Summary: In 602 adults with type 2 diabetes in a Taiwanese shared-care programme, adding a single one-hour diabetes Conversation Map (CM) session to usual care produced a small but statistically significant reduction in HbA1c at three months (standardised β = -0.101). Mediation analysis indicated the benefit was carried by improved diet behaviour (β = -0.126, P < .001), whereas the exercise pathway was not significant (β = -0.023, P = .465).
PICO Summary
| Element | Detail |
|---|---|
| Population | 602 adults with type 2 diabetes enrolled in a diabetes shared-care framework in Taiwan; open-label parallel-group RCT. |
| Intervention | One additional 1-hour, theory-driven diabetes Conversation Map group-education session on top of usual shared-care service (n = 300). |
| Comparison | Usual shared-care service only, without the Conversation Map session (n = 302). |
| Outcome | At 3-month posttest, the CM group had significantly lower HbA1c than control (standardised β = -0.101, baseline-adjusted multivariate autoregression). The CM group also showed greater improvement in diet (β = 0.261) and exercise (β = 0.239) behaviours. Mediation of the HbA1c benefit was significant via diet (β = -0.126, P < .001) but not via exercise (β = -0.023, P = .465). No 95% CI, absolute HbA1c change, ARR or NNT were reported in the abstract; no significant adverse effects were described. |
Diabetes Conversation Map & HbA1c
RCT · type 2 diabetes · 3 months
A single one-hour Conversation Map session modestly lowered three-month HbA1c, and the benefit was carried by improved diet rather than exercise.
Expert Commentary
This adequately powered randomised trial offers a modestly encouraging signal: a single one-hour Conversation Map session, layered onto an existing shared-care programme, was associated with a statistically lower HbA1c at three months. The verdict, however, should be read cautiously. The effect was expressed as a standardised beta of -0.101 rather than an absolute HbA1c change, so the clinical magnitude is uncertain and is plausibly small. The more interesting finding is mechanistic: the glycaemic benefit appears to be mediated by dietary change, while the exercise pathway, despite a measurable behavioural shift, did not reach significance as a mediator. The principal limitation is durability. Outcomes were captured at three months only, and behaviour-change effects from brief educational sessions frequently fade, so it is unknown whether this benefit persists. The intervention was also unavoidably open-label, which can inflate self-reported behaviour. Can I use this with my patients? Cautiously yes, as a low-risk adjunct for motivated adults already in structured diabetes education, with the honest caveat that it is unlikely to substitute for pharmacotherapy. Future work should report absolute HbA1c differences with confidence intervals and longer follow-up. I would welcome a pragmatic replication outside a single national shared-care system before drawing firm conclusions.
References
Lo HJ, Huang JH. Effects of a diabetes conversation map intervention on HbA1c and the differential mediating roles of diet and exercise health behaviors: findings from a large RCT in Taiwan. Ann Behav Med. 2025;59(1). doi:10.1093/abm/kaaf033