Journal list menu
Evaluation of integrated diabetes and mental health care for type 1 diabetes and disordered eating (T1DE): the ComPASSION service
Abstract
Type 1 diabetes and disordered eating (T1DE) is often poorly understood and treatment is inadequate in many cases, with little or no co-ordination between physical and mental health teams. In 2019, NHS England launched two pilot programmes to increase understanding of T1DE, and to trial an integrated diabetes and mental health pathway for assessment, referral and treatment. This paper sets out key findings from the independent evaluation of one of these pilots, the ComPASSION service in Wessex, including baseline patient characteristics, clinical outcomes, and patient and staff experience. Among 11 patients with at least 12 months follow-up, the median change in HbA1c was a reduction of 31mmol/mol. Copyright © 2023 John Wiley & Sons.
Introduction
People with type 1 diabetes are at higher risk of eating disorders than the general population, partly because of the attention to diet and lifestyle required for diabetes management.1 It has been estimated that up to 30% of young people with type 1 diabetes have disturbed eating behaviour.2 The clinical consequences of eating disorders can be particularly severe in type 1 diabetes. Many patients with type 1 diabetes and disordered eating (T1DE) have very high HbA1c for prolonged periods, arising typically from insulin omission to control weight. Consequently, diabetes complications can arise at a young age. T1DE is associated with increased risk of microvascular complications including retinopathy and neuropathy.3-5 There are also short-term consequences including electrolyte imbalance, volume depletion, muscle atrophy and diabetic ketoacidosis.
T1DE is often poorly understood, and there is no consensus on diagnostic criteria or appropriate care. Psychological and medical treatment, where provided, often occur separately with little or no coordination. Psychological care is frequently unavailable.
In 2019, NHS England launched two pilot programmes, in Wessex and London, to increase understanding of the characteristics and care needs of people with T1DE, and to trial an integrated diabetes and mental health pathway for assessment, referral and treatment. Insight Health Economics was appointed by NHS England to conduct a formal evaluation of the programmes. This paper sets out key findings from the evaluation of the Wessex pilot, which was run as a multidisciplinary multi-site hub and spoke model providing both outpatient and inpatient care, and given the name ComPASSION. The hub was in Bournemouth, and the spokes in Dorchester and Portsmouth. Further details are given in supplementary material (available in Practical Diabetes online at https://wchh.onlinelibrary.wiley.com). Pilot oversight was provided by a project team with clinical and project management inputs.
The hub clinical team produced a screening tool and ran education sessions to increase clinical awareness of T1DE. Referred patients were assessed using diagnostic criteria developed for the service (provided in the supplementary material). Patients who met the diagnostic criteria were admitted to the service.
A description of the ComPASSION service, and of learning from the first six months, has previously been published.6, 7 Here we describe characteristics and outcomes for a patient cohort enrolled over 20 months, plus patient and staff experience.
Methods
Patient cohort, baseline and outcomes analysis
Between 1 March 2019 and 31 October 2020, 30 patients were assessed, of whom 24 were diagnosed with T1DE and enrolled in the service. Twenty-three consented to inclusion in the evaluation.
Demographic data (age, gender, ethnicity, deprivation) and baseline and follow-up clinical data (HbA1c, BMI and health-related quality of life [HRQOL] scores) were provided by the project team where available. Not all fields were completed for all patients; clinicians reported that COVID-19 impeded data collection, and it was challenging to administer questionnaires to such vulnerable patients.
Demographic and baseline characteristics were summarised using counts and descriptive statistics. Outcomes were analysed in relation to HbA1c and HRQOL scores. Baseline HbA1c values were compared with the most recent follow-up values, for patients with a follow-up reading at least 12 months after baseline. In supplementary analysis, baseline and follow-up HbA1c levels were compared for patients with a follow-up reading at least three months after baseline. In many cases, follow-up HbA1c values were estimates derived from glucose monitoring devices as blood samples could not be taken for laboratory analysis owing to COVID-19. (Studies indicate that approximately 72% of laboratory HbA1c readings are within 0.5% of the derived estimate, and 97% of readings are within 1% of the estimate.8) Wilcoxon signed-rank tests were used to test the significance of changes in HbA1c between baseline and follow-up.
Interviews and surveys
Eight semi-structured individual staff interviews were conducted. Staff were selected for interview to ensure representation from the hub and both spokes, from all professional groups involved in delivering the service, and from the project team. Interviews covered service design, implementation, perceived impacts, challenges and facilitators, experience and general learning. Three patient interviews were conducted, covering experience and impact. Owing to the vulnerability of the patient group, clinicians were asked to nominate patients for interview. All patients who agreed were interviewed. All interviews were conducted between July and November 2021, and were recorded and transcribed. Reflexive thematic analysis was conducted to identify and explore key themes, based on Braun and Clarke's six-phase process.9
Anonymous online surveys were conducted with clinical staff and patients. The staff questionnaire covered experience, skills and knowledge, impacts, challenges and facilitators. The project team was asked to send an email link to the questionnaire to all T1DE team clinicians, other clinical staff providing care to T1DE patients, and referring clinicians in June 2021.
The patient questionnaire asked about experience, perceptions and impact of the service. An email link to the questionnaire was emailed to all patients by the clinical team on discharge, after 12 months of care or in June 2021 (whichever came first).
Survey results were summarised with counts and the comments were collated and analysed alongside the interview data.
Survey questions are in the supplementary material available online.
Results
Baseline characteristics
Between 1 March 2019 and 31 October 2020, 11 patients were enrolled in the T1DE service at the Bournemouth hub, seven at Portsmouth and five at Dorchester.
Patient baseline characteristics are summarised in Table 1. Twenty-one of 23 patients were female, almost two-thirds of patients were aged ≤35 years. Almost three-fifths of patients had HbA1c ≥97mmol/mol at initial assessment. More than two-thirds of patients had BMI in the healthy range.
| Variable | No. | % | Range |
|---|---|---|---|
|
Age (years) – n=22 16–25 26–35 36–45 46+ |
9 5 5 3 |
41% 23% 23% 14% |
17–68 years |
|
Years since diabetes diagnosis – n=22 0 to 10 11 to 20 21+ |
11 7 4 |
50% 32% 18% |
0–62 years |
|
HbA1c (mmol/mol) – n=21 47–53 54–74 75–99 100–124 125+ |
2 3 6 7 3 |
10% 14% 29% 33% 14% |
47–145mmol/mol |
|
BMI (kg/m2) – n=19 Underweight (<18.5) Healthy (18.5–24.9) Overweight (25–29.9) Obese (30+) |
1 13 3 2 |
5% 68% 16% 11% |
18–35kg/m2 |
Ethnicity was recorded for 10 patients; all were White British. Indices of multiple deprivation quintiles are shown in Figure 1.
One or more complications of diabetes were recorded for five patients (22%). The most common complication was retinopathy (four patients). One or more comorbidities were reported for 13 patients (56%). The most common comorbidities were depression (four patients), anxiety (two) and osteoporosis (two).
Patients reported that they had an eating disorder for median 2.5 years (range 0–30 years, n=23). Seven patients (30%) reported previous treatment for disordered eating, including cognitive behavioural therapy, iTalk, counselling and (for one patient) multiple inpatient stays in eating disorder units.
- Diabetes Eating Problem Survey (DEPS-R).
- Eating Disorder Examination questionnaire (EDE-Q).
- Eating Disorder Quality of Life (EDQOL).
- Problem Areas in Diabetes (PAID).
- Generalised Anxiety Disorder Assessment (GAD-7).
- Patient Health Questionnaire 9 (PHQ-9).
- Work and Social Adjustment Scale (WSAS).
The number of questionnaires completed per instrument ranged from 8–15.
Four-fifths of DEPS-R scores were above the threshold used to indicate T1DE risk.10-12 Two-thirds of EDE-Q scores were clinically significant,13, 14 and 70% of EDQOL scores were indicative of severe eating disorder symptoms.15-17 Scores on the other instruments suggested relatively severe diabetes-specific emotional problems, functional impairment, anxiety and/or depression.
A summary of mean scores and ranges for these questionnaires is provided in the supplementary material available online.
Outcomes
Of the 23 patients enrolled between 1 March 2019 and 31 October 2020, 15 were still receiving care from the T1DE service on 31 July 2021, with mean treatment duration of 17.2 months. Nine patients were discharged by 31 July 2021 (of whom one was re-referred), with mean treatment duration of 11.9 months.
Median HbA1c reduction for 11 patients with follow-up ≥12 months was 31mmol/mol (p=0.01); (Table 2). For a further eight patients, HbA1c readings were available but follow-up was 3–11 months. The median estimated change in HbA1c for all 19 patients was a reduction of 9mmol/mol (p<0.01). Figure 2 shows change in HbA1c in relation to baseline for patients with ≥12 months follow-up.
| Patient cohort | No. | Change in HbA1c (mmol/mol) | Follow-up (months) | ||||
|---|---|---|---|---|---|---|---|
| Median | P-value | Mean | Range | Mean | Range | ||
| Follow-up ≥12 months | 11 | -31.0 | 0.01 | -25.6 | -78 to +4 | 19.8 |
12.7–28.0 |
| Follow-up ≥3 months | 19 | -9.0 | <0.01 | -16.3 | -78 to +5 | 14.3 |
3.6–28.0 |
Follow-up HRQOL scores were available for only a small number of patients (5–8 depending on the instrument) and are provided in the supplementary material available online.
Patient experience
Fifteen patients (65%) completed the evaluation survey. Results are summarised in Table 3.
| Strongly agree | Agree | Neither agree nor disagree | Disagree | Strongly disagree | |
|---|---|---|---|---|---|
| The Type 1 Diabetes/Eating Disorder team have helped me manage my insulin and/or my eating | 33% | 40% | 7% | 13% | 7% |
| The Type 1 Diabetes/Eating Disorder team have helped me manage my feelings and emotions around body image | 20% | 33% | 13% | 20% | 13% |
| The Type 1 Diabetes/Eating Disorder team have helped me manage my feelings and emotions around food | 47% | 33% | 7% | 7% | 7% |
| Overall I was satisfied with the treatment I received | 53% | 20% | 13% | 7% | 7% |
Eighty percent said they would be likely or extremely likely to recommend the T1DE service to a friend or family member who needed similar care.
When asked whether any additional elements would have made the service better, the components most frequently selected were better understanding from other health care professionals such as GPs and A&E teams (57% of responses), and out-of-hours support (50% of responses). Further details are provided in the supplementary material.
In comments on the survey and in interviews, patients spoke of the impact of the service: ‘This is just a lifesaver, I don't think I'd be sat here talking to you now if it wasn't for the service’; ‘The team have been outstanding, their commitment to helping me… has saved my life’.
Patients identified integrated care as a key factor: ‘It's really crucial that we integrate mental and physical health services together’; ‘The integration of the diabetes and the eating disorder is something I've been after for so long’; ‘I had never spoken to any health care professionals who talked through these issues with me. The ComPASSION team are brilliant at raising the topic and talking through difficulties’.
Some spoke of their own learning: ‘It has helped to notice the emotional patterns attached to my eating habits’; ‘It's given me the confidence to speak up when I'm struggling. Before, my blood sugar levels would run high and I would avoid it. Now I see that my blood sugar levels are high and I do something about it’; ‘The team provide you with all the tools and correct mindset to help overcome issues gradually over time. Ultimately the biggest change comes from yourself, but the team provide patience and consistent support until that point’.
Staff experience
Twenty-two clinical staff members in Wessex completed the staff survey (of 51 invited). Seventeen respondents had worked in a T1DE team.
Ninety-five percent of respondents reported increased confidence in supporting T1DE patients since the ComPASSION service began, with the proportion reporting considerable or moderate levels increasing from 9% to 86%; 91% of respondents reported an increase in their knowledge and understanding of T1DE, and 77% reported increased confidence in recognising patients with T1DE. Further details are provided in the supplementary material.
In interviews and in survey comments, staff reported increased skills, confidence and job satisfaction: ‘I feel really privileged to be part of this’; ‘It's been one of the highlights of my career’; ‘It's been a real joy’.
Diabetes and eating disorders staff spoke of a steep learning curve, and of gaining knowledge from each other and from patients: ‘Joint working has been huge in developing our understanding and improving our skill set’; ‘Patients taught us so much about the way diabetes can be sabotaged or manipulated… They have been so honest with us, teaching us… what to look for’.
Staff spoke of the emotional challenges they faced: ‘I really did struggle to start with. I lost a lot of confidence’; ‘It's unbelievably intense. It needs a lot of strength of character… the patient group really push the boundaries’.
Weekly multidisciplinary team meetings were identified as a key forum for learning as well as for planning care, while monthly clinical peer supervision sessions were seen as essential for sharing problems, gaining alternative perspectives and relieving pressure.
Discussion
While the ComPASSION service in Wessex is small in scale and patient numbers are low to date, this evaluation provides early learning on baseline characteristics of T1DE patients, diagnostic methods and the potential for multidisciplinary working across physical and mental health to improve clinical outcomes for this high-risk clinical group.
It is noteworthy that neither BMI nor HbA1c follow a consistent pattern in T1DE patients; a majority had healthy BMI, and not all had elevated HbA1c. This is because T1DE patients exhibit a range of different behaviours; some who restrict insulin may have a healthy BMI, and some who have restrictive eating but appropriate insulin use may have HbA1c at target level. Existing eating disorder questionnaires do not identify all those with T1DE. These observations may help explain the difficulties these patients often experience in accessing care.
ComPASSION has demonstrated that it is possible to provide effective care for people with T1DE through multidisciplinary working across physical and mental health, even in the context of COVID-19 which interrupted patient recruitment and impacted service provision. Improvements in the quality of care were achieved through collaboration, knowledge exchange and mutual support between diabetes and eating disorder teams.
Statistically significant reductions in HbA1c indicate the potential for services of this kind to improve clinical outcomes for a group of predominantly young patients at very high risk of early-onset diabetes complications. Previous studies have indicated that even short-term improvements in blood glucose control can have long-term impacts on the risk of complications.18 However, to understand the true impact of a service such as ComPASSION longer-term follow-up is required, and a larger study cohort.
ComPASSION has produced diagnostic criteria, treatment protocols, risk assessment documents and learning materials, and has established a network for knowledge exchange. It is hoped that these resources will provide a foundation for the development of similar services elsewhere. It will be important for these to be supported at local level by increased knowledge and understanding throughout the health care system, to ensure early identification of patients at risk, appropriate referrals and high-quality care after discharge.
Acknowledgements
Funding for ComPASSION and for the evaluation were provided by NHS England. ComPASSION team members provided data, interviews and administrative support for the evaluation.
Declaration of interests
KEY POINTS
- We present the findings of an independent evaluation of the ComPASSION service in Wessex, providing integrated diabetes and mental health care for T1DE
- Almost three-fifths of patients had HbA1c ≥97mmol/mol at initial assessment
- The median change in HbA1c for patients with at least 12 months’ follow-up was a reduction of 31mmol/mol
- In survey responses, 80% of patients said they would be likely or extremely likely to recommend the service to a friend or family member
- 95% of staff survey respondents reported increased confidence in supporting T1DE patients
