OBJECTIVES : Government policy in Denmark has been to push monitoring of diabetes patients away from specialized hospital clinics into general practice settings. This study investigates composition of hospital- and GP-monitored diabetes patients, and treatment cost-efficiency implications of moving monitoring into primary care.

METHODS : Using administrative databases of hospitalizations, primary care services, and prescriptions, we algorithmically identified diabetes patients, and whether they were hospital-monitored. Socio-demographics, comorbidities, and annual treatment cost were compared across monitoring settings. Propensity score matching (PSM) was used to estimate impacts on annual healthcare costs of being monitored in hospital (Average Treatment Effect on the Treated - ATET).

RESULTS : In 2015, 288578 diabetes patients were identified, and 41976 were identified as being regularly monitored in hospital. Hospital-monitored patients: had significantly higher Charlson Comorbidity Index scores (3.4 vs. 1.9); were younger (58 vs 63), but with longer histories of diabetes (14 vs. 8); more likely to be male (58% vs. 51%); had higher levels of education; and were less likely to be immigrants. Hospital-based patients had higher annual costs for medications, hospital inpatient and outpatient admissions, and non-hospital-based specialists; only general practice costs were lower. Estimated ATET on total costs for hospital-monitored patients was 7873DKK; positively driven by outpatient cost (9120DKK) and prescriptions (1266DKK); and negatively driven by ATET estimates for general practice (-520DKK) and inpatient costs (-2047DKK). Estimated ATET for costs of hospitalizations originating as emergency visits were negative (-1103DKK). No significant impact on costs for non-hospital-based specialist services were identified.

CONCLUSIONS : This study finds evidence that average hospital-monitored diabetes patients in Denmark are in poorer health than GP-monitored patients, and the quality of healthcare delivered to hospital-monitored patients is higher. However, PSM results suggest that there are possible net cost savings to be realized by moving care of some hospital-monitored patients to primary care.