MEDEA study: Organizational models for the care of type 2 diabetes mellitus in Andalusia - The endocrinologist's contribution.

INTRODUCTION

Diabetes mellitus is currently a major health care priority due to its high prevalence, complexity, and associated morbidity and mortality. The organization of care for type 2 diabetes mellitus (T2DM) is essential to deliver high-quality care aimed at optimizing clinical outcomes. However, the effectiveness of different health care delivery models remains under evaluation.

MATERIALS AND

METHODS

The MEDEA project was a multicenter, quasi-experimental study that assessed three different models of care for T2DM. The comparison groups consisted of an intensive intervention group, a moderate intervention group, and the remainder of the health care system as a control group. The groups differed according to the degree of hospital-based specialist involvement in primary care. Based on electronic health record data, at-risk patients were identified according to predefined criteria: poor metabolic control (HbA1c >8%), inadequate lipid control (LDL > 100 mg/dL), and risk of hypoglycemia in older adults (patients older than 75 years with HbA1c <7% receiving potentially hypoglycemia-inducing treatments). Outcomes were measured as the percentage of patients achieving targets across these three domains.

RESULTS

No significant differences were observed in mean HbA1c at baseline and at the end of the study among the three intervention groups. However, the proportion of patients achieving HbA1c <8% increased significantly in the intensive intervention group (17.9%) compared with the moderate intervention and control groups. A total of 3,513 patients were included in the intensive intervention group. In the centers belonging to this group (CS1 and CS2), the target HbA1c <8% was achieved in 56.9% and 68.02% of patients, respectively. Regarding deprescribing, pharmacological withdrawal was achieved in 41.9% and 26.9% of eligible patients in CS1 and CS2, respectively. There was a significant increase in the number of patients treated with more than one oral antidiabetic drug (P < 0.001), with increased use of DPP-4 inhibitors (P = 0.000) and SGLT2 inhibitors (P = 0.002). A reduction in basal-bolus insulin therapy was observed (P = 0.01), alongside an increase in the proportion of patients treated with basal insulin analogs (P = 0.01).

CONCLUSIONS

Intensive intervention leads to improved health outcomes in patients with diabetes and is clearly superior to on-demand intervention, which does not demonstrate improvements compared with routine clinical practice. Centralized data more closely reflect the reality of individual centers and enable the identification of at-risk patient groups; however, health records lack specificity and require individualized validation, data cleaning, and clinical contextualization to ensure that analyses accurately represent real-world practice. Mean HbA1c may not be an optimal parameter for assessing intervention impact due to its population distribution. The proportion of patients with HbA1c >8% decreased significantly following coordinated intervention combined with structured external planning and monitoring. Therapeutic inertia also affects deprescribing in older patients; targeted interventions facilitate the withdrawal of potentially hypoglycemia-inducing treatments in vulnerable populations.