A 12-month quality improvement project at a socioeconomically deprived Hampshire practice offered monthly group education to adults with type 2 diabetes on insulin or GLP-1 agonists. Fifteen patients attended at least one session. Mean HbA1c fell 12.4% among those attending four or more sessions, 5.6% after three, 2.4% after two, and rose 1.8% after one. All nine questionnaire respondents reported better understanding; peer support was the most valued element. Attendance was lowest among patients with multiple comorbidities and mental health needs.
The dose-response pattern looks persuasive but should not be read as proof that sessions lowered HbA1c. With only fifteen participants, no control group and descriptive analysis, the likeliest explanation is selection: motivated, engaged patients both keep attending and improve their glycaemic control, while the one-session group may simply have disengaged. Regression to the mean and concurrent medication changes could also contribute. Larger structured programmes, such as the UK's DESMOND, have shown modest HbA1c benefits in randomised settings, and peer-support trials generally show small but real effects, so the concept is plausible. The more useful finding is the engagement gap: the highest-need patients, who stand to gain most, attended least. The authors' proposed hybrid model pairing group and individual support addresses that, but remains untested. This is a preprint, not yet peer-reviewed, and the results may change. It is an incremental, hypothesis-generating service report, not evidence of efficacy.