Structured, intensive education maximising engagement, motivation and long-term change for children and young people with diabetes: a cluster randomised controlled trial with integral process and economic evaluation the CASCADE study

Diabetic ketoacidosis

In T1D, diabetic ketoacidosis (DKA) results from either insufficient insulin or lack of insulin efficacy (e.g. during intercurrent illness). DKA carries a significant mortality rate (0.2%), largely from cerebral oedema and hypokalaemia. DKA is often present at diagnosis. Approximately 25% of all newly diagnosed children are admitted in DKA at diagnosis (35% in those of < 5 years). 6,7

Hypoglycaemia

The most common short-term complication of T1D is low BG (hypoglycaemia). Hypoglycaemia is a result of an imbalance of insulin to both ingested CHO and the current BG level. Hypoglycaemia (BG level of < 4.0 mmol/l) initially causes symptoms associated with activation of the sympathetic nervous system (i.e. adrenaline release). Severe hypoglycaemia (BG level of < 2.5 mmol/l) results in insufficient glucose for neuronal activity (neuroglycopenia), resulting in impaired consciousness, bizarre behaviour, seizures, coma and death. Persistent or frequent recurrent hypoglycaemia may impact on short- and long-term neurocognitive functioning. 8

Technological approaches

Advancing technologies such as the artificial pancreas are providing promising results in adults with diabetes. 15 Technological advances underpin increasingly sophisticated insulin delivery and self-monitoring devices. 15,16 Recent reviews report a wide range of current options. 17–22 Individual and group technology-based interventions include digital devices, video games, online chat rooms and social networks designed to engage, motivate, support and inform young people, although effect on glycaemic control remains modest. 23–30

The multidisciplinary team

Safe and effective diabetes care for children and young people requires a well-resourced multidisciplinary team (MDT) competent in integrating clinical, educational, dietetic, lifestyle, mental health and foot care aspects of diabetes. 31

Intensification of treatment supported by the MDT can produce dramatic improvements in HbA_1c. Poor control is common when children are asked to take responsibility for self-care without sufficient cognitive and social maturity. 32 Involving families,33 supporting parental monitoring34 and facilitating shared responsibility for disease management contribute to good glycaemic control and adherence. The challenge is to find patient-centred models of care that can be implemented in clinics and create similar improvements in HbA_1c level to those achieved in the DCCT, with subsequent reductions in the development of diabetes-related complications.

Reactions to a diagnosis of diabetes

Initial reaction to a diagnosis of diabetes can be devastating. A loss of spontaneity and restrictions on activities create a sense that things will never be the same. 45 Adjustment improves with time after diagnosis. However, parents were clear that they never fully ‘accept’ the diagnosis. 46,47 Episodes of grief are described 7 years post diagnosis triggered by regimen changes, injections, hospitalisation, discussions about diabetes control and worry about complications, attending clinics and meeting new medical teams – reminding them that their child is different. 48

Parents worry over the long term, uncertainty about the future, medical procedures and finance. Loss of pay and costs of travelling to appointments creates additional stress. 45 For some, balancing a career and managing the complexities and demands of a child’s treatment regimen becomes untenable.

Adherence to the treatment regimen

Adherence is a common problem in adolescents in general. 49,50 Although adherence to different components of the regimen may be unrelated to each other,51 insulin omission is common in adolescent girls who are concerned with body weight issues. 52–56 Failure to monitor BG or adjust insulin is also common. 57 Demographic, psychosocial and health-care system factors all influence adherence. Poorer adherence is reported in children and adolescents with T1D from ethnic minority backgrounds, of low socioeconomic status (SES) and from single-parent families. 58 Adolescence is also a risk factor for low adherence and poor glycaemic control. This is related to the effects of diabetes on major developmental changes that occur in adolescence and include hormonal changes associated with puberty, resulting in decreased insulin sensitivity negatively affecting BG metabolism, leading to increases in BG levels. 59 Alongside these physical factors, self-management, increasing independence, emerging sexuality and increased stress from peer and academic pressures are all associated with deteriorating glycaemic control in adolescence. 60

Healthier family functioning just after diagnosis and greater family support are associated with better adherence over time. 61–63 Good family communication patterns,64 low family conflict65,66 and good problem-solving67 are also associated with better adherence. Frustration and guilt at failure to achieve optimal outcomes can be exacerbated by existing psychological issues within families.

Psychosocial variables

High self-esteem, appropriate health beliefs (e.g. perceived threat of diabetes low and perceived benefits to cost ratio high)68–70 and the ability to cope with the stress of negative life events (ranging from a new diagnosis to daily hassles) predict better adherence. 58,71 Peer support is positively related to adherence, particularly for dietary and exercise behaviours. 72 However, pressure to conform to social situations and be accepted by peers is related to decreased glucose monitoring. 49,73,74 Young people who see themselves as having little internal control over their health and who attribute negative events to external sources (i.e. out of their control) have lower adherence. 75

When young people identify other people’s reactions to their self-care as negative, they are also more likely to have adherence difficulties. 72,76,77 Maladaptive coping, such as risk behaviours or withdrawal from family or peer interaction, is associated with poor adherence. 78,79

Managing the demands of a diabetes regimen

Stress management programmes using cognitive-restructuring and problem-solving strategies have limited impact. 80,81 In contrast, individual and group interventions focusing on peer group support, problem-solving and coping skills impact on short-term glycaemic control, showing positive effects on adherence including improved self-perception and increased knowledge about diabetes and decreases in diabetes-related conflict. 82,83 A problem-solving skills programme improved BG monitoring but failed to show improvements in problem-solving or HbA_1c level. 84 In contrast, social problem-solving, social skills training, cognitive–behaviour techniques and conflict resolution skills for children transferring to intensive diabetes regimens reported improvement in glycaemic control and QoL, maintained at 1-year follow-up. 85–87 The groups did not affect adverse outcomes of hypoglycaemia, DKA or weight gain in boys but decreased the incidence of weight gain and hypoglycaemia in girls.

Diabetes summer camps offering peer interaction, sports and recreational activities alongside diabetes self-management training are highly evaluated; however, only show short-term changes in self-management and knowledge and fail to demonstrate lasting improvements in HbA_1c level without ongoing support. 88,89

Focusing on psychological therapies

The effectiveness of psychological and psychoeducational interventions to improve adherence, glycaemic control, psychosocial functioning and QoL is debated, not only in T1D,90,91 but also in paediatric chronic illness in general. 92 Systematic reviews offer limited evidence that psychological interventions in children and young people with diabetes improve adherence and glycaemic control, with small to medium effects on physical and psychosocial outcomes. 91,93 Interventions that result in modest improvements focus on increasing knowledge/skills, addressing specific psychosocial issues and addressing self-management behaviours, but are often not sustained. Continuous support delivered by a diabetes MDT may be equally as important as short-term interventions. DeWit et al. 94,95 reported that asking young people about QoL during clinic visits improved QoL and satisfaction with care but had no effect on glycaemic control.

A recent meta-analysis of 15 randomised controlled trials (RCTs) of adherence interventions in 997 adolescents with T1D found a mean effect size of 0.11 (95% CI −0.01 to 0.23). 96 Modest improvements in glycaemic control were found in interventions that included emotional, social or family targets in addition to specific behavioural goals. 96 Interventions are most likely to be effective if they help the adolescent connect the many different aspects of diabetes management with all aspects of daily living. 97

Family interventions

Because of the complexity of diabetes management and the importance of family involvement and support, most interventions include a family component. In a review of family interventions, a positive effect was reported in 5 out of 19 studies, suggesting that family interventions may improve diabetes-related knowledge and glycaemic control. 98 A teamwork intervention increased family involvement and prevented an expected deterioration in glycaemic control. 98,99 However, the quality of family relationships may not be causally related to adherence, therefore decreasing family conflict and improving parent–child relationships alone may not result in improved adherence or glycaemic control.

Multisystemic therapy and an intensive home and community family intervention, incorporating developmentally appropriate negotiated responsibility,100 significantly improved adherence to BG testing, improved glycaemic control and decreased the number of inpatient admissions. 101,102

Behavioural family systems therapy (BFST) was developed for families of adolescents with clinically significant conduct-related problems. 103 The intervention includes cognitive restructuring of irrational beliefs and targets problematic family characteristics, family communication and problem-solving. Initial adaptations for children and families with diabetes addressed general developmental issues, such as managing curfews, chores and focused less on diabetes treatment-related treatment adherence or management. 104–106 BFST enhanced family communication, improved parent–adolescent relations, reduced behaviour problems and general and diabetes-related family conflict. 104–108 The effects on psychological adjustment depended on the adolescent’s age and gender but overall there was little effect on adjustment to diabetes or diabetic control. 104,105

A revised Behavioural Family Systems Therapy–Diabetes integrated empirically supported intervention strategies focusing on specific behaviours, as well as the social context of diabetes treatment-related behaviour including:

1. targeting at least two or more diabetes problems

2. behavioural contracting109

3. self-monitoring BG82,110

4. parental simulation of living with diabetes100

5. involving peers, siblings, and teachers, and running sessions in different locations.

Family conflict decreased as well as improvements in adherence. HbA_1c level was significantly reduced, particularly among adolescents with poor metabolic control. Change in treatment adherence correlated significantly with change in HbA_1c level at each follow-up. 105,106,111

These approaches show encouraging results; however, they are resource-intensive, requiring highly skilled practitioners who are able to travel into the family home. Although effective, this strategy is unlikely to be incorporated into UK clinical practice until shown to be clearly cost-effective.

Cognitive–behavioural therapy

Cognitive–behavioural therapy (CBT) can be delivered only by trained practitioners and usually requires 6–12 sessions. The client connects thoughts, feelings and behaviours. This requires a degree of engagement and participation that is often missing in adolescents. Although effective, standardised interventions for behaviour problems or depression in adolescents are available, no studies have targeted specific psychological disorders in adolescents with diabetes. A review found that the majority of studies used non-standardised approaches. 112

Motivational interviewing

Motivational interviewing is a person-centred therapy that works to resolve ambivalence about behaviour change. It creates a shared agenda, inviting the client to identify his/her goal (e.g. young people may say they want to talk about problems at school, whereas parents may want to talk about blood tests). The clinician negotiates how to explore these issues during the consultation. MI invites the client to be in charge of making changes, and looks for statements about intention to change and optimism about the future. The client identifies importance of change, confidence in their ability to change and when change will be a priority.

Motivational interviewing in paediatric populations has been increasingly explored. 113–115 Early work demonstrated the potential of MI to decrease harm-reduction behaviours in relation to alcohol and polysubstance use. 113,114 A review of nine RCTs in health-related domains, including diabetes, reported seven with positive findings on the effectiveness of MI. All RCTs that specifically addressed T1D related to the adolescent age group. 113,114,116 Two studies using MI alone117,118 found a significant reduction in HbA_1c level, reduced fear of hypoglycaemia, and improved QoL and positive well-being. MI combined with other therapeutic approaches, such as dietary advice or CBT, has reported reductions in HbA_1c level, a greater sense of control, improved perceptions of diabetes and improved self-esteem. 119–121

Attempts to incorporate MI into other formats, for example a six-session ‘personal trainer’ intervention delivered by non-clinical practitioners,122 have shown promising results, lowering HbA_1c levels in older teenagers.

One response to the scarcity of mental health resources in the UK has been to ‘skill up’ staff and introduce components of successful interventions into routine consultations. MI122,123 and other brief therapy approaches have the potential to offer diabetes teams ways to communicate with young people and encourage greater self-management. 124,125 However, a recent study that attempted to deliver MI components as part of routine consultations proved challenging to implement and had no impact on either HbA_1c level or psychosocial measures. 123–125

Solution-focused brief therapy

Solution-focused brief therapy (SFBT) relies on the clinician noticing what works successfully during conversations and doing more of this. 124 SFBT assumes the clients are the expert in their own situation and invites them to describe their preferred future and focus on what is already working. This creates opportunities to notice small changes and identify what makes this possible. 126 The clinician does not attempt to find the cause or ‘take the problem away’. The clinician and family work collaboratively to find already existing solutions to manage injections, finger-pricks or eating difficulties. They work together to stop diabetes getting in the way of family communication and find ways to manage sadness or anger. The family takes the lead in relation to ‘non-adherence’, focusing on what is working for them.

Solution-focused brief therapy is an effective approach for most, including those with severe and chronic problems. 127 The latest outcome evaluation research128 reviewed 109 SFBT studies including two meta-analyses and 19 RCTS. Nine RCTs found that SFBT has a greater effect than other approaches, for example CBT. Of the comparison studies, 34 out of 43 favoured SFBT. Solution-focused practice has been recommended within clinical practice guidelines. 120,129,130 SFBT combined with MI has been successfully integrated into diabetes clinical service,124 reducing HbA_1c level in children and adolescents. 121

Solution-focused conversations create a different experience for families. Young people who have been blamed or criticised and described as non-adherent or manipulative ‘grow visibly taller in their chair’ given the opportunity to talk about their strengths and abilities and describe positive steps they are already taking to get their lives back on track.

This approach is increasingly relevant to the management of long-term chronic illness and models of empowerment. Nurses trained in SFBT showed positive changes in their practice and improved willingness to communicate with troubled people. 124,131 Changes to practice centred on the rejection of problem-orientated discourses and reduced feelings of inadequacy and emotional stress.

SFBT techniques may be relevant to nursing and a useful, cost-effective approach to the training of communication skills . . . provides a framework and easily understood tool-kit that are harmonious with nursing values.

Bowles, Mackintosh and Torn131

Quality assurance

The programmes DAFNE and DESMOND have developed robust quality assurance processes;145,146 however, no standard template exists for paediatric and adolescent diabetes teams to assess patient knowledge and skills. The structure of quality assurance is crucial to ensure a rigorous process, with clear, written standards that can be monitored, regularly reviewed and updated. The quality assurance process include three main elements:

• A defined programme with clear content, structure, curriculum and underlying philosophy. The training programme for educators should be included within the quality assurance process.

• Quality assurance tools based on the programme structure, and a set of observable behaviours required to deliver the programme.

• Internal and external processes to assess delivery and programme organisation.

Developmental level and perceived learning needs

Children of < 12 years use concrete thinking to combine, separate, order and transform objects and actions. An eight-year-old relies on parents to help test BG, supervise injections or manage hypoglycaemia at school or out with friends. Educational material and conversations about health care need to be simple, concrete and involve shared goals with the parent and child.

Adolescents start to develop a sense of identity, increasing their need for independence with a move from parental to peer influence. Another change is the beginning of complex thinking processes (formal logical operations), including abstract thinking (thinking about possibilities), reasoning from known principles (form own new ideas or questions), the ability to consider many points of view according to varying criteria (compare or debate ideas or opinions) and the ability to think about the process of thinking. Adolescence is a transitional time with expectations of increasing responsibility for independent diabetes self-management.

Age and developmental status are powerful contextual variables that influence diabetes self-management. An evolutionary concept analysis identifies three attributes of diabetes self-management in children and adolescents. 148 The first is process, which is proactive, flexible, and involves a shift of shared responsibility between the child and family, and collaboration with health-care providers (HCPs). The second is performance of simple to complex activities related to the adjustment of regimens, including self-adjustment of insulin. The third is where children and parents engage in this process to accomplish certain goals. Successful completion of process, performance and goals are developmentally influenced.

The CASCADE intervention was designed to be accessible for children and young people between 8 and 16 years. Four modules lasting approximately 120 minutes each are delivered to groups of three to four families with children and young people aged 8–11 years or 12–16 years over 4 months. Everyone in the group is included in all discussions, and encouraged to share ideas and thoughts and develop their own solutions to their goals by evaluating decisions made in the past and think about possibilities for the future.

Interventions that incorporate parents are generally associated with favourable outcomes. 145,146,149–151 Young people are invited to bring parents or significant family members to CASCADE unlike other current structured education courses. 137

Learning methods

Families work in the large group or individual family or peer groups (young people and parents). Delivery is non-didactic. Families are invited to discuss diabetes information from a position of their own knowledge and expertise. Their right to choose different behaviours is acknowledged. The literacy level for written material was age appropriate with additional visual and written information (handouts, charts, diagrams, flow charts). Children, young people and parents are invited to consider attitudes towards changing self-care behaviours and complete exercises that look at the pros and cons of changing behaviour and assess readiness to change. It is more important that the HCP ‘hear and understand what the child has to say, than the child hears and understands what the healthcare worker is telling them’. 152

Measuring change using a competencies framework

The aim of patient education is for people with diabetes (or their carers) to improve and put into practice knowledge and skills with confidence, enabling them to become experts at managing their (or their child’s) diabetes on a day-to-day basis. HCPs need to collaborate with education specialists to formulate appropriate and reliable evaluation methods that can be used in addition to measuring changes in HbA_1c. 153

The CASCADE modules are informed by an eight-level competency system that assesses skills and knowledge designed by Children’s Hospital in Los Angeles ( Table 1 ). 154

Families must demonstrate a minimum of competency level 5 to start on pump therapy. The assessment, completed in a clinical interview, in conjunction with insulin pump therapy showed significant and sustained reductions in HbA_1c level over a 2-year period. 154

Underpinning philosophy and delivery techniques

Psychological approaches used in CASCADE are MI and SFBT (see Chapter 1 ). Specific components integrated into each module to enhance engagement, develop confidence and motivation are described below.

Key communication skills

Key skills in motivational conversations are asking open-ended questions, affirming, reflective listening and summarising. Educators use open questions to elicit information and broaden and deepen conversations. Affirming positive actions and intentions of young people and families focuses on positive outcomes. Educators use reflective listening and simple summaries to show that they are listening.

Identifying skills abilities and strengths

Solution-focused brief therapy assumes that people possess the knowledge and resources they need to tackle their problems. 126 At the beginning of each module, educators invite descriptions of what has gone well, listening out for examples of skills, abilities and resources. Children, young people and families note positive change and identify what they are doing differently. The aim is for both educators and families to have ‘determined curiosity and relentless optimism’.

‘How come?’ and ‘What else?’

Describing examples of when they have been in charge of diabetes since the last session orientates people towards what is already working and acknowledges when their preferred future is happening. Educators ask families what they had noticed had worked (‘What else?’) and ask what made the positive changes possible (‘How come?’).

Focusing on the future

Families imagine themselves in the future having overcome difficulties or problems and describe this problem-free preferred future. Families think about what they will be doing differently, or more, or less of, and how this will have changed their relationships with friends and family. ‘Stepping into the future’ invites families to identify ‘their’ goal for change.

Scaling

Scaling helps families think about how close they currently are to their preferred future and break this into small and achievable steps. Scaling reinforces positive change and prevents young people from feeling stuck when change is difficult to notice. Asking how someone had moved a half or a whole point towards a goal reinforces a sense of personal agency having taken a step forward (or prevented themselves from slipping backwards).

Considering the pros and cons of behaviour change

Motivational interviewing assumes that young people and families are, to a greater or lesser extent, ambivalent about current self-management behaviours and the degree to which they put pre-existing knowledge into action. The majority of young people with high HbA_1c levels will have had many conversations with diabetes consultants, nurses and dietitians, and possibly psychologists or counsellors. Uncertainty about implementing change can be expressed across emotional, practical and behavioural domains. In each module the young person and their parent/carer are invited to identify advantages and disadvantages of behaviours related to diabetes self-management (e.g. BG monitoring, using different insulin regimens, correcting for high or low BG) to help the educator understand the young person’s position and clarify the ‘pros and cons’ of change for the young person.

Establishing the importance and confidence of change

Importance and confidence are central to exploring the young person’s ambivalence to change. 156 Young people and parents scale (between 0 and 10) importance of change, confidence changing and how ready they feel they are to do things differently after completing a pros and cons exercise. The family identify what needs to be different for change to happen. If confidence is low the young person can think how to build confidence before making changes. If importance is low the group can think about what would need to happen for change to be more important.

Evocation not education: avoiding giving a choice or answer

The majority of young people with high HbA_1c are aware of what they ‘need’ to do but are not willing, able or ready to put this knowledge into practice. Educators ask if it is ‘OK’ to discuss information about different topics and invite the group to talk about existing knowledge first. These ‘scaffolding’ questions help individuals discover further information for themselves, or ‘draw it out’. 157 Additional learning is a collaborative effort between individuals and trainers, reducing the sense of an expert imposing knowledge and moving towards a shared venture. This active rather than passive approach has been shown to be effective in eliciting behaviour change in other areas. 158

Module 1: the relationship between food, insulin and blood glucose

Readings: (1) educator notes about CHO foods, BG and a healthy diet and (2) ISPAD guidelines on nutritional management in childhood and adolescent diabetes. 159 The introduction identifies strengths, resources and abilities. Focusing on the future, identify how things will be in the future if families get what they want from CASCADE. They scale how close they currently are to this. The session focuses on understanding food groups, particularly CHO. It talks about the role of insulin and different insulin regimens. Finally, families and young people consider the pros and cons of matching insulin to food to attain better glycaemic control.

Module 2: blood glucose testing

Readings: Implications of the DCCT13 and assessment and management of hypoglycaemia in children and adolescents. 160

After reviewing module 1, educators discuss the recommended target HbA_1c. The group identify factors that cause BG to rise and fall and explore individual hypoglycaemia definitions, reviewing symptoms according to severity. Families discuss ways to treat hypoglycaemia and assess the pros and cons of BG testing.

Module 3: adjusting insulin – pros and cons

Reading: insulin analogues in diabetes care,161 using CHO counting162 and guidelines on assessment and monitoring of glycaemic control. 163

Following the review families discuss symptoms of hyperglycaemia and ketones. A brainstorming session considers when, how and who to contact for help with managing hyperglycaemia. The session focuses on managing high BG using temporary insulin changes and explains how to correctly calculate correction doses. The group explores the advantages of identifying trends in relation to permanent insulin dose changes and considers the advantages and disadvantages of CHO counting as a way of improving glycaemic control.

Module 4: living with diabetes

Reading: guidelines on exercise164 with a set of PowerPoint slides used in the workshop to explain the principles of exercise and management of insulin and CHOs.

Following the review, families identify the effect that low and high BG has on performance and concentration, and discuss effective strategies to bring BG into target range before starting exercise. Participants work in family groups to identify how different activities affect BG and discuss the timing of insulin injections in relation to exercise before considering the advantages of using CHO before, during and after exercise (to keep their BG stable during different exercise).

At the end of module 4, young people and families complete a ‘blueprint for success’. This marks the end of the sessions and acknowledges the steps into the future the young person has already made. It creates an opportunity to review the programme and strengthens long-term motivation to change by reviewing previous successful goals.

Trained educators

The Department of Health and DUK have both highlighted the importance of structured training for educators involved in delivering patient education. Members of UK paediatric and adolescent diabetes teams have a varying amount of training in relation to age-specific and developmental milestones for children and adolescents with no current consensus regarding uniformity of roles and qualifications for different members of teams delivering structured education.

Each diabetes team in the CASCADE intervention arm identified two primary site educators, one of whom had to be a PDSN. The 2006 Royal College of Nursing (RCN) document on specialist nursing services for children and young people with diabetes states that a PDSN based in hospital or community, working as a member of the team specialising in the management of childhood diabetes, should be able to provide:

. . . a source of specialist advice for children, young people and families on the nursing care and management of diabetes, including the provision of basic dietary advice and the management of acute complications.

. . . individual specialist teaching for children, young people and families, facilitating the development of self care skills and knowledge, at time of diagnosis and in planned, ongoing, age appropriate education, both individually and in groups.

Royal College of Nursing165

The second trainer could be any HCP within the team. These two trainers were expected to deliver the groups to families within their service. Teams were invited to bring along additional members to the training if they so wished.

Pilot for intervention manual and training workshops

The intervention manual was piloted with a family known to the University College London Hospitals (UCLH) team. Content, delivery and resources were discussed to ensure that the principles and messages were clear and understandable. Training workshops were piloted with two UCLH PDSNs not involved in the intervention development, and five local PDSNs from centres not recruited to CASCADE. Feedback on the length of training (2 days), content and delivery was positive. Small changes were incorporated before training of the intervention sites began.

Training workshops

Key messages, based on the fundamental principles of CASCADE ran through all of the workshops:

1. Assume basic diabetic knowledge already provided at, and subsequent to, diagnosis.

2. Non-didactic educational principles and psychological models underpin behaviour change.

3. Not additional work for staff but working differently.

4. No right or wrong ways of saying things – just more or less helpful ways of talking.

5. Builds on families’ knowledge, skills and abilities, and uses their experience and expertise.

6. Assumes families have knowledge about how to manage diabetes and want to change their behaviours.

7. Use open questions; affirm and positively connote behaviours.

8. Reflect on what is heard and summarise to show good listening skills.

9. Explore the advantages and disadvantages of changing.

10. Focus on advantages of change and disadvantages of the status quo.

11. Encourage staff to be interested in the person rather than the problem.

12. Identify ways that people are already doing what they want to do by focusing on what is going well rather than what is not going well.

Training took place on two separate days with at least two intervention sites invited on each day. The following basic principles were emphasised:

• Follow the manual as much as possible to maintain treatment fidelity.

• Always have two people delivering each module.

• Divide the different components in the module in whatever way works best for site educators.

• Read key readings for each module to ensure knowledge of educational content.

Session 1 (morning of day 1)

The background to CASCADE was introduced with key delivery principles. These were to:

• Deliver the intervention to all recruited families in each intervention clinic.

• Offer four modules monthly to groups of three or four families with children in the same age group.

• Offer groups if only one or two families turn up for a module. If only one family turns up, reschedule or complete the module on their own and then join another group for future modules.

• If families miss a module they can continue with the next module and, where possible, be helped to catch up by discussing the module content with one of the site educators and/or being given the handouts from the module they missed.

The session introduced the basic principles of SFBT and MI and provided opportunities to try out the techniques embedded throughout the four modules.

Session 2 (afternoon of day 1)

This session introduced modules 1 and 2. Delivery and communication style was modelled by the UCLH trainers. Site educators participated in each exercise, giving answers or comments that they thought children might offer or the information they thought correct. The UCLH trainers worked together; one delivering the activities, while the second trainer commented on the process and performance, and invited discussion of content and delivery style. The trainers highlighted SFBT and MI techniques used in each activity.

Session 3 (morning of day 2)

Modules 3 and 4 were described using the same format as session 2.

Session 4 (afternoon of day 2): envisaging the future

Session 4 explored practicalities of running groups, organising sessions and how to engage and motivate participants to attend. After a brief slide presentation on groups, educators then worked in their clinic groups. They were asked to imagine themselves in 12 months’ time, having completed all of the CASCADE groups, and consider:

• What made completing CASCADE possible?

• What aspects of the process had gone well?

• What had been possible challenges and barriers that they had overcome?

They were invited to identify their skills, abilities and knowledge that contributed to this ‘ideal outcome’ making the groups possible. As each ‘outcome’ was described, trainers asked questions to generate details about solutions, such as ‘How did you achieve this?’ ‘Which days worked best for you?’ The shared ideas from each clinic generated practical and achievable solutions to potential problems in advance to support group organisation (e.g. in one training session good attendance was ‘achieved’ by staff phoning families a week before the start as a reminder). Other ideas highlighted during this session included:

• Enthusiasm for the project, successful recruitment and practicing sessions in front of peers for feedback.

• Resources in place for delivering the programme; finding an appropriate venue for running sessions with staff and time to plan, organise and deliver sessions.

• Good communication between educators, supportive management and other team members who would be receptive to the approaches and non-directive in communicating with young people.

• Sessions delivered at the right pace, working as a team with families with good attendance and age groups that bonded well. Involving everyone in the group with good use of techniques and listening skills.

• Positive feedback from young people and families. Families happy, more confident and accessing services appropriately.

• All HbA_1c levels of < 7.5%.

One-day refresher training

A 1-day refresher workshop was offered to all members of each intervention site when all workshops were finished. The morning session provided the training in the delivery process with a description of the four modules in the afternoon.

Ongoing support

Site educators could contact trainers at any time during the delivery of groups with queries or concern about timing of the groups, clarity in relation to aspects of the manual or what to do if a family did not turn up to training or composition of groups.

Trial objectives

The primary trial objectives were to determine whether a structured, intensive educational programme could be provided within a standard clinic setting for a diverse range of young people with T1D and to assess whether the CASCADE intervention improved long-term clinical outcomes (HbA_1c levels).

Secondary objectives included determining the impact of the intervention on other markers of diabetes control, psychological wellbeing and costs (see Outcomes, below). The logic model which informed evaluation of the intervention is shown in Figure 2 .

FIGURE 2.

CASCADE logic model.

Ethics committee approvals

The trial was performed in accordance with the recommendations of guiding physicians in biomedical research involving human participants adopted by the 18th World Medical Assembly, Helsinki, Finland, 1964, amended at the 59th World Medical Association General Assembly, Seoul, October 2008. The study was approved by the University College London (UCL)/UCLH REC reference number 07/HO714/112 (see Appendix 4 ). Subsequent amendments to the original ethics approval are detailed below. Site-specific approval was granted at each site.

Primary outcome

The primary trial outcome was glycaemic control, assessed at the individual level using venous HbA_1c value. Intravenous HbA_1c samples for patients and questionnaires from patients and carers were collected at baseline. The original protocol stated outcomes would be collected 12 and 24 months post intervention. A substantial amendment to collect follow-up data 12 and 24 months after the date of baseline blood sample was approved and reported in protocol v6 301010 (HTA progress report 5). Where possible, data were collected within 3 months either side of the expected collection date.

Secondary outcomes

Secondary outcomes were outcomes directly and indirectly related to diabetes management, including hypoglycaemic episodes and hospital admissions, choice of diabetes regimen, knowledge and skills associated with diabetes management, responsibility for diabetes management, compliance with intervention and clinic utilisation. Information about psychological functioning in terms of emotional and behavioural adjustment and general and diabetes-specific QoL was measured using well-validated, reliable self-report and parental measures.

Psychosocial outcomes

Validated instruments appropriate for age were used wherever possible. Parental and self-report of health-related QoL was measured using the Pediatric Quality of Life Inventory (PedsQL). 167 The scale has good internal consistency, reliability and validity for both the generic and diabetes modules. The physical psychosocial health summary score produce an overall total score. The T1D module has four scales: Treatment I and II, Worry, and Communication, which combine to produce a total diabetes score.

The five-item ‘Impact Supplement’ of the Strengths and Difficulties Questionnaire (SDQ) (parent and child report versions) assessed the impact of identified emotional and behavioural difficulties on the young person’s life. 168

Young people were also asked to rate how happy they were with current weight and if they had skipped insulin using two self-report questions designed specifically for the study.

Diabetes outcomes directly related to patient management (parental report and case note review)

• Diabetes regimen (insulin dose/number of injections/insulin types).

• Hypoglycaemic episodes (frequency, severity) in last 6 months.

• Admissions to hospital in last 6 months and reason (e.g. episodes of ketoacidosis, hypoglycaemia).

Diabetes outcomes indirectly related to patient management

• Knowledge and skills associated with diabetes. 155

• The Diabetes Family Responsibility Questionnaire (DFRQ)169 (parent and child report) reflects the sharing of diabetes-related responsibilities (such as deciding what to eat at meals and snacks, rotating injection sites and telling teachers about diabetes) within families. Parent and child versions of the questionnaire are scored separately. Higher responses (on both) indicate greater responsibility taken by the young person for their care. A dyadic score for the parent–child pair reflects agreement about who is taking responsibility. High dyadic scores when neither reports taking responsibility is linked to poor health outcomes.

• The Diabetes Self-Efficacy Scale assesses confidence managing diabetes-related tasks. 170

Compliance with intervention/control

• Attendance at intervention sessions.

• Service utilisation rate.

• Clinic attendance.

• Contacts with PDSNs and diabetes teams.

Data were collected through process evaluation (PE) and case note review (CNR).

Service user perspectives

Service user perspectives were sought on the structure and content of all the questionnaires. Initial questionnaires were reviewed by a young woman with diabetes as part of the trial management group. One individual with research responsibilities for a diabetes charity and the mother of a girl with diabetes who was also research trained provided comments about the length, general tone and specific items on the questionnaire. Early data collection proved problematic as questionnaires were too long and were often only partially completed. After consulting a patient organisation and a carer researcher, the self-efficacy scale was removed and questions assessing knowledge and skills about self-management were removed from the 8- to 12-year-old child version.

Primary outcome data collection

Arrangement for collection varied widely between and within sites. Research nurses or a member of the local clinical team trained in venesection took the sample in the paediatric outpatient clinic or occasionally at the patient’s home or GP surgery. Hospital phlebotomists also collected samples. Approximately 4 ml of blood was collected and stored in a tube prefilled with diluent and preservative. The samples were sent to a single UK laboratory (UCLH Special Biochemistry, London) for measurement of HbA_1c concentrations. This ensured direct comparability of results from all clinics. Current legislation requires that non-infectious diagnostic specimens are transported in accordance with guidelines UN 3373 and need to conform to secure packing instructions P650 (The European Agreement concerning the International Carriage of Dangerous Goods by Road; Regulations concerning the International Carriage of Dangerous Goods by Rail; the International Air Transport Association). The research team provided the appropriate packaging. Samples were transported using a courier or more often delivered by hand by researchers or via the Royal Mail (identified as a biological substance, category B). HbA_1c assays were carried out on a Tosoh G8 instrument (Tosoh Bioscience, Tessenderlo, Belgium) with International Federation of Clinical Chemistry aligned calibrators and results reported directly to the data manager following adjustment against the DCCT international standard.

When a sample was lost or spoiled in transit, the research team used their discretion with regard to requesting a repeat sample. A small number of young people were recruited in whom it was subsequently found they had variant haemoglobin, meaning that HbA_1c levels could not be accurately measured. This issue came to light when the UCLH laboratory reported abnormal results for baseline blood samples. The clinic principal investigator (PI) was informed. These patients remained in the study and provided completed questionnaires but their HbA_1c data were not used and no further study blood samples taken.

Secondary outcome data collection

For collection of secondary outcome data the researcher provided patients and carers with a copy of the appropriate questionnaire. Participants were encouraged to self-complete the questionnaires independently (i.e. parents and children were asked not to confer). The researcher provided help when required. If a young person and/or their carer was not able to complete the questionnaire in the clinic, either owing to insufficient time or the parent not being present, the researcher would go through the questionnaire with the young person/parent on the telephone. Patients recruited by site staff or research network nurses were contacted by telephone to complete the questionnaire.

The PI from each clinic ensured appropriate access to advice and psychological support should any participant express any concerns or worries with regard to taking part in research activities. In exceptional circumstances, for example disclosures of significant risk (to self or others), feedback to the family/carer and clinical team would be made by the research team. Participants were reminded that DUK provides an independent voice of support, through ‘Careline’, a dedicated helpline for all people with diabetes, friends, family, carers and HCPs.

Eligibility criteria for clusters

• NHS paediatric diabetes service – clinic had to be staffed by at least one paediatrician and paediatric nurse with an interest in diabetes.

• Located within London, the south-east (SE) and the Midlands (maximum train journey time 2 hours).

• Not running a group education programme at time of recruitment or during trial.

• Not taken part in a similar paediatric diabetes trial within the last 12 months.

• At least 40 eligible patients on current patient list.

Inclusion criteria for patients

• Diagnosis of T1D with duration ≥ 12 months.

• Aged 8–16 years.

• Mean 12-month HbA_1c value ≥ 8.5 mmol/l.

• Under the care of a paediatric and/or adolescent diabetes clinic conducted by a specialist, or general paediatrician with an interest in diabetes.

Exclusion criteria for patients

• Significant mental health problems unrelated to diabetes requiring specific mental health treatment.

• Significant other chronic illness in addition to diabetes that may confound the results of the intervention. Patients with coeliac disease or hyperthyroidism were included.

• Significant learning disability or insufficient command of English to enable full participation in the planned intervention. Young people with good command of English but whose parents have poor command of English would be eligible to attend by themselves if they wished as long as parents had given informed consent. Another relative who was one of the primary diabetes carers (e.g. sibling, aunt or uncle) who had good command of English could participate instead of the parents.

• Participated in diabetes treatment trials in the 12 months prior to collection of baseline data.

Additional strategies to support recruitment process

Five additional sessional workers were used to support recruitment on an ‘as required’ basis to increase capacity and maximize flexibility within the team.

Where possible clinics were asked to cluster eligible patients into specific clinics to minimise the number of trips researchers had to make. Reimbursement was offered to families for any expenses incurred through travel for research purposes.

The study was accepted as part of the NIHR portfolio and adopted by The Medicines for Children Research Network (MCRN) and the Diabetes Research Network (DRN). Research nurses from MCRN Trent and MCRN West Midlands Local Research Networks supported recruitment in clinics that fell within their geographical localities. Nurses connected to the London South East, North Central and East Research Network supported recruitment in a number of London clinics.

Protocol amendments were approved to allow local clinicians to recruit and take consent from young people and their families. For example, £100 was paid to the trust for each patient recruited. Baseline questionnaires were distributed and subsequent data collected over the telephone for patients recruited by local clinicians (progress report 1, 30 October 2008). Research nurses and site staff were trained in study procedures particularly with regard to minimising risk with respect to allocation. All patient recruitment took place on hospital sites with the exception of one site where a member of the clinical team recruited a patient at home. Ongoing support for staff assisting with the recruitment was provided by e-mail, phone and face-to-face contact. A DVD was made to support recruitment in clinics. A copy of the DVD was provided to each clinic and the file was also uploaded to YouTube (www.youtube.com/watch?v = XFgVKBYdCbU). A substantial amendment was submitted to allow the DVD to be used by site staff to support recruitment (progress report 2, 30 April 2009).

A regular newsletter was sent to clinics to maintain motivation to recruit.

Aim

The PE was designed to assess the feasibility and acceptability of delivering CASCADE groups as part of standard care within an NHS setting. Should the CASCADE intervention fail to be effective, the PE was also designed to assess the extent to which theory or implementation was responsible.

Specific aims of the PE were to:

1. compare and contrast standard NHS paediatric/adolescent diabetes service across control and intervention sites

2. describe the UCLH training workshops for the CASCADE intervention and assess quality, fidelity and satisfaction

3. report on feasibility and acceptability for staff organising and delivering CASCADE groups in the sites

4. review feasibility and acceptability of the CASCADE intervention for families

5. describe parent/carer and young people’s perceptions of impact of participation in CASCADE on themselves and their families and views of staff regarding the impact on young people and on their own practice.

Design of the process evaluation

A mixed-methods approach to data collection was used. Data were collected from key stakeholder groups (UCLH trainers, site staff, young people and parents) at specific time points in the trial. Questionnaires, observation and CNR were used to collect qualitative and quantitative data. Qualitative data were collected through semistructured interviews, informal discussion following observation sessions and fieldwork notes. Participants were informed that all information would be confidential with only the research team able to access individual data. All data used would be anonymised. Permission to record interviews was sought. Interviews were transcribed verbatim. If permission to record was not granted then notes were taken by the interviewer during or immediately after the interview. All PE instruments used for data collection are available on request.

University College London Hospitals training workshops

CASCADE groups

Interviews with site staff

Researchers interviewed one site educator from each intervention site and a key member of staff who had contact with young people in the control clinics (usually the lead nurse), as well as other key stakeholders working within the clinics. A purposive sampling approach ensured that those expected to have key information were included. Interviews were conducted either in person or over the telephone with the exception of the structured questions at the beginning, which could be e-mailed and completed in advance of the main interview. Interviews were conducted after all groups had been completed in the intervention sites. Control site interviews lasted about 30 minutes, and about 60 minutes for intervention sites. The schedule included structured and semistructured questions about service structure, current challenges regarding diabetes care, and current training/education. For the intervention sites semistructured questions were asked about the training and delivery of the CASCADE intervention.

Interviews with young people and parents

The main aim was to gather in-depth information on how young people and their families look after their diabetes, and the education they had been offered and/or received about diabetes. Control site interviews focused on standard practice, whereas in the intervention sites they included experience of the CASCADE modules.

The original protocol stated that these data would be collected from focus groups. Given the challenges experienced in the study with accessing young people at time points additional to routine clinic appointments, a decision was made to use individual interviews with young people and their carers instead. This minimised inconvenience for all stakeholders as interviews could be conducted in the clinic setting at the same time as follow-up data were collected or, alternatively, over the telephone. Interviews with up to four young people in each intervention site and two in each control site (plus parents) were planned. Saturation point (the point at which no new information was given) was reached after 30 intervention and 18 control site interviews were completed therefore target numbers were reduced to two and one per site respectively. To ensure a range of key characteristics within the sample, interviewees in each site were purposively selected using a sampling grid constructed for each site ( Box 1 ). The aim was to achieve diversity across as many characteristics as possible in each site and across the two arms of the trial.

Individuals were invited to take part in an interview by a member of the research team, with parents/carers invited to take part once the young person consented. Young people were interviewed even if parents/carers declined.

Semistructured interview schedules used learning from data collected in the study and lasted approximately 20 minutes. They were piloted with a few young people and then further refined. All children and young people were asked about their experience and management of diabetes. Intervention site informants were also asked about their experience of CASCADE and any perceived impact. Those who had opted out of sessions were asked about their reasons for opting out. Carers were asked about views of their child’s management of diabetes, and, in the intervention sites, experience of CASCADE.

In control sites, interviews were completed at the time of first follow-up data collection. In intervention sites, interviews were carried out within 3 months of the young person completing groups. Most participants were invited to take part by letter that explained the main aim of the interview, and informed that they would receive a £10 voucher as a ‘thank you’ and compensation for their time.

Five researchers carried out most interviews face to face. Parent and young person interviews were designed to be conducted separately but a request to do the interview together was acceptable. Some were carried out over the phone when a face-to-face encounter was difficult to arrange. If a selected participant declined to be interviewed then a replacement participant was selected from the sampling grid.

Questions in secondary outcome measure questionnaires

A small number of PE questions were incorporated in the follow-up questionnaires for intervention families. This allowed data capture from the whole intervention sample (as opposed to just those interviewed). In the 12-month follow-up questionnaires the focus was on experience and perceived impact of the CASCADE groups. Those who had declined to take up the groups were asked to give reasons for this. In the 24-month follow-up questionnaires, questions were based on interview data from young people and carers focusing particularly on views of sustained perceived impact.

Field notes

Detailed field notes were kept during the 4 years of data collection.

Case note review

A CNR was conducted at each site around the time of the participants’ 24-month follow-up appointment. CNR for withdrawn patients was completed up to when they last provided data for the study. The same data were collected for participants at both control and intervention sites and entered on to a secure database with controlled access and no patient identifiable information. Data were used to describe standard care, specifically the level of nursing contacts in both control and intervention sites. Data were compared using Microsoft Excel version 2003, version 11 (Microsoft Corporation, Redmond, WA, USA).

Analysis

For large quantities of qualitative data analysis was supported by the use of NVivo software version 9 (QSR International, Southport, UK). Where the quantity was small, it was done by hand. Qualitative analysis of the interview transcripts identified key topics and issues that emerged from the data through familiarisation with transcripts. 179 Pertinent excerpts were coded and memos written to summarise and synthesise emerging themes. Researchers refined their analysis ensuring that themes were crosschecked with other data, first within and then between transcripts so that the validity of emerging explanations were tested and improved so themes were ordered within an analytical framework based on the data. To maximise the validity of the findings, a sample of the qualitative data were separately analysed and reviewed to agree a framework for full analysis of all data.

Quantitative data were analysed using an Excel spreadsheet and the Statistical Package for the Social Sciences (SPSS) version 19 software (SPSS Inc., Chicago, IL, USA) for statistical tests.

Composite fidelity delivery scores from individual observer and self-rated scores were created for each pro forma for content and for technique. This allowed the range and median score for content and technique to be calculated. A further composite variable was then calculated, which summed the content and technique scores for each site across all four modules. If any scores were missing for a module the total scores would be divided by the number of completed modules. This allowed comparison across sites and modules, as well as comparing the individual influence of content and techniques.

Analysis of the UCLH training workshops observation was carried out by a separate researcher reading through the notes made by the observers and identifying key themes and fidelity issues emerging from the data.

Characteristics of the clinics

Clinic-level characteristics clinics are shown in Table 2 . The two groups were similar at trial entry, with all clinics mixed with regards to age – 86% in each arm were general hospitals with similar numbers of eligible patients in each arm.

Baseline characteristics of the participants

Demographic information on 327 children and young people and 324 parents who completed the baseline questionnaire are shown in Tables 3 and 4 . Groups were similar at baseline, with a mean (child) age of 13 years (SD 2.1 years), and time since diagnosis of diabetes of 5.1 years in the intervention arm and 5.6 years in the control arm.

Table 5 shows the numbers providing data at each point of the analysis. Main reasons for loss to follow-up included an inability to contact participants at follow-up points (attempted directly and with assistance of clinic staff), and withdrawal (1) as moving away, (2) for reported reasons including needle phobia or pressure of GCSEs, and (3) for unexplained reasons. Some participants unavailable at follow-up 1, were re-engaged for data collection at follow-up 2. Table 5 shows some participants provided one type of follow-up data (blood sample or questionnaire) but not the other. Participants in the intervention arm who changed their mind about taking part in the CASCADE groups were informed that they were welcome to remain in the study despite choosing to opt out of the intervention.

Missing data

As there was no differential attrition by arm and no reason to believe the data were missing not at random, an intention-to-treat analysis based on randomised participants who provided primary outcome data was undertaken for all outcomes.

Secondary young person outcomes

The secondary outcomes as reported by the participants at 12 months and 24 months are shown in Tables 11 and 12 . They are similar between the two arms of the trial. At 24 months the intervention group has a higher DFRQ scores, which indicates that child/young person is taking more responsibility for self-management. There is also reduced self-reported happiness in body weight. However, it should be noted that with a 5% level of significance this number of significant findings is consistent with chance.

Changes in knowledge and diabetes regimen were additional secondary outcomes. These data proved difficult to collect with accuracy. We present the results here ( Table 13 ), but these data should be interpreted with caution, as they are unadjusted for baseline and have a large number of missing participants. Young people showed no increase in knowledge at either follow-up.

Secondary parent/carer outcomes

Secondary outcomes reported by the parent/carer are shown in Tables 14 and 15 . They are similar between the two arms of the trial. At 12 months, the carer reports an increase in the family responsibility score, which indicates the child/young person is taking more responsibility for self-management of diabetes. As stated this significant finding is consistent with chance.

The decision to take part in the trial

Staff interviewed were mainly nurses. They had become involved in CASCADE mainly because of a consultant’s interest in being part of the study. About half of them remembered it being a collaborative team decision to take part:

Dr X said ‘I think we should do this, what do you think?’ And I remember two ladies coming down [presentation from the research team]. It sounded interesting, so we agreed to do it.

Nurse I11

For a significant minority, involvement was not something they had actively chosen, either because the current consultant, with little or no discussion, had instigated it or a previous consultant or other member of staff who was no longer part of the diabetes team had agreed to take part:

By that point I knew [my predecessor] wasn’t gonna come back but I felt that I couldn’t stop the process because it had gone too far and I’d sort of inherited it all and I just thought, ‘oh my goodness, how am I gonna cope with all of this?’

Nurse I4

Views on becoming involved in CASCADE

Becoming involved in CASCADE was an exciting prospect for many. They saw it as an opportunity to develop aspects of their service they wanted and would be expected to offer in the future.

I was quite keen because we’d had to stop giving the diet sessions because we hadn’t got . . . enough dietetic time and I just thought it would be great to have some structured education . . . and I thought it was something that we could be involved in and hopefully give something to benefit the kids.

Nurse I13

Many, however, including those who were generally enthusiastic about being involved, expressed concern they might not have the time or resources to do what would be required. Nurses were unsure whether consultants who had agreed to be part of the study were aware of the extent of the work involved.

I was a bit anxious at the time because I felt that the consultant didn’t truly understand who did what [in CASCADE] and I think she believed that everybody [from the central team] was gonna come in and deliver the sessions which wasn’t the thing so that made me a little bit anxious.

Nurse I1

Feelings about the outcome of the site randomisation process

Staff had varied reactions to which arm their site was randomised to. About half in the intervention arm were pleased. An equal number were anxious and concerned often owing to the delay that resulted from the extended patient recruitment process, which meant that the timing was no longer good for their site primarily because of lack of resources or staff shortages.

Horrified, well, because of the time delay and by the time we were starting it was just not a good time for me, so yes, I was horrified.

Nurse I1

Well, initially I think we were a little anxious because at the time we were having staff shortages and we think from a practical point of view it would have been better to have been in the other arm but it didn’t matter . . . we were excited to be in it anyway.

Nurse I5

Those randomised to the control arm were also evenly split between relief and disappointment at the zoutcome.

I was quite relieved because having to do the education [would have been] really difficult for me . . . actually carrying that out without any support [would have been] very difficult.

Nurse C2

We’d have liked to have been in the intervention arm partly because we were hoping that people would like to see what we could do to change our practice to improve the teenagers, particularly 12 and upwards [when they] start to get more independent and not wanting to comply.

C8

Attendance at the training

Workshops 1 and 2 were run on six occasions each between 18 August 2009 and 21 May 2010 at UCLH. The UCLH trainers also travelled to deliver training to eight staff from two hospitals (in one site) over 2 days. The remaining workshops 1 and 2 were run between 1 and 7 weeks apart, with the majority ≤ 2 weeks apart. Each site sent the required minimum of two staff, with some sites sending more than two. A total of 43 staff attended. A few sites found it challenging to ‘free up’ staff to attend. Seven additional staff from intervention sites attended the 1-day refresher workshop. The numbers of participants at each workshop ranged from two to eleven. The number of clinics at workshops ranged from one to four ( Table 20 ).

The majority attending were PDSNs. Dietitians from 10 intervention sites attended, eight of whom were intending to be the second educator. One site, unable to release a second member of the diabetes team for training, sent a research nurse based at the site. One consultant attended both workshops. Three consultants attended one workshop or the 1-day refresher. Two nurses who attended the initial workshops also attended the 1-day workshop.

Fidelity across training workshops

Observation data demonstrated a high degree of fidelity across both workshops. The same trainers delivered all workshops and maintained a high degree of consistency in both content and style. The training approach was highly interactive with regular opportunities to clarify concerns in question and answer sessions.

Satisfaction and acceptability of training

The manual and resources

Site educators were positive about the manual and the resources. They found the manual easy to follow with useful prompts to promote wider discussion.

Numbers of groups

A total of 30 groups were run across 14 intervention sites. The modal number of completed groups was two, with a range of zero to six groups offered by clinics ( Table 21 ). Twelve sites delivered at least one complete group of four modules. One site only offered the first session and one site failed to offer any groups.

Sites were not given a formal target number of groups but were aware that all recruited participants should be offered the opportunity to attend a group. A post hoc calculation suggested that 44 groups should have been run across the 14 intervention sites. Only three clinics completed the maximum number of groups possible ( Table 22 ).

Size of groups

The aim was to deliver a group to three to four young people of a similar age. This was not achieved in most modules ( Table 23 ). The attendance ranged from one to five young people with a mode of two.

Time between modules

The average interval between modules was slightly greater than 1 month ( Table 24 ) though the average length across all groups (from start to finish) was approximately 4 months. The main reason for delays was problems organising dates, as well as site educators being on sick leave or study leave.

Time between training and delivery

As soon as site educators completed the UCLH training they were free to start preparing the groups. On average, sites ran the first CASCADE module 3 months after completing the UCLH training ( Table 25 ) with module 4 completed 7 months after training.

The training was very good, the only criticism is between training and then actually getting up and running was about 6 months which is . . . I think the training needs to be done a month before you go and launch your site.

Site educator I12

A key reason for delay setting up the first groups was delay in patient recruitment. Training began in August 2009; however, patient recruitment was not completed until September 2010. Earlier trained clinics often had to wait several months before they could start to organise and deliver groups. Many site educators cleared their diaries in anticipation of delivering groups; however, the delay meant that they were unable to sustain this with other commitments encroaching on the time required for the project. Where there were enough patients to make up a group, organising and delivering groups started before patient recruitment was completed. For sites with a low number of recruited patients this strategy prevented groups being organised by age.

Time between sets of groups

Where sites had sufficient participants groups were mostly run concurrently, organised by age, at different times to suit particular families. One site had a very long gap between running their first set of groups and their second owing to a site educator being on sick leave.

Days and time of CASCADE modules

Groups were run on a full range of weekdays. They were offered within school time (morning or afternoon), immediately after school (e.g. 1600), and evening (e.g. 1800). The most frequent time for running modules was the early afternoon ( Table 26 ). Saturdays and holidays did not appear to appeal to participants more than weekdays. One attempt at running a module on a Saturday ended when no participants arrived.

Preparation for running groups

In the majority of sites a single site educator did the bulk of the preparatory work particularly when the second trainer had limited time allocated to paediatric diabetes (e.g. dietitians and mental health practitioners). Preparation involved deciding which participants should be grouped together, setting dates and times and booking rooms. This was felt to be particularly burdensome in sites where there was just one PDSN.

Initial decisions on how to group participants were considered straightforward, except in sites with small numbers of recruited patients. Where numbers were small, compromises were made about grouping by age. For sites with larger recruited numbers this compromise was usually made as they reached the last groups and the remaining pool of participants was small.

Setting dates and times was the most challenging aspect reported. The initial challenge was anticipating and/or determining time preferences for participants. Approaches used included contacting families to ask for individual timing preferences and offering a range of different days. Some staff were prepared to work outside normal working hours – including evenings and weekends; others were not.

The first one I think we held at about six in the evening having spoken to the families first and that was the preferred time for them and then we negotiated it as we went along. The first group was more flexible and were prepared to come in . . . let’s say less outside hours time. The second group was a nightmare because . . . there was no [flexibility]. ‘We want it at this time and I’m giving up this to get here’ and there was no concession at all that [my colleague] and I were giving up our [personal activities].

Nurse I11

Other sites made calculated guesses about the best times for holding the modules.

We did the sessions in the morning and I think the reason for that was really I didn’t feel they’d want to, or be able to, take on that after school when they’re tired.

Nurse I2

I think we did 6.30 to 8.30 [pm] for all our sessions, we thought it was a good choice because we didn’t want the young people to be missing school.

Nurse I8

The process of trying to reach an acceptable date was considered to be laborious and time-consuming. Site educators communicated via a combination of letter, telephone, and (occasionally) text. No site used e-mail or web-based diary systems to help organise dates. For most sites agreeing a date involved a series of negotiations to ensure that the greatest number of potential participants could attend.

We tried, we sent out an initial letter with a date like say, for example, the 8th of July as so those that didn’t respond I rang them again and said ‘We haven’t heard from you’ and they’d say ‘I’m on holiday that day’. So then I offered them an alternative date the following week and some of them were offered three or four opportunities to come before we abandoned them. So everybody was offered a selection of dates. It wasn’t through want of trying, it was a shame really.

Nurse I14

Group 2 was the problem. We probably offered something like August, October, November and December . . . really the people couldn’t make any of those dates.

Nurse I13

Following the setting of the date and time, most sites sent out a standard confirmation letter (provided by UCLH). Some sent a reminder letter (or made a reminder telephone call) prior to the date, which improved attendance. Teams that did not send a reminder expressed regret that they had not, thought it should not have been required or did so after early groups failed.

A common challenge was participants not attending agreed dates. When attendance was poor, site educators either cancelled or rescheduled to achieve better take-up or ran the session with one family who were then invited to join remaining modules of a different group. One of the sites that ran three complete groups offered seven module 1, five module 2, four module 3 and three module 4. Only one of the three groups ran with the same participants in (or at least invited to) all four sessions.

The majority of site educators began with great enthusiasm and goodwill trying to accommodate everyone but became more pragmatic about what was achievable over time. Additional administrative support would have made the task considerably easier for many sites, especially where site educators had little administrative experience and had to do this work in addition to normal work responsibilities. No site had any significant level of administrative support for organisation and running of groups.

Oh, it would have been very useful, to have somebody else running around making sure people could make certain dates . . . reminding them, cause one of the issues we had was that people claimed the letters had never gone out so I then started reminding everybody a week before.

Nurse I11

Staffing constraints

Staff turnover, absence and lack of flexibility around work hours also influenced the number of groups organised. In two sites a site educator left while groups were being delivered. One site found a replacement who had attended the refresher workshop and the groups recommenced relatively smoothly. The other site was unable to find a replacement and completed only one group (with the original educators), cancelling the second planned group. One site did not run any modules. The lead trainer went on extended sick leave and on return was deployed into a different role. In another site, educators ran module 1, twice in the morning, on a specific day of the week – inviting the same four families on each occasion. On the first date only one young person and parent turned up, on the second the previous attendee and parent returned along with one other participant and parent. Neither returned for module 2. No further groups were offered owing to the small number of recruited patients in this site.

The added burden of the research context

Problems setting up groups were exacerbated by CASCADE being a research project, not just an educational intervention. Staff felt pressured by additional trial-related tasks, such as organising research blood samples and liaison with research staff.

Restricting groups to a subset of recruited patients, instead of offering groups to the entire clinic list was perceived as making the organisation of the groups more challenging. This was particularly the case in sites with a small number of recruits and/or where a hospital served a large geographical area making travel arrangements complex. It also meant that natural groupings of patients (by age, or geographical area, etc.) were more difficult (and at times not possible). Site educators were aware that they would be observed delivering some of the modules which added to their concern about achieving reasonable group numbers.

Use of time

The theory behind use of time was explained in the training workshops:

How do we find the time to organise and run the sessions?

CASCADE is about organising time differently, not taking more time. For example, a home visit might take an hour and nothing changes afterwards. But, if you run a CASCADE session, you can say something to a group of children that you otherwise might have to say on an individual basis.

However, a great deal of frustration was expressed about the time required to organise CASCADE. In many sites complications around organising the modules took up any time that site educators had allocated to practice and preparation of the programme.

It suddenly hit me about how much commitment and organisation it was going to take and I do remember saying to [a colleague] ‘Gosh, I don’t know how we’re going to fit this in with all the other things that we’ve got!’.

Nurse I4

Staff perceptions were that so much time was spent on organisation it was unlikely to be outweighed by time saved.

I didn’t notice that it saved me any time because I was constantly chasing them [families] up to be there.

Nurse I1

The UCLH trainers were aware that organisation would be challenging and allocated time to discussing this in the training. The extent to which it was to emerge as a barrier had not been anticipated.

There have been no email, telephone requests back to [the trainers] . . . about the content of the manual, how to teach a particular element or uncertainty about ‘could you go through this again?’ I’ve had nothing. Nothing. The main one has been . . . ‘I can’t get people to come to this group, but also the admin organising side of it’.

UCLH trainer

Teamwork

Sites that were the most successful organising and running groups were similar. Two exemplary sites had dedicated MDTs who were flexible and willing to modify schedules in order to provide additional tailored care for the young people in their clinics. These sites had shown greater whole team dedication to the CASCADE project than some of the other sites, for example with greater consultant involvement.

Site educators

All four CASCADE modules should have been delivered by at least two trained site educators with the recommended level of diabetes knowledge. 165 Delivery in most sites was carried out by two trained educators (range 1–3). Staff delivering the modules were primarily PDSNs. In some sites the second (or third) site educator was a dietitian who shared the training equally at some sites, whereas in others covered only dietary topics. In one site an educator was a Child and Adolescent Mental Health Service (CAMHS) practitioner ( Table 27 ).The majority of site educators reported that they worked well together.

All sites had continuity with at least one trained site educator. Difficulties with second educators were experienced in a number of sites where they deviated from recommended CASCADE principles. In one site a dietitian who had not attended any of the CASCADE training assisted with the delivery of a module 4 session. In another site two members of staff who had only attended workshop 1 provided a support role in some modules. The rationale behind including untrained or partly trained staff was that site educators felt they did not have specific expertise that the untrained site educator had (e.g. dietetics or psychology). In a third site, two groups were delivered predominantly single-handed as the second site educator was sick. The site educator who ran these solo sessions said the lack of a second educator made delivering the modules very challenging.

Venues

Staff often had limited choice over venues. All groups, bar one, were held on hospital premises, usually in a seminar/training room. Most rooms had tables and chairs suitable for a teaching environment. The groups ran most smoothly when venues were comfortable, light and airy, and the set-up was conducive to interactive small group activities and discussion. One site held evening sessions in a pavilion in the community, creating a relaxed and informal environment away from the hospital setting.

Some sites managed with less suitable venues such as hospital wards and consulting rooms. One site held a module in an open waiting room area of children’s outpatients, with hospital staff passing through while the group was in progress. In another outpatient setting, there were no tables and the telephone in the background rang continuously. Distractions such as these disrupted module delivery.

Measuring fidelity of delivery

Activity completion was measured through site educator self-report and observer ratings. Self-report for each activity in all four modules had a median score of 7 (out of 7) for every activity (see Appendix 7 , Table 52 ). Observers scored activity completion lower than site educators, although still generally quite high. When observers rated an activity as being incomplete, the reasons given were mainly not using techniques enough or appropriately, rushing, missing a part of the activity or deviating from the protocol.

Time spent on activities

The manual recommends a guide time for each activity within each module. Modules were supposed to last between 90 and 110 minutes. Most lasted between 75 and 120 minutes (range 55–150 minutes).

Educators could be flexible with timing, depending on numbers in the group, group dynamics, family knowledge and interest in different topics. Although there was variation between sites, there were no activities that were either consistently truncated or that over ran excessively (see Appendix 7 , Table 53 ). A key factor affecting timing was the size of the group. Smaller groups took less time to complete an activity.

Very few observed modules incorporated a planned break. Observers noted modules flowed better when there was a scheduled short break to prevent children becoming restless and bored. Offering refreshments also appeared to help group mood and improved focus.

Content

CASCADE content was consistent with the recommended level of knowledge and skill of a PDSN. 165 However, there were areas of the curriculum that some site educators reported were beyond their expertise, for example use of CHO for exercise management. The educators spent less time than intended on these sections when delivering the modules. Educators were also more comfortable with activities where they felt families were engaged and/or knowledgeable.

The one thing that had a real impact on all of them was there was a chart that showed your HbA_1c and the rise. The graph . . . and how when your HbA_1c was 14, how much more at risk you were with your eyes and . . . they all sort of took a sharp intake of breath when they saw that and they’ve all said since, so it really makes you sort of think. We definitely would love to use that in the future.

Nurse I14

Site educators were less able to encourage discussion when there was initial hesitation from families or the family found the topic difficult to understand – e.g. matching food and insulin. When this happened some educators struggled to keep families engaged or interested.

The group introductions and reviews at the start of the modules were particularly challenging. The aim of introductions in module 1 was to help group members connect with each other, to increase engagement and encourage attendance. Despite the importance of this activity some site educators spent only a few minutes on it (see Appendix 7 , Table 53 ). This was particularly the case if they felt they knew the families well. Managing this activity was also difficult when groups were very small.

Module 1 . . . I thought the beginning of it where you’re trying to identify people’s strengths and weaknesses is a very difficult starting position.

I13

Educators also did not find it easy to facilitate reviews at the start of modules 2, 3 and 4. They struggled to ‘bring forth’ positive events and activities during the reviews specifically designed to build participants’ sense of knowledge, skills and abilities.

Reviewing the previous session, people couldn’t remember, it was too far away and I think some of that maybe because we’d have to re-schedule our sessions so it might have been longer than the month but still I do think a month is too long, you’ve forgotten . . . I couldn’t tell you what I did a month ago . . .

Nurse I13

Use of techniques

Self-report from the site educators suggested that they felt they had used the SFBT and MI techniques ‘quite a bit’ on average, but less than had been hoped. ‘Asking open questions’ and ‘focusing on positive solutions’ were rated by site educators and observers as being used the most consistently (see Appendix 7 , Table 54 ). They were less confident in ‘avoiding giving choices’ and asking ‘how come?’ and ‘what else?’, which were designed to evoke responses from families. Observers rated these techniques as only ‘somewhat used’.

Overall completion of activities and use of techniques

An alternative way to calculate fidelity was derived to explore the degree to which individual site educators delivered the four modules as intended. From each observer activity score overall activity completion and technique fidelity scores were calculated. The scores only reflect observed modules and do not provide data about changes in fidelity over time for sites that ran more than one group.

The composite activity score was based on all 24 activities delivered in the modules (as listed in Table 28 ). With possible ratings of ‘1’ (not completed) to ‘7’ (completed) the highest possible composite score was 168. The range of composite activity scores achieved by the 12 sites that completed at least one full group of the four CASCADE modules was from 116 to 161 (or 69–96% of the possible maximum score). This indicates that there was a high level of activity completion in eight of the twelve sites.

The techniques fidelity composite score was based on the six techniques used over four modules. With possible ratings of ‘1’ (not used at all) to ‘5’ (used extensively) the highest possible composite score was 120. Sites scored between 65 (54%) and 113 (94%). Half of the sites were able to use the techniques to an acceptable extent (> 70%) in delivery of the intervention ( Table 29 ).

There appeared to be relationship between technique fidelity and activity delivery fidelity scores – with sites having similar high, medium or low scores for both. However, there were a few anomalies – one of the sites that scored lowest in the activity completion rated high on use of techniques, whereas, in contrast, one of the most successful sites in terms of activity completion rated lower on the use of techniques. Both these sites took time to use the manual and plan the sessions.

Barriers and facilitators to delivery

Number of young people attending any CASCADE modules

Ninety-six of the 180 young people recruited to the intervention arm (53%) attended at least one module. Significantly more children (8–12 years) attended at least one module compared with teenagers (13–16 years) (64% vs. 44%, p < 0.01). Those who attended had significantly lower mean baseline HbA_1c levels than those who did not attend (9.52 mmol/l vs. 10.33 mmol/l, p < 0.01). There were no significant differences in attendance by gender, length of time since diagnosis of diabetes, ethnicity or level of deprivation in the area in which they lived.

Non-attendance

Eighty-four young people failed to attend any module. There were three main categories for non-attendance at any CASCADE modules.

1. Seven (4%) dropped out of the study before the CASCADE workshops were offered across six sites.

2. Eleven (6%) were never offered modules to attend by their site either because the site did not run any modules or had stopped offering modules.

3. Sixty-six (37%) were offered the opportunity to attend the modules but opted not to do so.

4. All sites, bar the one that failed to offer any modules to anyone, had young people who chose to opt out.

There was no significant difference by intervention site when all three reasons for complete non-attendance were included. However, when the seven who withdrew from the study were removed, attendance compared with non-attendance was significantly different across the intervention sites (p < 0.04). Table 30 shows that only two sites kept non-attendance at less than one-quarter of the recruited young people from their site.

The two sites with the lowest non-attendance rate had the lowest mean baseline HbA_1c levels for their recruited patients and completed the target number of groups for their site based on number of recruited families. The six sites with the highest number of non-attendance struggled with the organisation of groups, had high proportions of teens (vs. children) and relatively high baseline HbA_1c means. There was a significant difference in mean baseline HbA_1c level for the young people in the site with the highest level of non-attendance (75% non-attendance, mean 10.38 mmol/l) compared with the site with the lowest non-attendance (10% non-attendance, mean 8.75 mmol/l, p < 0.03).

Reasons for opting out of the CASCADE education sessions

Young people were asked at follow-up for reasons for non-attendance of CASCADE modules. For the 36 non-attenders that returned the questionnaire the most frequently cited reason was not wanting to miss school (n = 10). This was especially the case for teenagers, many of whom were preparing for GCSEs. Eight young people said they were not interested or did not think it would help. Other reasons are detailed in Appendix 7 , Table 55 .

I couldn’t be bothered. I just wanted to do the blood tests they asked me for.

Teenager I1

Parents provided very similar responses. Some explained work constraints hindered attendance, which made it more difficult for their child to as well. Only two parents suggested that they thought that CASCADE would not help their child.

[My daughter] does not understand diabetes. I hoped that she could attend to learn. I regret her not going.

Parent who had work conflicts I13

Site educators also commented on parents’ work being a barrier.

I think the child would have liked to have come but it’s more to do with one of parents, it was her dad. He was a very busy dad, making a lot of money I think, had a lot on his plate really.

Nurse I6

A few parents mentioned their child didn’t want to attend. One said her son was at a stage where he did not want to talk about his diabetes all the time. Some parents tried to persuade their children to go, others let this be their decision.

I tried to encourage my son to attend, so did the diabetes nurse. He just didn’t want to go. He said that if sessions had been during school time, he would definitely have attended so he could miss some school time!

Parent of teenager I14

When it came to the actual crunch, she goes ‘I don’t wanna do it’ and I was thinking ‘Okay, if you don’t wanna, don’t want to force you, don’t do it then’. So I wasn’t that bothered about it either.

Parent of child I12

Staff commonly cited lack of motivation when discussing the challenges of trying to encourage attendance by the most difficult to engage young people.

I would say the people that opted out I would have thought at the beginning wouldn’t have done it anyway so I was probably very surprised that they agreed to do the CASCADE study in the first place . . .

Nurse I6

I don’t think they were motivated enough, it’s just classic, isn’t it? That’s why they were invited in the first place, they were not motivated enough and they weren’t motivated enough to come.

Nurse I5

This view is consistent with the higher HbA_1c levels among non-attenders.

Some staff also commented they thought attendance at CASCADE would be better for those who were closer to their diagnosis of diabetes, as these families tended to prioritise health needs over any other needs (including school work), thus facilitating uptake.

I think when you’re newly diagnosed, they tend to attend better, if you say you need to come to clinic because you’ve just been diagnosed, they’ll find the time to come that very first Monday whereas when they’ve been diagnosed 2 or 3 years, they’ll phone up and say . . . ‘I can’t get out of work’ or something, ‘can’t get out of school so I can’t come to clinic’. They’re a bit sort of more ‘do as they’re told’ for the first 12 months, they’re more likely to attend and perhaps take it on board, it gets them in the right frame of mind early.

Nurse I14

Attendance at modules

Attending the first module increased the likelihood of attending three or four modules. Of the 96 attendees, 57% attended the full intervention of four CASCADE modules ( Table 31 ) and 78 (82%) attended three or four modules. There were no significant differences by gender, age group, ethnicity, deprivation score, mean baseline HbA_1c levels or length of time since diagnosis between those who attended one or two modules and those who attended three or four modules.

Regular attendance did not significantly vary by site. Seventy-five per cent or more participants attended three or four sessions in 10 of the 12 sites that offered all the modules. The remaining two achieved 67% retention. Delivery fidelity by the site educators did not influence attendance.

The main reasons given by young people for missing modules were not wanting to miss school, illness, having no one to accompany them, snow and forgetting. Only three young people said that they had stopped attending because they were not interested or did not want to go. Parents’ responses were similar citing specific reasons why individual modules had to be missed, and disappointment that they were. A few parents suggested that attendance was a struggle. One parent commented:

We attended all the sessions but (my daughter) didn’t want to go but I insisted she went.

Parent of teenager I6

Staff agreed most participants seemed engaged and lack of attendance did not indicate a dislike of the programme.

I think it was personal issues really, one teenager had an exam on the last session, it was circumstantial rather than not wanting to come I think.

Nurse I6

Number of parents/others attending

The CASCADE intervention was designed for young people and family members to attend together. Young people were accompanied at 297 of the 327 modules (90%). The majority were accompanied by at least one parent (mostly mothers); however, other significant family members and one girlfriend also attended (see Appendix 7 , Table 56 ). In 50 modules (17%) more than one member of the young person’s family attended the session with them. One young man attended three modules on his own. Three modules were attended by family members but not by the young person.

Acceptability of CASCADE to young people and parents who attended

At 12 months, young people and parents who had attended modules were asked if they would recommend the CASCADE intervention to a friend in their situation. Overwhelmingly parents and young people said they would. Only one parent (out of 85) and two young people (out of 94) said that they would not make this recommendation.

I felt they were of little use to me as I already knew everything however this kind of session would be useful to someone who had just been diagnosed.

Teenager I9

I think they were very helpful and informative. The atmosphere at the sessions was very nice and friendly and I would definitely recommend it to friends.

Teenager I10

In general, responses were positive to specific questions about module content, group dynamics and the structure and length of sessions ( Table 32 ).

Group dynamics

Most young people and parents liked being together in the same groups. In the questionnaires two-thirds said that they’d learned something from other people in their group with positive endorsement for group work from the interviews. Very small group sizes were criticised by parents and young people with mixed age groups described as problematic by some young people.

I quite liked the fact that there were other kids with the parents there, I think it takes the pressure off the individual child so for them to be with other kids that have the same problem and bringing it up there . . . their issues.

Parent of teenager I8

If more children and parents had turned up, then it might have been [easier to come]. It would be different; you needed more [participation]. Just one or two doesn’t make it beneficial, you need feedback from more, at least a good half dozen would’ve been good, I think.

Parent of child I5

[Being in a mixed age group] kind of restricted the things to talk about but if you make CASCADE each have a different age group, it would be better, bit more open.

Sort of sexual matters and drinking matters and sort of drugs and stuff like.

Teenager I11

Good [things about CASCADE] – meet other [people with] diabetes your age, with same probs and how they deal with it; Bad [things about CASCADE]- time consuming (coursework, homework . . .) and aimed at a younger age group.

Teenager I8

Teaching styles and length

Nearly all parents were positive about the length of sessions and methods used by trainers. In contrast nearly one-quarter of the young people were critical about the length of the sessions and said they would have preferred more sessions that were shorter.

I would have said . . . cause sometimes with like a concentration lapse, maybe [the sessions could be] a bit shorter . . . Yeah, cause like say if [the module] started at 3, they did then carry on till 5, all of us are just thinking about our stomachs.

Child I1

I thought it was OK, but I was a little bored, but my mum thought it was good.

Child I3

It was well planned . . . yeah, it was like humour as well . . . They gave us activities to do and then like we’d all go through it which helped like to get a good understanding . . . They stood out the front and did their bit and then they like handed it down to us, to let us like discuss our views on what like we thought about it.

Yeah . . . I think so, like we would all have different methods of like different things, so . . . yeah, it made the sessions better that we could all like discuss our views in the middle of the sessions.

Child I1

Most young people thought it was a good thing to know the site educator in advance of the modules. A few were critical of this, however, because they felt the nurse had preconceived ideas about them.

I think it makes it more personal [if you already know the site educator] . . . I think I would have felt less comfortable [if I hadn’t know them] cause I was meeting new people in the group and I would have had to have met new nurses as well, it could have been quite awkward cause we all knew [the site educators], it was all quite mutual.

Teenager I9

Content

About half the parents and young people thought the CASCADE content covered things they already knew, but many commented positively about having an opportunity to revisit this information.

Quite a lot was a reminder of what we might have known but it was still very helpful. I enjoyed it a lot

Teenager I6

I thought it was really useful – right balance between sciency and everyday terms and it was applicable.

Teenager I11

Well I didn’t enjoy it that much because they’re basically saying things like I already kind of knew and when I got diagnosed I got taught about all the things that they were doing.

Child I5

Perceived impact of CASCADE

Nearly all parents who completed follow-up questionnaires (95%) and/or took part in interviews perceived CASCADE to have had some positive impact. When parents reported no real changes, this was sometimes because they said they had always felt confident about managing the diabetes, and/or encouraging their son/daughter to participate in sports and other activities. During interviews, a few described some change since attending the groups but were unsure whether or not this should be attributed to CASCADE rather than other factors, such as increasing maturity or other life experiences of their child. One parent reported that although useful at the time, attendance at the sessions had no continuing impact. Another parent expressed disappointment that it hadn’t helped as she had hoped.

Like the parents, almost all young people identified some impact of the CASCADE sessions. Many indicated that changes had been both positive and sustained but often perceived as modest.

A little bit.

Cause like if it’s too high I give myself a few units and then I go and do like take the dog for a walk or something like that.

No.

I have, just like little bits? . . . Um, matching my food to insulin.

Teenager I1

Specific themes of how CASCADE was perceived to have had an impact are detailed below.

Improved knowledge and understanding

Ninety per cent of parents who attended the CASCADE sessions reported a greater understanding of the need for CHO counting with others reporting increased knowledge relating to insulin and CHO ( Table 33 ). In the interviews many parents reported increased knowledge as a result of CASCADE, either for specific topics or it had reinforced or supported their existing knowledge base.

At 12-month follow-up approximately three-quarters of young people identified increased knowledge and improved understanding in diabetes management as a positive outcome of attending CASCADE (see Table 33 ). CHO counting, matching insulin to food and exercise were most commonly cited as areas of increased knowledge and many young people gave examples of applying this knowledge to daily management.

I still eat but now I’m starting to inject the right amount.

[before CASCADE] I’d sometimes miss the injections.

Yeah, cause I understand a bit more.

Teenager I14

Many examples were given of how the young person had applied this knowledge and understanding to the management of their illness with subsequent outcomes.

Well I’ve learnt new things . . . yeah, I could do some things but it’s just like improved the way that I do it . . . I’d probably say like carb counting, and yeah, I used to have like a rough idea, now I know like things that are already in my head, like how much a piece of bread . . .

Child I1

Another young person was now able to make their own adjustments to match insulin to foods and explained the nurses had ‘shown her different ways’.

At 24-month follow-up 84% of parents who attended the modules continued to report that CASCADE had increased their knowledge ( Table 34 ).

I found the sessions very informative, especially where growth hormones and ketones were concerned. I thought I was quite clued up on diabetes until I attended these sessions.

Parent I9, 24-month questionnaire

Greater confidence in managing diabetes and making decisions

At 12 months, over two-thirds of parents who attended any CASCADE module reported greater confidence in making decisions in the day-to-day management of their child’s condition; 78% felt that they were better able to control their child’s BG levels (see Table 33 ). Parents who attended the groups commonly reported anxiety and worry had decreased. Greater confidence was frequently attributed to an improved understanding of diabetes and its management, and a belief that their child’s diabetes was being managed more effectively. In a few instances parents identified a particular component of CASCADE which they had found helpful, for example a log book as a tool for assessing foods. Young people also reported greater confidence making decisions about insulin needs and application of this knowledge in day-to-day management of diabetes.

At 24 months, 61% of parents who had attended still reported more confidence managing their child’s diabetes. Over half said they were happier about how their child felt about their diabetes and just under one-third said they worried less now about their child’s diabetes as a result of CASCADE (see Table 34 ). Ambivalence in relation to putting knowledge into action was highlighted by nearly all the parents (93%) reported a greater desire to stop their child’s BG going too high or low; however, less than half (43%) said they wanted to test their child’s BG level more often.

Reported impacts on disease management and lifestyle

Improved management of diabetes was described by a number of interviewed parents. In one instance, a parent identified CASCADE as the reason for their daughter not having had an admission to hospital since the sessions. Improved BG levels were also reported in a number of cases.

Additionally some of the interviewed parents suggested practical and lifestyle impacts of attendance at CASCADE modules. Specific points made by parents included changes in shopping habits, closer reading of food labels, understanding of CHO counting, and managing sugars with exercise. In one instance, the parent reported that as a consequence of the participation in CASCADE, she was now able to leave the home for a day or two without worrying.

Changes observed in the young person by parents

Young people were described in interviews by parents as having greater awareness and understanding of their condition, and being more confident and comfortable with managing their condition. Examples given to illustrate this were changes to regimens and taking more responsibility.

Especially in school time, he is more responsible.

Parent of teenager I12

One parent described how her daughter, previously, apparently disinterested, was now motivated to manage her condition effectively and took responsibility for many tasks with which she had been previously less co-operative.

Social impacts were also reported. One parent observed that her daughter now felt able to discuss diabetes with friends and participate in social activities she had previously avoided. Another parent reported how her son had become more accepting of his condition and was more open with his friends.

Young people’s observations of self-change

Greater understanding of long-term complications of disease motivated one young person to change eating habits. This young person also reported

Keeping a lot more close eye on my BG . . . I think it made me realise . . . keep it under control now so I’ll be healthy when I’m older.

Teenager I3

Closer attention to, and/or more frequent testing of, BG levels was reported by many young people as well as the belief that they could make more appropriate adjustments particularly in relation to food and sport.

I suppose I’m taking it more seriously . . . I’m looking at more sort of how I should be eating and how like exercise and other things, how it all affects . . . you know . . . I change my units up and down based on what I’m eating and what I’m doing.

Teenager I11

Young people becoming more proactive in the management of their condition was common, for example one young person reported doing her own research:

[Regarding different foods] I noticed that they’re not good for me and I also got a list of things that were good and bad for me off the Internet.

Child I12

Changes of attitude and emotional factors

Sessions came along at the right time for this young person:

I had to finally accept and think that it was enough of actually going into hospital and getting into a lot of trouble.

Teenager I14

At 12 months young people felt more in control (68%). At 24 months they described a change in attitude, including ‘greater acceptance’ of the condition and the need for injections. Forty per cent reported that they were happier with their diabetes as a result of attending CASCADE and more confident (43%) and comfortable in the management of their condition (see Table 34 ).

[After CASCADE] I look after it much more than I used to. I feel good, just to know that I know.

Teenager, I5

I’ve been more happier . . . yeah, like around the house I’ve been more happier. Not so many strops . . . Cause my readings are better and we’ve been given a lot more information about the ketones and how to treat it . . . I found it really good . . . like useful the way they presented it.

Child, I13

In some cases young people also reported that they could take more responsibility:

I can do more things now, like by myself.

Child, I11

. . . if I’m in a low I don’t have to . . . nobody else has to worry about it that much cause I can handle it.

Teenager, I5

Partnership between parents and young people

At 12 months, three-quarters of parents said they had a better understanding of how diabetes affected their family following CASCADE. Participation had a positive influence on parents’ relationship with their son/daughter and their interaction managing diabetes.

Attending modules with their son/daughter was identified as beneficial by a number of parents. They described a shared understanding of the condition, which made communication at home easier. Parents also described how after sessions both parents and young people had a greater appreciation of each other’s perspectives and concerns. One parent also reported how they enjoyed spending this dedicated time, including the travel time and attending the sessions, with their son/daughter.

She’s not thinking ‘shut up’, she knows why I’m saying it, she understands . . . why I was saying it.

Parent of child I6

I think we’ve both been told the same . . . we’re both more on an even keel.

Parent of teenager I12

A number of parents reported they learnt more about their son/daughter by attending CASCADE with them. For example, they discovered that their son/daughter had greater knowledge and understanding of their diabetes than they previously thought, or gained more insight into how their son/daughter felt about having diabetes.

At 24 months, 71% of parents still reported it was helpful to have been at the sessions with their child and just over one-third said they get on better with their child regarding their diabetes now, as a result of CASCADE (see Table 34 ).

Young people also reported better co-operation with parents in managing diabetes, as a consequence of having a better understanding of the need for good management.

I just like, my relationship . . .

Like I can become less angry and stuff and like less agitated and stuff.

Teenager I14

. . . my mum used to say inject and I always did it like an hour later cause I was feeling it won’t affect me, but they showed me the graph and whenever mum told me to inject, I did it like seconds after.

Child I14

. . . she [mother] probably worries less because she knows I know a lot more about it.

Teenager, I12

One young person said his father now “paid more attention” and would engage in discussions to help achieve appropriate sugar levels.

Communication with diabetes team

Pre-existing relationships and accessibility were generally felt to be very good, therefore, the majority of parents and young people reported no change in the way they might contact or relate to clinic staff. However, at 12 months parents (82%) reported they were more able to contact health services in regards to their child’s diabetes after CASCADE. At 24 months, over half (60%) said that they liked spending time talking to the diabetes team without being in a rush. Reported changes were always in a positive direction, for example feeling more able to ask for help, or being more prepared to ask for explanations or to ‘challenge’ answers to questions. Positive changes for young people were that they found staff more approachable, were more confident raising issues for discussion and prepared to disclose their feelings:

I used to never tell the truth about my bloods because I was scared they would shout at me . . . But now . . . if I’m like high I’ll tell them . . . and they sort it out.

Child, I14

I wouldn’t say changed it [feelings about talking to nurses and other staff], but like I say, I have become more confident with knowing what I’m actually talking about.

Teenager, I11

Long-term impact on blood glucose control

Over half of parents reported that CASCADE had a long-term impact on motivating their child to keep trying to keep their BG well controlled. Just under half of young people reported some long-term impact on how hard they tried ( Table 36 ).

Why CASCADE might have had less impact

A couple of parents said the information covered was known already. A few felt that group dynamics (differing characteristics of participants – e.g. ages, regimen, small group size) reduced the potential impact. Several parents acknowledged the difficulties of trying to change behaviour in young people.

I think being a ‘teenager’ has interfered with her understanding. She knows what she should do but doesn’t always do it. Hopefully this is something that she will grow out of and get back on track.

Site I2, 24-month questionnaire

Useful to have time to talk through issues surrounding management, but my child is at an age now where she will not always listen to advice.

Site I1, 24-month questionnaire

Two parents also suggested that although they liked CASCADE, the intervention did not seem enough on its own to overcome entrenched problems.

I will always worry about my child’s diabetes. My child is very angry about having diabetes and I don’t think the amount of lessons about diabetes will stop my child from asking ‘why me?’! . . . My child tries hard but feels disappointed when goes to clinic and the results aren’t what she expects. Then leaves feeling low and not wanting to do anything including blood readings.

Site I6, 24-month questionnaire

The most frequently mentioned reason for CASCADE failing to have an impact for young people was that they already knew the information covered. Some said they were bored or the sessions were too long or the content was boring. Young people also said that mixed groups with different ages and genders and varied regimens made the groups less helpful for them.

Data sources

• Staff interviews after the CASCADE groups were completed.

• Parent/carer and young people’s interviews.

• Detailed fieldwork notes.

• Data collected for the minimisation criteria.

• CNR.

Patient numbers

The number of registered patients per clinic ranged from 78 to 220 (control) and 90 to 280 (intervention) at the start of the study, increasing slightly across both arms during the study.

Consultants

The majority of sites had one paediatric consultant. Six had two consultants and one had three. In 27 sites the ratio of consultants (including staff grade) to patients ranged between 1 : 80 and 1 : 120. In one site (intervention) it was 1 : 170. In one intervention site the consultant was on long-term sick leave for the majority of the period from patient recruitment to completion of delivery of the intervention.

Nursing

Approximately half of the 28 sites had one PDSN, one (control) had four, with the remainder having two or three ( Table 37 ). Nurse to patient ratio ranged from 1 : 51 to 1 : 170, with seven having a ratio of ≥ 1 : 100. The RCN recommends a ratio of 1 : 70. 165 The majority worked part time. At the start of the trial there were 29 nurses in the control clinics [22.4 whole-time equivalents (WTE)] and 24 in the intervention clinics (17.9 WTE). In control sites there was little change in nursing time over the course of the study. Two intervention sites had a 100% increase in nursing WTE. All control clinic nurses and over 90% of the intervention clinic nurses described themselves as being employed as PDSNs, with the majority having at least one nurse who had worked at the site, in their current role, for over 5 years.

Dietetic support

In control sites, 13 out of 14 received approximately 0.1 WTE dietetic support compared with 0.2 WTE in intervention sites. The dietitian was not always a paediatric dietitian. The majority of dietetic time was used to cover routine paediatric diabetes outpatient clinics and newly diagnosed patients.

Psychological support

Very few sites had any integrated psychology provision in the early stages of the study. Lack of psychological support was a major concern for many staff. In general, referrals go to local CAMHSs. This was almost unanimously considered to be very unsatisfactory with long waiting lists and the threshold for accepting young people often too high for most referrals from the diabetic team.

Unless it’s a suicide attempt it’s difficult to get anyone to see children.

Site educator I12

Two intervention sites appointed a psychologist during CASCADE, at which point patients began to see a dietitian and psychologist at every clinic. Two other intervention sites referred patients to the UCLH diabetes team to access psychology services.

Team working

In most sites all members of the MDT were in an acute setting; however, there were a number of sites where the PDSN was based in the children’s community nursing team – working alongside generic community nurse colleagues.

We’ve no longer got a Band 7 [Nurse] so we’ve been incorporated into the Children’s Community Nursing team. It does mean that we don’t have anybody leading diabetes and keeping us as one team which we’re meant to be. A lot of things that are challenges to us at the moment are due to the fact that we don’t have a Band 7 because the Children’s Community Nursing team leader who is currently our team leader has no concept how diabetes as a specialism works really and doesn’t really understand how we work.

Nurse/site educator I3

Paediatric diabetes clinics

Most sites ran a consultant-led clinic at least once a week. At least two-thirds of control sites ran additional regular nurse-led clinics. Fewer intervention sites ran nurse-led clinics. Nurse-led clinics aimed to be more flexible and informal than consultant-led clinics and were primarily a key source of support for patients with persistently poor control and their parents/carers. In most sites nurses also attended consultant-led clinics. This meant attending 8–10 clinics per month in addition to home visits, school visits and telephone contact. Sites without nurse-led clinics recognised the need for them and had plans to start running them. Some nurse-led clinics were held in community venues.

Frequency of nursing contacts from 2008 to 2012 was obtained from a detailed CNR. The mean length of time data were collected was 49.7 months (range 44–54 months). There was huge variability in how contact information was recorded. Nurse contacts included home visits, nurse-led clinics, telephone calls, school visits and ward visits excluding routine clinic visits with the consultant. Data were obtained for 164 control and 162 intervention patients. The average number of nurse contacts per patient per year in control sites was 5.08 (range 1–121) compared with 7.12 (range 1–168) in intervention sites.

Clinic ‘did not attend’ rates

Clinic staff were asked to estimate the average number of young people who ‘did not attend’ (DNA) clinic appointments. The rate reported was 5–42% for controls (with most around 15%) and 5–20% for intervention clinics (with most around 10%).

All clinics reported higher DNA rates for adolescents and young adults (20–50%). One nurse had offered nurse-led clinics to encourage clinic attendance, which had no impact on attendance. One nurse said it was ‘a regular few’ who usually did not turn up. In one clinic, a nurse felt the low rate of 5% was because clinics were scheduled in the afternoon so young people did not miss too much school. Some staff mentioned working towards a policy on how to deal with DNAs as part of the preparation for the Best Practice Tariff (www.diabetes.nhs.uk/networks/paediatric_network/best_practice_tariff_for_paediatric_diabetes/).

Annual review clinics

All clinics carried out annual reviews – the number of clinics reporting what was offered (at the end of the intervention period) is listed in Table 38 .

Education at diagnosis

All sites provided initial diabetes education at diagnosis on the ward. PDSNs were primarily responsible for education with contributions from other members of the team, including consultants and dietitians and ward staff.

Well, I do [the initial education] myself . . . obviously the parents are seen by the consultant within the 24 hours and he tells them a little bit, and then I have a package that I work through . . . that’s all we do really before they get home but they do see a dietitian as well for a brief description.

Nurse C6

Consultants had a minimal role in education beyond a brief description of the diagnosis. One intervention site offered an appointment with an Emotional Wellbeing Practitioner based in the primary care service, whereas another referred to a psychologist at diagnosis to assist families with adjustment and problem-solving.

Initial education at diagnosis usually followed a checklist. It was generally regarded as important to deliver information in manageable ‘chunks’ so as not to overwhelm families who were dealing with the initial shock and implications of the new diagnosis. The checklist covered the ‘basics’ about diabetes including a general overview, managing BG levels, living with diabetes at home and MDI. One clinic uses a DVD from DUK for families, as well as books and leaflets.

It’s quite intensive really . . . how to do an injection, how to prepare it . . . We like them to have a decent understanding of hypoglycaemia because that’s the biggest emergency . . . high blood sugar levels and DKA . . . We call those the initial survival skills.

Nurse C7

Dietitians covered topics related to food and CHO. Only one control site introduced CHO counting from the beginning.

We try and do that right from the start so they’ve got a good idea of adjusting their own doses because it’s much harder to change once they’re doing something different.

Nurse C1

Nurses kept in close contact with families after their return home, giving them further information, supporting them through home visits and telephone contact before their first clinic visit. Families were encouraged to telephone nurses with queries or concerns. Some patients had their first hospital appointment 1 week after diagnosis, but for others first appointments could be up to 6 weeks later.

Contact between nurses and families with a newly diagnosed child varied depending on the perceived individual needs, as well as available resources. Many nurses offered several home visits to deliver more information about managing diabetes. Other nurses offered telephone support, often daily, in the first week. Following that, telephone contact was on an as-needed basis for ‘trouble-shooting’.

It’s very much on an individual basis because I say to the families, ‘I’m here to support you’, and some families will need daily visits for maybe the first 4 or 5 days . . . I’m very much guided by what the families want cause some need more support than others.

Site educator/nurse I4

Usually go and do education on the ward, follow them up with a home visit and then do an evening session for their family and friends and then do as much education (as required) for their swimming teachers, scouts, brownies, community education . . . I usually follow them up once a week initially and then once a month until they feel confident but I phone them every day and we talk about what the day is like and then put into place anything we need to for the following 24 hours.

Nurse C2

Nurses also visited schools to deliver training to staff often within a week after diagnosis and before the child returned to school. This could involve several visits to implement a health-care plan and to train staff, including first aiders and school nurses. One nurse mentioned training ‘whole school employees’ including receptionists. This could be more intensive for primary school children who were more dependent on staff.

I go into the schools when they’re newly diagnosed if they want me to educate the staff and (help with) health care plans. If they change to different teachers, then I’ll go in and see the new teacher to educate them. I’ll go into whole schools like the training days if they let me and just educate all the staff for that and like if they change to Basal Bolus or a pump, then again I’d go in and help support the staff.

Nurse C3

A minority of sites were less involved with schools. In one intervention site, the nurse acted in an advisory capacity but considered the care plan to be a contract between school nurses and parents. A nurse at one control site said they were having difficulty implementing comprehensive care plans because schools were reluctant to take responsibility for administering injections and blood tests.

Most staff rated education provided for newly diagnosed families as ‘sufficient’, ‘good’ or ‘very good’. The majority of staff had developed education programmes over time using teaching methods that they felt worked best for them and their patients. This was rarely formally evaluated. In contrast one site was infuriated by the impact of lack of dietetic time.

What are the key challenges for your service at the moment, would you say?

What do you have in the way of that?

Is that a change?

How do you get round it?

So is that a better service?

Ongoing diabetes education

The majority of ongoing education was delivered by PDSNs with some input from dietitians and consultants. Dietitians were particularly involved in teaching CHO counting. When nurses felt that children and families had grasped ‘the basics’ further information would be discussed depending on individual need. Education was delivered on an informal basis in consultant-led and nurse-led clinics, during home and school visits. Clinic appointments were time limited and were regarded by some as not the best venue for delivering education, which tended to be unstructured and ad hoc.

I would say if you’re attending a clinic appointment we do try and fit in education around that but it’s very, very limited.

Nurse C4

I think we cover everything, it’s pretty comprehensive, there’s no time limit to do it, some do it all in 6 months, some do it in 6 weeks. We do an awful lot of . . . we do training for members of the family as well.

Nurse C6

Most clinic staff went into schools on an annual basis, or on request, to update schools. Although parents were the first port of call for schools in emergencies, PDSNs had ongoing contact with schools by telephone, fax and e-mail and through school visits. A nurse at a control site was frequently called to visit schools that did not have school nurses. In one intervention site staff went into schools four times at the beginning of a school year to train staff in giving injections, to explain regimens, to offer advice on school trips and managing hypoglycaemic attacks and to make sure they had adequate supplies. In another intervention site the nurse had also taught other school children about diabetes.

I’ll go in and do education sessions to the other children so that they know what little J’s having to go through and if there’s issues of teasing or bullying I’ve been into schools as well to talk about insulin and . . . it’s not fun.

Nurse/site educator I13

Two sites offered training for health-care and education professionals. One site had organised a group session for up to 40 teachers; however, schools were reluctant to pay the £5 the clinic was charging to cover costs.

Group education for patients

Many sites had run groups at some point. These tended to be occasional and ad hoc rather than a formal or regular programme. A few sites had run more formal programmes with a planned curriculum that generally focused on CHO counting. One intervention site nurse and dietitian ran a series of 3-hour sessions for a group of 4–6 people on CHO counting over a period of 3 years prior to CASCADE.

Limited resources were seen as a barrier to the development of group work. Although some site staff reported positive experiences with good attendance and good feedback from attendees, others had found them difficult to organise with poor attendance.

[Organising groups has been] a bit ad hoc. It’s always taken ages; different languages (spoken by families) and times that people can make it are so different.

Nurse C9

One nurse said she had run groups in the past but felt they did not work as well as one-to-one training where individual needs, abilities and stages could be catered for.

About half of the control sites ran some kind of education groups during the trial, which were described as informal, loosely structured and ‘ad hoc’. A few had attempted to use a theoretical base and a curriculum.

Examples of groups offered in control sites during the trial are listed below.

1. One PDSN organised educational and social events, affiliated to DUK, for children and families to meet and gain support in living with diabetes. Age-banded events were held including a meal at a restaurant for the teenagers, a party for the younger ones and an activity day for 7- to 17-year-olds. A parents workshop was facilitated by an experienced counsellor. A Healthy Eating family workshop was facilitated by a dietitian. The PDSN also organised a ‘boys-only night’ for boys aged 11 years and their Dad, brother or best friend (male). A few adults with diabetes share their experiences of life with diabetes – from leaving home to getting a job and what happens when things go wrong. Boys with diabetes who attended got a £10 voucher.

2. A group has been run by a dietitian for the past five years on CHO and insulin adjustment based on DAFNE. 145

3. ‘If it’s about CHO counting, you know, the grandparents are struggling, then we get the scales out and the food and start weighing so they understand and it’s much easier with practical sessions’ (Nurse C2).

4. A series of 10 education sessions over 12 months (in 2010) was offered through a quarterly newsletter. Topics included ‘What is diabetes and what does it mean for me?’; CHO counting; exercise and travel; sick day management; pumps; insulin dependent devices; sex, drugs and alcohol; and transition to adult services. Attendance at sessions was poor.

Almost all group sessions were run by nurses and dietitians with minimal involvement of consultants at three control sites. Two control sites mentioned the involvement of employees from companies that provide pumps in CHO counting training.

We also have a rep that comes from his company . . . helps with the carb counting so that’s why he’s been involved and he’s offered financial help as well.

Nurse C13

Education groups were primarily run on hospital premises. The majority took place outside school hours in the late afternoon, evening or during the school holidays, and occasionally during weekends.

Certainly every main holiday and often every half term as well, it depends how many youngsters have expressed an interest and obviously the availability of the dietitian as well but certainly every Easter, summer, Christmas holiday we would do one session and often we would do one in between at half term.

Nurse C8

In the majority of cases, attendance was a patient’s choice. Sessions were open to those who ‘expressed an interest’ or who wanted to come along to a particular session. Families were invited to group education on the basis of their ‘stage’ in the clinical pathway.

We have an ongoing list so that’s decided as we get to know the patients and their families and we know what the next stage is for them or if they’re going to be going on to an insulin pump that’s part of pump (pathway).

Nurse C10

Groups varied by site according to type of education offered and whether it was targeted to a specific group or the entire caseload. PDSNs were aware that mixed age groups did not work well so organised groups on the basis of age or gender with age regarded as more important than gender. Parents were involved in some groups.

Rather than doing a group of 12, it’s probably better to do a group of 6 so we’re looking at splitting that one but when we actually came to it we discovered the older teenagers weren’t particularly impressed, so we’ve actually had to split the very young from moderately young.

Nurse C8

A lot of stuff around sport and exercise . . . I would say one of the best attended ones because we’ve had an outside speaker and she’s excellent and we’ve had two or three sessions on that. I would say we’ve had 60 people there at those sessions. Yeah, it makes it worthwhile for her coming as well

Nurse C4

Group attendance varied from two to three individuals up to groups of 60. The percentage of invited people who attended was generally quite low and was often difficult to predict.

Design strengths

A randomised cluster design was used to good effect in this study. An appropriate number of clinics and additional participants were recruited to allow for dropout without compromising statistical power. We achieved a sample size that allowed adequate power to test our hypotheses. All centres recruited and randomised into the study completed the structural and training elements of the study.

Delivery by a large number of practitioners minimised the influence of individual practitioners’ skills and personality. Secondary outcomes used included several shown to perform well in measuring QoL in children with diabetes in a range of contexts. 203 Assessment of practitioner fidelity used extensive observation of a large sample of workshops delivered at different time points, as well as self-assessment by practitioners after delivery of every group. 204

Data loss

Many young people had issues with providing the blood samples, which was a major barrier to both recruitment and subsequent collection of HbA_1c data after consent. Considerable negotiation was often required as to how and when the sample would be collected. Clinics often struggled to find an appropriate HCP to collect the blood samples. Primary (HbA_1c) and secondary (questionnaire) measures were sometimes collected on separate occasions at both follow-up points. Trying to organise telephone questionnaires, when they could not be completed in the clinic, was a key contributor to data loss.

Internal validity

Measurement bias was minimised by the use of a central laboratory for all HbA_1c assays. The two trial arms were well balanced at baseline. One limitation of pragmatic cluster trials is possible bias due to trial arm allocation knowledge. Allocation was concealed until after clinics had consented and a first participant recruited to minimise selection biases at entry of clusters to the trial. Checks for bias revealed no obvious imbalances in demographic data. It is unlikely that allocation knowledge bias affected recruitment rates as there was no difference between trial arms in number of eligible patients or final recruitment rates.

Intervention development

A strength of the CASCADE trial was incorporation of recommendations from previous systematic reviews of psychoeducational interventions in childhood diabetes. 91 The intervention had a firm theoretical basis that drew significantly on principles of MI and SFBT previously shown to have potential in childhood diabetes117,121 and incorporated key components of structured education programmes following NICE guidelines. 4

Educational content

The study highlighted different levels of knowledge in practitioners for aspects of key diabetes content. Problematic topics were related to food groups, matching food and insulin, CHO counting and insulin sensitivity ratios. A small number of secondary site educators did not have a thorough grounding in diabetes knowledge. In interviews only half of the parents and young people said they had been familiar with the content of the groups even though the majority of the content should have been covered in the average 5 years since diagnosis. This confirms a lack of consistent and effective structured education currently being offered as part of standard care by the majority of diabetes teams.

Training to deliver CASCADE

Overall the training programme was positively evaluated although motivation to attend the workshops varied, often depending on how site educators had become involved. Training (from the same two trainers) took place over 2 days, primarily at UCLH. The majority of site educators regarded 2 days of training as feasible and acceptable, although some felt there was too much to take in. All site educators received substantially the same or very similar training, with some variation depending on group dynamics and types of questions and discussion. Site educators generally thought that the training was very good and were initially excited and motivated.

Running groups

Despite the post-training enthusiasm of the site educators, only 30 out of a possible 44 groups were run (68%). One site did not manage to run a full set of sessions and one site failed to run any groups owing to lack of staff.

Experience of setting up and running groups by site educators was often minimal. Most sites found organising the groups very difficult and time-consuming, with little use of electronic media to facilitate this process. Some found it very difficult and time-consuming to agree dates and times with families. Practical issues, such as finding appropriate rooms to house the groups and identifying time for nursing staff to deliver the group, were also barriers to delivery. Some thought that more time could have been spent on this in the training, despite 25% of the training being focused on getting clinics to think about how to get the groups up and running as well as practical issues like arranging rooms. Site educators suggested that more administrative support would have been helpful. Key barriers to running groups included lack of administrative support for organisation of groups, low use electronic media for communication with patients, competing priorities for potential participants and factors specific to being part of a research study.

Age-appropriate groups

The educators found it difficult to keep the young people focused when more challenging topics were introduced. This may have been because the language used by the educator was not age appropriate and was particularly challenging in groups with mixed ages. The UCLH training team emphasised the need to have similar age groups. However, for pragmatic reasons compromises were sometimes made in terms of the CASCADE model for optimum group composition. Rather than grouping by age, groups were often mixed. Educators found it difficult to keep mixed age groups focused on activities and to engage the entire group at once. Additionally, the age mix of the groups was felt to be problematic by some young people.

Delivery

The intervention appears to be deliverable within the context of routine care without major impacts on service structure with some provisos. Although at least two practitioners from each study site attended both workshops, there was a failure to use trained educators in a few sites in the delivery of some groups, which compromised the integrity of the programme. Two sites that swapped site educators across the modules started with higher fidelity scores in the first two modules, which tailed off in modules 3 and 4.

Fidelity of delivery of activities was generally high across all sites. Site educators delivered activities as described in the manual and used aspects of the approaches demonstrated in the training sessions. However, less time than was recommended was often spent on some activities. Some of the key exercises that would have had the greatest impact on HbA_1c level were the ones on which less time was spent. Key factors that influenced effective delivery included having an appropriate venue and having planned breaks. In addition, advance preparation for sessions by both site educators and skilled group management using the psychological approaches with confidence made the groups run smoothly.

Although fidelity of CASCADE techniques was good, it was not as high as desired. Site educators were aware that they needed to use CASCADE approaches more often (e.g. evoking knowledge by asking ‘What else?’) but found it difficult to do in larger or small groups or where there were group dynamics they found difficult to manage. Educators found it challenging at times to work with parents and children and keep both of them focused on the task and would have benefited from more training in how to manage the group when there was a dominant member of the group, usually a parent. Some site educators found it difficult to move away from didactic teaching and struggled with ‘avoiding giving a choice or giving people the answer’ and asking ‘how come?’. They said it was ‘impossible’ to keep asking ‘what ifs’ if only one family attended a session. The frequency with which groups would comprise single families had not been anticipated in the training, hence this was not prepared for. The training workshops seemed to have ‘deskilled’ the PDSNs, as in normal practice they are more used to working with single families than groups. As PDSNs reported lacking confidence delivering the CASCADE groups to a single family this would have to be addressed in future training.

Uptake

Staff thought that many of the young people who were eligible for CASCADE were the least motivated in terms of their diabetes, and that these remained the least motivated to attend groups, even if recruited. Of the 180 families recruited, 66 (37%) failed to attend any of the groups offered, but were happy to provide baseline and follow-up data. The majority of reasons for opting out of groups were a desire not to miss school (young people) and not missing work (parents). Older teenagers and young people with higher baseline HbA_1c levels were significantly less likely to attend than those with lower baseline HbA_1c levels. The mean HbA_1c value of the patients at baseline was higher than the population average as reported by the NPDA. This was because the study selected young people with a higher HbA_1c level. The data suggest that the intervention does not appeal to young people with baseline HbA_1c levels of ≥ 10.4 mmol/l and an alternative, more personalised and/or intensive, intervention may be required for this subgroup of patients.

Staffing problems and difficulties organising groups prevented 14 of the expected 44 groups being offered. The most suitable timing for when groups are best offered and which specific age groups should be targeted remain unclear. Some staff felt that CASCADE would have been more effective if offered during first 12 months after diagnosis when young people and families are more likely to take on information and are more in ‘the right frame of mind’ (site educator, I14). Children (8–12 years) were significantly more likely to attend CASCADE than teenagers (13–16 years), which would suggest that parents found it easier to ‘bring’ younger children to the session. In addition, reported perceived impact was significantly higher for younger children than teenagers.

We contend that the generalisability of these results is high for diabetes patients in the UK aged 8–16 years and their carers with participation of a large number of diverse clinics, serving diverse populations and suggest it is feasible to set up and run structured education with adequate support. A low dropout rate and good attendance for the subgroup that attended the intervention, as well as reported impact by this group, strengthens the findings of the study; however, it suggests that uptake and effectiveness might be improved if offered to younger individuals with lower HbA_1c levels.

Although the start-up cost associated with the intervention could be spread over more patients were CASCADE to become routine clinical practice, about four-fifths of the costs of the intervention are closely related to the number of patients. Thus, the costs will differ by number of patients and will be lower for larger clinics. However, these differences will not be large.

Pragmatic RCTs tend to demonstrate high external validity and assess effectiveness rather than efficacy. The intervention has been evaluated in ‘real-life’ and representative settings, with all the challenges that this presents, ultimately providing invaluable information on barriers and opportunities regarding future similar interventions.

Professor David Dunger, Professor of Paediatrics

Department of Paediatrics

University of Cambridge

Box 116, Level 8

Addenbrooke’s Hospital

Hills Road

Cambridge

CB2 0QQ

Professor Jacqueline Barnes, Professor of Psychology

Department of Psychological Sciences

Institute for the Study of Children, Families and Social Issues

Birkbeck

Malet Street

London

WC1E 7HX

Professor Tim Cole, Professor of Medical Statistics

Centre for Paediatric Epidemiology and Biostatistics

University College Institute of Child Health

30 Guilford Street

London

WC1N 1EH

Professor Christopher JH Kelnar, Honorary Professor of Paediatric Endocrinology

University of Edinburgh

9 Easter Belmont Road

Edinburgh

EH12 6EX

Professor Darren Ashcroft, Professor of Pharmacoepidemiology

University of Manchester

School of Pharmacy and Pharmaceutical Sciences

1st Floor, Stopford Building

Oxford Road

Manchester

M13 9PT

Professor Robert Coe, Professor of Education

Director of the Centre for Evaluation and Monitoring (CEM)

Durham University

School of Education

Leazes Road

Durham

DH1 1TA

Dr Chris Patterson, Reader in Medical Statistics

Centre for Public Health

Queen’s University Belfast

Room 3.014, ICS Block B,

Royal Victoria Hospital

Grosvenor Road

Belfast

BT12 6BA