and impact on self-managementA desire for person-centred care with an emphasis on continuity | Experiences differed between participants; for some, having diabetes and a SMI is linked to numerous mental and physical barriers to self-care and diabetes self-management. The importance of personalised support from people who know you and understand all your conditions was emphasised |
| 6 | To compare diabetes care provision for people with and people without SMI, and to estimate costs for these | People with SMI and diabetes have significantly more primary care consultations and more mental health inpatient stays (of a longer duration) than people with diabetes alone They have a similar number of non-mental health inpatient stays but patients with SMI stay for significantly longer People with SMI and diabetes use consistently more health-care resources than people with diabetes alone | People with diabetes and SMI are seeing many different health-care professionals as a result of their two primary conditions and often multiple comorbidities. Mental and physical health-care services are often seen as separate entities and function in this way (themes 2 and 8) Links between mental and physical health symptoms and severity have been highlighted (themes 3 and 4) | People with diabetes and SMI are understandably using more health-care resources because of multiple conditions and comorbidities The lack of joined-up care described in the qualitative data may account for the increased consultations for this group. Links between mental and physical health symptoms could mean that physical health issues leading to inpatient stays were exacerbating mental health issues, resulting in extended illness periods and longer hospital stays |
| 7 | To identify which health-care interventions may be associated with better diabetes outcomes for people with SMI and diabetes | People with SMI had increased chances for health checks, compared with controls, in people without macrovascular complications (as outcome) People with SMI had reduced chances for health checks, compared with controls, in people with macrovascular complications (as outcome) | The importance of support for diabetes and mental health management, both informal and from health-care services, was emphasised. Ongoing support from a familiar person was important to people with diabetes and SMI. Individualised care that focused on all aspects of a person and their condition was wanted (themes 5 and 8) Fluctuating nature of moods, and therefore diabetes management, was identified, which is unlikely to be captured in 6- to 12-monthly checks (themes 3, 4 and 6) | Ongoing, regular support and continuity of care were discussed in the qualitative study as being key to diabetes and mental health management. The fluctuating nature of mood (not just SMI symptoms) and the close links between mental state and diabetes management suggest that closer monitoring of this group could help improve self-management and outcomes |
Results
Six themes were identified from the integration process. These are described in the following sections, with reference to individual study findings when relevant.
Diagnostic overshadowing
From the analysis of patient records, we found that, despite people with SMI and diabetes having more frequent contact with their GP and more diabetes checks (e.g. cholesterol, blood pressure, HbA1c, BMI) than people with diabetes alone, they still have a much higher risk of mortality (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results). We also found that, although the rates of macrovascular and microvascular complications were similar in both groups (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results), people with SMI are more likely to present to A&E for treatment (i.e. to have unplanned care) than people with diabetes alone, who have higher rates of elective care (see Chapter 5, Regression results for the impact of severe mental illness status on physical and mental health outcomes, mortality and health checks). Both of these findings suggest that, although there is a greater opportunity to detect and treat diabetes complications in people with SMI, this is not happening. Findings from the qualitative study suggest that diagnostic overshadowing, which occurs because of the foregrounding of SMI (see Chapter 6, Theme 1: mental illness affects everything) in a health-care system that treats mental and physical illnesses in silos (see Chapter 6, Separation of mental and physical health care), may help to explain this.
The qualitative study also found that diagnostic overshadowing is influencing patient, as well as staff, behaviours. So, although participants in all three groups advocated for care that was person-centred and co-ordinated (see Chapter 6, Theme 8: person-centred care), and described the separation of physical and mental health services as a barrier to this, they also expressed beliefs that primary care treats physical health and mental health services treat mental illness. Participants with SMI and diabetes also described encounters with staff during which these beliefs influenced what they spoke about as patients (see Chapter 6, Separation of mental and physical health care). Staff accounts reflected these beliefs too: although they agreed that SMI and diabetes interact, limited time and resources tended to restrict their focus according to these beliefs (see Chapter 6, Separation of mental and physical health care). Lack of training in both conditions reinforced the practice of focusing on a single condition (see Chapter 6, Training needs), and, in the co-design workshops, participants recommended that all health-care staff be trained in multiple conditions and multimorbidity.
The foregrounding of SMI affected patient and staff behaviours further, as they focused on addressing the greater impact of psychosis relative to diabetes (see Chapter 6, Foregrounding mental illness), which included addressing social issues, such as housing and benefits (see Chapter 6, The pervasive effect of severe mental illness). This, and the difficulties attributing symptoms to either illness (see Chapter 6, Theme 3: interacting conditions, overlapping symptoms), resulted in the symptoms of diabetes being hidden, and, for many participants, becoming a lesser priority. This may explain why we obtained such limited responses about diabetes care from service user participants (see Chapter 6, Theme 6: diabetes health care), which contrasted significantly with the detailed accounts of mental health care.
Depression and anxiety
In the context of SMI, this study suggests that, like diabetes, depression and anxiety may be overlooked because of the focus on managing and preventing psychotic symptoms, and living with the experience of these symptoms (see Chapter 6, Theme 1: mental illness affects everything). Depression and anxiety were highlighted by participants as being difficult to attribute to SMI or diabetes (see Chapter 6, Theme 3: interacting conditions, overlapping symptoms), and this was confirmed by co-design participants who consistently linked mood symptoms to both conditions. There was also a suggestion that these symptoms may not be prioritised for treatment (see Chapter 6, Fluctuating moods, fluctuating management); when we asked about this in the co-design workshop, service users reported generally not seeking help for low mood, and clinicians agreed that depression is not always considered as a potentially separate condition to manage.
This additional overshadowing is important, because many service user participants in the qualitative study described low mood, anxiety and stress as recurring features in their lives (see Chapter 6, Theme 4: the effect of mood on diabetes management), in addition to the symptoms of psychosis and mania that were associated with a relapse. In the quantitative study, we found that people with SMI who took antidepressant medications had an increased risk of macrovascular complications (see Chapter 5, Regression analysis results for the impact of type 2 diabetes on health outcomes and health-care interventions), even though having depression and anxiety was associated with better attendance for health checks that could help to prevent such complications (see Chapter 5, Regression analysis results for the adjusted impact of severe mental illness on levels of physical health checks). We also found that people with SMI and diabetes had an increased risk of depression and anxiety, compared with people with diabetes alone (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results). This risk was highest for those diagnosed with diabetes before SMI; for these patients, their SMI diagnosis was more likely to be depression and psychosis, rather than other types of SMI (see Chapter 5, Objective 2: factors associated with variation in diabetes and mental health outcomes in people with severe mental illness and diabetes, Results).
Service user participants seemed to relate their anxiety and low mood to worries about diabetes (see Chapter 6, Worries about diabetes) and the impact of poor physical health on their lives (see Chapter 6, Theme 3: interacting conditions, overlapping symptoms). At the same time, depression and low mood were perceived to contribute to worsening physical health, through lower motivation to manage diabetes and other health problems (see Chapter 6, Fluctuating moods, fluctuating management).
This complex relationship between depression and anxiety, SMI and diabetes suggests that low mood may result in poor diabetes outcomes in people with SMI and may help to explain why people with both conditions are more likely to experience a SMI relapse. Low mood is more likely to occur in people who are marginalised and less likely to be noticed in primary care, and people with SMI who develop T2DM are more likely to live in the most deprived neighbourhoods than those with SMI who do not develop T2DM (see Chapter 5, Objective 1: factors associated with the development of diabetes in people with severe mental illness, Results). This relationship may form part of the mechanism linking poor social conditions and worse physical and mental health.
Better diabetes control, or hidden fluctuation?
We found that, although people with SMI and diabetes have poorer outcomes, including increased risk of mortality, than people with SMI alone and people with diabetes alone, they consistently have lower HbA1c and blood pressure levels (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results), which are indicative of better diabetes management. Our integrated findings offer three potential explanations.
The first explanation relates to diagnostic overshadowing (see Diagnostic overshadowing); despite there being a greater opportunity to detect problems [through more testing and contact with GPs (see Chapter 5, Objective 6: comparing diabetes care provision and estimating health-care costs for people with and people without severe mental illness, Results)], there is disproportionately less action taken, with the consequence of underdiagnosis of complications and delayed referral to specialist services for people with SMI (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results). The second explanation is that diabetes control may fluctuate more widely in people with SMI because of the additional challenges to managing their diabetes. Such fluctuations would not be evident in primary care records without more regular monitoring of glucose. Co-design participants recommended glucose monitoring as a future intervention for people with SMI and diabetes to help increase awareness of whether a symptom was due to diabetes or SMI, and to improve diabetes management. The third explanation is that people with SMI who have the greatest risk of poor outcomes from diabetes may be the least likely to attend diabetes checks and receive specialist diabetes care. For example, we found that, among those with macrovascular complications, people with SMI were less likely to receive health checks than people with diabetes and no SMI (see Chapter 5, Objective 7: identifying health-care interventions associated with better outcomes for people with diabetes and severe mental illness, Results). Moreover, people with SMI and diabetes had more psychotic relapses than those without diabetes (see Chapter 5, Regression analysis results for the impact of type 2 diabetes on health outcomes and health-care interventions), and we know from the qualitative study that psychotic symptoms can lead to a foregrounding of mental illness and feelings of helplessness, and negatively affect self-care, engagement with services, social support and health management (see Chapter 6, Theme 1: mental illness affects everything, and Theme 5: ‘the most critical person in my care’ – the role of informal support networks).
Tied intimately to these potential explanations is that social and economic deprivation contribute to the health gap. People with SMI living in deprived areas or who are of black or Asian ethnicity were more likely to develop diabetes, and these factors were the strongest predictors of developing diabetes (see Chapter 5, Objective 1: factors associated with the development of diabetes in people with severe mental illness, Results). Compared with controls with diabetes only, people with SMI and diabetes lived in more deprived areas, which was a significant predictor of worse health outcomes (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results, Descriptive statistics). In addition, social deprivation was associated with fewer health checks (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results).
Multimorbidity
Both studies highlight the importance of situating comorbid SMI and diabetes within the framework of multimorbidity. People with SMI and diabetes had more comorbidities than people with SMI and no diabetes; this included depression and anxiety as well as physical illnesses such as cardiovascular disease and hypertension (see Chapter 5, Objective 3: comparing health-care interventions and health outcomes in people with severe mental illness and diabetes with those in people with severe mental illness without diabetes, Study population). When controlling for other key variables, comorbidities were also associated with both the onset of diabetes in people with SMI (see Chapter 5, Regression analyses results for the risk factors for development of type 2 diabetes) and poor outcomes for people with diabetes, whether or not they had SMI as well (see Chapter 5, Regression results for the impact of severe mental illness status on physical and mental health outcomes, mortality and health checks).
Surprisingly, we found that, when comparing outcomes between people with diabetes with SMI and people with diabetes without SMI, people with SMI had fewer physical comorbidities and similar rates of cardiovascular disease, even though they had a significantly increased mortality rate (see Chapter 5, Regression results for the impact of severe mental illness status on physical and mental health outcomes, mortality and health checks). Social deprivation may, in part, explain this, because people with SMI were more likely to live in a deprived area, and this was associated with a higher risk of mortality (see Chapter 5, Regression results for the impact of severe mental illness status on physical and mental health outcomes, mortality and health checks). Diagnostic overshadowing (see Diagnostic overshadowing) may also offer an explanation, with physical health problems being attributed to SMI; therefore, they are not diagnosed or treated. The qualitative study sheds light on other potential reasons too, as many service user participants talked about physical health problems such as poor sleep, musculoskeletal problems and poor mobility, which we could not reliably explore in the CPRD data, but which affected their ability and motivation to manage diabetes (see Chapter 6, Theme 2: not just two illnesses – multimorbidity and diabetes management). The sense of being overwhelmed by poor health from multiple conditions, each with their own treatment and management burdens (see Chapter 6, Theme 1: mental illness affects everything, and Theme 2: not just two illnesses – multimorbidity and diabetes management), led many participants to respond by prioritising SMI as their dominant concern, potentially to the detriment of optimal diabetes outcomes.
In the context of multimorbidity, this study additionally highlights the importance of ordering of diagnosis. Unexpectedly, we found some evidence that people who are diagnosed with diabetes first may have poorer health outcomes (see Chapter 5, Objective 2: factors associated with variation in diabetes and mental health outcomes in people with severe mental illness and diabetes, Results, and Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results). The qualitative study included very few participants who received their diagnoses in this order and offers little insights into how different their experiences are, but, considering the overwhelming nature of having SMI, it may be that the layering of psychosis onto poor physical health has a greater impact on an individual’s ability to manage their diabetes (see Chapter 6, Participant characteristics).
The role of antipsychotic medication
We found that, after controlling for other variables, treatment with antipsychotic medication increases the risk of developing diabetes in people with SMI (see Chapter 5, Objective 1: factors associated with the development of diabetes in people with severe mental illness, Results). For people with both conditions, use of antipsychotic medication increases the risk of mortality (see Chapter 5, Regression analyses results), and, for people with both conditions, atypical antipsychotics are associated with worse outcomes. The qualitative study also highlights concerns about the side effects of antipsychotic medication, including weight gain, increased appetite and lethargy (see Chapter 6, Mental health medication and diabetes management). Some participants associated these with both the onset and management of diabetes, and all three side effects were identified as barriers to eating healthily and being physically active (see Chapter 6, Mental health medication and diabetes management). Fewer people with diabetes and SMI than staff identified a link between antipsychotic medication and diabetes, and many expressed a resigned acceptance that living with the side effects that came with such medication was a necessity (see Chapter 6, Persistence and powerlessness). Very few participants talked about antipsychotic medications in relation to diabetes treatment; for example, no participant talked about addressing the side effects of antipsychotic medications to help their diabetes, or the possibility of switching medication as potential solutions (see Chapter 6, Theme 6: diabetes health care).
The value of relationship-based care that supports management of both conditions
The quality and quantity of relationships with others, both friends and family, and health-care providers, were identified as important resources for health management, and were also reported as being deficient by many participants (see Chapter 6, Theme 5: ‘the most critical person in my care’ – the role of informal support networks). When talking about person-centred care, the positive aspects that participants highlighted were often about the relationships they had established with individual mental health care co-ordinators, and the benefits of quite general support provided to help with their lives, such as getting out and about (see Chapter 6, The value of person-centred care). This mirrored the narrative about the benefits of interactions from informal social networks (see Chapter 6, Theme 5: ‘the most critical person in my care’ – the role of informal support networks), which included a focus on the whole person and help with navigating the care system.
In supporting people with SMI and diabetes, offering continuity and opportunities to build relationships (see Chapter 6, Theme 8: person-centred care), and compensating for the lack of informal support we observed among many participants, were highly valued by service users and staff in the qualitative study, and also by the co-design participants as they discussed recommendations for future care. Challenges to this were also identified, including lack of time and continuity to provide person-centred care, and limited knowledge among staff about either diabetes or SMI, which limited their ability to provide specialist advice or care for both conditions. Poor information-sharing between different services was identified as a further barrier to accessing specialist care and navigating the health-care system (see Chapter 6, Theme 8: person-centred care), and, although the unreliability of referral data limited our investigation of diabetes and mental health care provision, we did find that frequency of care alone (e.g. more GP contact, health checks, hospital stays) was not indicative of better care or health outcomes (see Chapter 5, Objective 4: comparing the health outcomes of people with diabetes and severe mental illness with the health outcomes of people with diabetes without severe mental illness, Results, Objective 6: comparing diabetes care provision and estimating health-care costs for people with and people without severe mental illness, Results, and Objective 7: identifying health-care interventions associated with better outcomes for people with diabetes and severe mental illness, Results).
In the final co-design workshop, service users and health-care staff took forward the study findings to make recommendations that they believed would improve outcomes for people with SMI and diabetes. Although they worked independently, all groups proposed similar changes and agreed on three key priorities: (1) relationship-based care that focuses on management of both conditions, (2) the sharing of patient records between primary care, mental health care and diabetes services and (3) access to specialist SMI and diabetes expertise for staff supporting patients, to increase knowledge of both conditions.