Hypothesis \& Aim:
The central aim is to determine whether communication of data structured against key diabetes process and outcomes directly to people with diabetes, with an emphasis on enabling, informing and empowering their engagement with service provision will drive improvement in the key parameters of structured diabetes care access, process and outcomes.
Thus we hypothesise that the communicating letter, which is proven to be readable, informative, valuable and desired in user acceptance analysis can be evidenced to drive engagement of users as demonstrated by measures of access and process and that subsequently attainment of access and process permit the achievement of diabetes clinical outcomes.
Project Design
The trial is designed as a prospective cluster randomised controlled trial of an intervention (structured reporting to patients) on key outcomes (as defined below). All patients with diabetes in Wolverhampton will be randomly allocated to receiving 3 mailings at 0, 3 and 6 months (Active Group) or 1 mailing at 3 months (Control Group) of their structured diabetes report (My Diabetes, My Information, My Plan). Using our pan district integrated diabetes information system, data parameters will be observed at 0, 3, 6 and 12 monthly intervals over the 12 months following the delivery of the report.
The objective will be to determine whether the repeated provision of structured, specific and annotated process and outcome information promotes patients to improve their own health care whether by their own actions and /or by seeking support, advice and help and this will be measured using a Process Score and an Outcome Score - emplacing that it is known that surrogate markers for clinical outcomes in diabetes (eg HbA1c) cannot be improved without service Access and completion of process.
The project will run for 2 years - 6 months set up, 1 year active implementation, 6 months analysis, publication, dissemination and closure. Throughout, evidenced effective methodology will be transitioned into standard NHS process.
Project Feasibility:
The Wolverhampton diabetes service is well integrated across primary and secondary care. A centralized pan district register is established. Data linkage between multiple systems is in place, including direct links between the local GP data warehouse (Graphnet) and all specialist / secondary care systems. Data management and governance systems are in place. Data quality is subject to rolling audit and is consistently at \>97% accuracy. Effective use of such data has been demonstrated in care pathways - e.g. the diabetes retinal screening programme (national EQA) and the innovative foot care pathway where Wolverhampton amputation rates are amongst the lowest in the country. The local service has a good track record in clinical research. The project has full support of all key NHS groupings including the local Diabetes GP Commissioning Lead (Dr M Sidhu) and the specialist services Clinical Director. All relevant significant others have been engaged in our local Diabetes Implementation Group - including our large local user group (Wolverhampton Action 4 Diabetes) and the regional office of Diabetes UK. We have run a pilot to test the feasibility and effectiveness of our key intervention (structured diabetes information report). This involved consultation with a diverse diabetes user group and conducting a questionnaire survey with 50 patients attending hospital diabetes clinics. Following translation of learning from that pilot into further NHS process, the project is now fully achievable as proposed.
Data Analysis:
Data will be analysed in SPSS for with and between group (paired and unpaired) effects of the individual process and outcome parameters and the summary scores. Thus access process and outcome scores (as defined) will be calculated at baseline and at 3 monthly intervals over 1 year and analysed to determine within and between group effects by 2 way analysis of variance (a group-by-time repeated measures ANOVA) Significance is set at the 5 % level. Professorial level academic statistical advice is available within the local diabetes research group. A power calculation to predict sample size is not warranted since the entire population (not a sample) is to be studied.
