The impact of second-level specialized care on hospitalization in persons with diabetes: a multilevel population-based study

Aims  We evaluated whether differences in the use of specialized care have an impact on rates of hospitalization for diabetes. Methods  In 2001 we determined the number of hours of second-level diabetes care provided by local health units (LHU) of the Piemonte Region (Italy) and created an indicator of the mean weekly number of hours of care per 1000 residents for each LHU. From the database of the Piemonte Hospital Information System, we extracted all hospitalizations for 20–75-year-old residents with a main discharge diagnosis of diabetes mellitus (n = 3457). For each LHU, we calculated the hospitalization rate, the percentage of unplanned hospital admissions, the mean length of hospital stay, the percentage of day-hospital admissions and the percentage of re-admissions for diabetes-related complications within 6 months. The association between the indicators of specialized care and of hospital care was studied using two-level generalized hierarchical linear regression models (level 1: patient; level 2: LHU), taking into account the clustered nature of the data. Age, educational level and an indicator of disease severity were used as adjustment parameters. Results  In the tertile of LHUs that provided the greatest number of hours of diabetes care, we observed, compared with the lowest tertile fewer unplanned hospital admissions [odds ratio (OR) 0.37; 95% confidence interval (CI) 0.20–0.67], greater day-hospital use (OR 1.99; 0.72–5.49) and a lower mean duration of hospital stay (coefficient −0.26; 95% CI −0.45 to −0.06), independently of the socio-economic level, which was a separate risk factor. Conclusions  The intensity of specialized diabetes care greatly influences the characteristics of hospitalization.