Assessing physicians' agreement and the completeness of the decision aid 'arriba Diabetes': a cross-sectional study

Guidelines for type 2 diabetes emphasise individualised treatment goals, yet implementation remains challenging in primary care. To address this, we developed the "arriba Diabetes" software, a patient-centred decision support tool. The software provides individualised recommendations for intensity of treatment based on four inputs: age, comorbidities, treatment burden preference, and risk reduction preference.

In a cross-sectional evaluation study in German primary care, we included 34 general practitioners (GPs) and 152 patients. The primary aim of this study was to evaluate the "arriba Diabetes" software by assessing the agreement between its treatment intensity recommendation and physicians' clinical judgement. Additionally, we explored the cases of disagreement, the distribution of patient-specific inputs and physicians' perspectives on the software's usability.

The "arriba Diabetes" recommendations aligned in 87% of cases with GPs' recommendation, and 87% of the doctors would use "arriba Diabetes" in the future. Patients had a median age of 68 years with a low comorbid load (median 3 on a scale 0-10). Patients expressed a moderate preference for higher treatment burden (median 6 on a scale 0-10) and a high preference to reduce organ complications in the future (median 8 on a scale 0-10). Acceptance of therapy burden correlated positively with the preference to reduce organ damage (Spearman correlation coefficient: +0.49).

Recommendations of "arriba Diabetes" were well aligned with GPs' recommendations. Implementation of the "arriba Diabetes" software has the potential to promote patient-centred and evidence-based diabetes treatment decisions in primary care.