Improving self-referral for diabetes care following hypoglycaemic emergencies: a feasibility study with linked...

Hypoglycaemia is a common and potentially life threatening consequence of insulin and sulphonylurea treated Diabetes. Some severe hypoglycaemic events result in emergency ambulance attendance. Many of these patients are treated at home and do not require immediate transportation to an Emergency Department. However only 27-37 % of patients then follow up their care with a diabetes specialist. Consequently repeat severe hypoglycaemic events occur.

The intervention was implemented for 8 months, using a prospective cohort design with a historic control, in one Scottish Health Board in 2012. Data was collected using postal survey questionnaires to patients and ambulance clinicians, telephone survey follow-up questions to patients. Scottish Ambulance Service electronic records were linked with the SCI-Diabetes database of patient records to enable objective measurement of follow-up behaviour.

Ambulance clinicians' (n = 92) awareness of the intervention was high and both the prompt card and telephone call components of the intervention were delivered to most eligible patients. The intervention was perceived as highly acceptable to patients (n = 37), and very useful by both patients and ambulance clinicians. However, comparison of patient follow-up behaviours using linked-data (n = 205), suggest that the intervention was unsuccessful in improving rates of patients' following up their care.

This study shows that the intervention is implementable, highly acceptable to patients, and considered very useful by both patients and ambulance clinicians. However, preliminary evidence of effectiveness is not encouraging. The study's novel use of linking existing clinical data for outcome measurement exposed challenges in the feasibility of using this data for intervention development and evaluation. Future research should examine challenges to the successful testing and effectiveness of the intervention. Revisions are likely to be required, both to study design and the optimisation of the intervention's content and components.