Ethical and governance approval
Multi-centre approval has been granted by Research Ethics Committee for Wales (07/MRE09/59). Site-specific approval has been granted by local RECs at all trial sites and all participating Acute Trust Research and Development Departments.
This is a multi-centre randomised controlled trial, for children aged 0-17 years who have been newly diagnosed with type 1 diabetes and their parents. Participants will be randomised to receive either hospital management (minimum of 3 overnight stays in hospital) or home management (no overnight stays in hospital).
Clinic and patient selection
This study will be carried out in eight UK paediatric diabetes centres. Criteria for selection are based on size of the paediatric diabetes centres (a minimum of approximately 30/40 newly diagnosed patients per year) and geographical placement, to ensure balanced distribution of socio-economic factors across the UK. The paediatric diabetes centres will be based within NHS secondary care (paediatric wards and outpatient clinics). The paediatric diabetes team at each centre will comprise at least one consultant paediatrician with an interest in diabetes, a paediatric diabetes nurse and a paediatric dietician. Collaboration and recruitment is already agreed with teams in the UK from Cardiff, Southampton, Hull, Liverpool, Cambridge, Belfast, Newcastle and Nottingham.
Within each centre, a study-specific Project Nurse will be employed, whose role will be to co-ordinate nursing management of participants and data collection.
Inclusion and exclusion criteria
Participants will only be entered into the trial if they meet the following inclusion and exclusion criteria (see Table )
Participant inclusion/exclusion criteria
Patients (aged ≥8 years) and all parents/carers of patients (aged 0-17 years), will be given information about the study by a member of the clinical team to read whilst in the assessment unit/paediatric ward. They will have time to consider the study while blood tests are taken to confirm the clinical diagnosis.
Once informed consent/assent is obtained, patients will be remotely randomised using an automated telephone system operational 24 hours a day. Patients will be randomised to either 'Home Management' or 'Hospital Management'.
Randomisation will be stratified by centre and balanced using randomly chosen permuted blocks. The randomisation ratio is 1:1
Process of care for all patients
Hospitalised and home-managed children and their parents will receive education and support from the project nurse and local paediatric diabetes team at each centre. Both cohorts will receive written information about diabetes, and a structured diabetes education programme. Education programmes, including dietary advice, will be standardized across centres as far as possible, but it is anticipated that minimal variability will be found concerning the educational content of programmes currently used at individual centres. Children will be commenced on an insulin regimen and delivery system that are deemed appropriate by teams at individual centres, and will be asked to undertake three to four blood glucose measurements a day before meals for the first 4 weeks of the study. All families will be given an appointment to attend the next appropriate diabetes clinic, will receive continued support from health professionals through telephone contact and clinic visits, and will be able to access help and advice out of office hours. Throughout the study, the paediatric diabetes team at each centre will comprise at least one consultant paediatrician, a paediatric diabetes nurse and a paediatric dietician.
To assist centres to deliver the study, a manual was developed providing guidance in key areas such as initial diagnosis, recruitment, home management and hospital management.
The study will use a pragmatic approach to home management in order to accommodate the individual needs of participating centres. It would be inappropriate to undertake the study using one specific model that, if found to be beneficial, could not be subsequently adopted by centres due to, for example, differing insulin regimens (e.g. twice daily insulin or multiple injection regimens) or particular geographical needs (e.g. centres covering large rural areas).
The standard elements of home management that will be common to all participating centres are:
- Discharge home on the day of diagnosis with no overnight admission to hospital
- All treatment, education and support will be delivered at home or on an outpatient basis (attending ward/clinic for no longer than 2 hours for supervision of injections as necessary according to local need) for a minimum of three days (at least six supervised injections)
- Dietetic education will be provided at home or as outpatients, with continued dietetic support provided in clinic.
Children will be admitted to hospital at diagnosis for a minimum of three nights (receiving at least six supervised injections while hospitalised). During their inpatient stay, families will receive treatment, education and support in the ward environment. The paediatric dietician will provide dietetic education to the child and family on the ward, and provide continued dietetic support at clinic visits. The project nurse at each centre will undertake at least one home visit (on or soon after the day of discharge at a time when insulin is due), with home visits continued as required according to the needs of individual families.
Frequency & duration of follow-up
There will be three follow up appointments carried out at 3, 12 and 24 months as part of the participant's routine clinic visits.
Questionnaires will be sent to parents, and age-appropriate questionnaires to children, at 1, 12 and 24 month intervals.
Data concerning length of stay at diagnosis, glycaemic control (HbA1c), growth (weight, height, BMI), readmissions, adverse events (e.g. hypoglycaemia), home and clinic visits, school attendance, self-care activity and parents' time off work and travel costs (in relation to the child's diabetes) will be collected at clinic visits for 24 months after diagnosis. This will be at routine visits, which take place every 3-4 months.
There will also be standardised follow up visits at 3, 12 and 24 months. At diagnosis and at months 3, 12 and 24, extra blood from the initial sample taken to measure the patient HbA1c (glycosylated haemoglobin) level, will be taken and sent to a centralised laboratory (Diabetes Research Network Wales Laboratory, Llandough Hospital) for measurement of HbA1c concentrations. See Table .
Primary & secondary outcomes
The primary outcome measure for the trial is HbA1c of patients at two years following diagnosis. HbA1c was selected because it is an objective measure of glycaemic control used to inform clinical practice and national and international policies and guidelines.
The secondary outcome measures for patients include growth, adverse events, psychological assessment of quality of life, coping with diabetes, diabetes knowledge, satisfaction and time off school. Secondary outcome measures for parents include anxiety, coping with diabetes, diabetes knowledge, satisfaction and time off work. Mean HbA1c at 3 and 12 months will be used to assess shorter term intervention effects. Additional qualitative outcome measures will be taken from health professionals' experience of both approaches to care. Finally, total health service costs, including hospitalisation, home/clinic visits and use of other NHS resources will be evaluated.
Quality of life measures to be used will be adapted from:-
- Issues in Coping with IDDM Scale - parent version[15
- Spielburger Anxiety Scale (short version)[16
- Diabetes Knowledge Scale - parent version [17
- Quality of Life (PedsQol) - parent proxy version [18
- Issues in Coping with IDDM Scale - child version[15
- Diabetes Knowledge Scale - child version[17
- Quality of Life (PedsQol) - child version[18
In addition, there is no available validated measure of the impact of a diagnosis of diabetes on social activity and independence. Therefore, a Social Activity and Independence Questionnaire (SAIQ) will be developed and validated for use as a secondary outcome measure. The development phase will include item generation through interviews with children and parents living with diabetes and those items will be included for all participants as part of the main outcome questionnaires with parents and children.
In order for a randomised trial to have 80% power to detect an effect size of 0.4 (difference in mean HbA1c of 0.5% with an SD of 1.3% [7
]) at a 5% significance level, a total of 200 patients would be required. To allow for loss to follow up of 17%, it is aimed to recruit 240 children. A previous study in Canada which evaluated reducing the amount of in-patient time at diagnosis showed a difference of 0.7% in mean HbA1c at two years[8
]. Loss to follow up for the primary outcome of HbA1c should be small, as all of these patients will be attending clinic on a regular basis where HbA1c is monitored for clinical purposes.
Primary analysis will be intention to treat and will compare HbA1c between the two groups at the 24 month follow-up time point using ANCOVA. Baseline HbA1c will be included as a covariate. These analyses will be corrected for any clustering of outcomes within a clinic. Secondary outcomes analyses will compare the two groups using repeated measures ANOVA. These analyses will also be corrected for any clustering of outcomes within a clinic and baseline levels. Secondary analysis of the primary outcome using repeated measures ANOVA will be carried out using the 3, 12 and 24 month HbA1c values for the two groups and will also involve a more detailed exploratory analysis of the impact of clinic level factors on HbA1c outcome using a two level linear multi-level regression model.
Qualitative data analysis is an on-going rather than discrete activity. Data from the qualitative interviews undertaken with a subgroup of children older than 8 years of age and parents will be subject to thematic analysis, which comprises a number of steps: 1) interviews will be audio-recorded and transcribed verbatim, 2) patterns of experience will be listed, which may arise from direct quotes or the paraphrasing of common ideas, 3) data relating to the already classified patterns will be identified, 4) related patterns will be combined and catalogued into sub-themes, and 5) themes will be developed. Themes that emerge from participants' accounts will be coded and subsequently collated to form a comprehensive picture of their collective experience. It is anticipated that this will provide a detailed, in-depth insight into the thoughts and experiences of the participants. Data will be explored to allow identification and comparison of similarities and differences between cases and between the two arms of the study. The analytical process will be undertaken in a way that ensures that the integrity of the original document is kept intact.
Cost Effectiveness Analysis
A cost effectiveness analysis will assess direct (management at diagnosis) and indirect (subsequent) health service costs against the primary outcome (HBa1C). All NHS resource use, including inpatient admissions, insulin use, contacts with the diabetes team, investigations, attendances at accident and emergency departments, ambulance journeys, contacts with general practitioners and other health professionals will be monitored prospectively and valued by standard methods [19
] using national unit costs supplemented where necessary by cost information from participating centres. Non-NHS costs such as parental time off work and travel are also being assessed but will be reported separately. Cost effectiveness results will be reported in the form of an incremental cost effectiveness ratio unless either form of patient management dominates (lower cost with greater effect). A series of one-way sensitivity analyses will test the sensitivity of the conclusions to changes in the main base case assumptions used.
Given the multiple objectives of a management at diagnosis service for children and the consequent importance of the secondary outcomes in this study (psychological adjustment, coping, adaptation to diagnosis), a costs-consequences analysis will also be undertaken.