UCSF

Background

Peer health coaching is an effective method of enhancing self-management support in patients with diabetes. It is unclear whether peer health coaching is equally beneficial to all patients with poor glycemic control, or is most effective for subgroups of patients.

Objective

To examine whether the effect of peer health coaching on hemoglobin A1c (A1c) is modified by characteristics that are known to be associated with diabetes control.

Design

Sub-group analyses of randomized control trial.

Participants

Two hundred and ninety nine patients with diabetes receiving care in public health clinics who participated in a randomized controlled trial of peer health coaches.

Main measures

We examined whether the association between study group and change in A1c was modified by differences in patients' demographic, behavioral or psychosocial characteristics. Analyses were adjusted for co-variables associated with change in A1c.

Key results

The effect of coaching on patient A1c was modified by patients' level of self-management and degree of medication adherence as baseline (p=.02, and p=.03 respectively in adjusted models). For participants with "low" self-management (one standard deviation below the mean score), the usual care group experienced a slight increase in A1c (0.3 %), while the health coaching group experienced a decrease (-0.9 %). For participants with "high" self-management (one standard deviation above the mean score), both groups experienced a similar decrease in A1c (usual care group: -1.0 %; health coaching group: -1.1 %). Participants with "low" medication adherence in the usual care group experienced an increase in A1c (0.5 %), while the health coaching group experienced a decrease (-0.8 %). Participants with "high" medication adherence experienced similar decreases (usual care group: -1.1 %; health coaching group: -1.3 %).

Conclusion

Peer health coaching had a larger effect on lowering A1c in patients with low levels of medication adherence and self-management support than in patients with higher levels. Peer health coaching interventions may be most effective if targeted to high-risk patients with diabetes with poor glycemic control and with poor self-management and medication adherence.