Summary: In a secondary analysis of the Japanese J-DOIT2 cohort, a prior history of dropping out of diabetes care was the strongest predictor of future dropout, and among those with no such history, very high HbA1c (≥10%) signalled raised dropout risk.
PICO Summary
| Element | Detail |
|---|---|
| Population | 996 patients with type 2 diabetes in the non-intervention group of the J-DOIT2 large-scale trial; Japan. |
| Exposure | Prior dropout history and elevated HbA1c. |
| Comparison | No dropout history; HbA1c 6.0–6.9%. |
| Outcome | Dropout history strongly predicted future dropout (adjusted HR 3.59; 95% CI 2.25–5.71). Overall HbA1c was not associated with dropout, but among the 855 without prior dropout, HbA1c ≥10% carried higher risk (HR 3.76; 95% CI 1.29–10.9). |
Expert Commentary
This is a practical, clinically intuitive analysis, and I value it because care continuity is an under-measured determinant of diabetes outcomes that we tend to neglect in favour of glycaemic targets. The central finding, that past dropout strongly predicts future dropout, fits the broader truth that disengagement reflects stable circumstances and beliefs rather than one-off events, and the secondary signal, that a very high HbA1c flags disengagement risk in someone not previously lost to follow-up, is a useful reframing of an extreme number as a relationship warning, not just a pharmacological one. My caveats are appropriate to the design. It is observational, so neither marker causes dropout, both are likely proxies for deeper factors such as distress, instability, or access barriers, and the definition of dropout and the Japanese care setting limit transfer to other systems. Can I use this with my patients? Yes, operationally. It supports flagging patients with prior disengagement or a markedly elevated HbA1c for proactive outreach, more frequent early contact, and attention to the barriers behind missed visits, rather than simply intensifying medication. Continuity itself deserves to be treated as an outcome.
References
Goto A, Hayashino Y, Takamoto I, et al. Association of dropout history and HbA1c levels with subsequent dropout risk in patients with diabetes: a secondary analysis of the Japan Diabetes Outcome Intervention Trial-2 Large-Scale Trial 008 (J-DOIT2-LT008). BMJ Open. 2025;15(8):e092013. doi:10.1136/bmjopen-2024-092013