Andre Ramalho - MEDCIDS - Department of Community Medicine, Information and Health Decision Sciences, Faculty of Medicine, University of Porto, Porto, Portugal
Julio Souza - MEDCIDS - Department of Community Medicine, Information and Health Decision Sciences, Faculty of Medicine, University of Porto, Porto, Portugal
Pedro Castro - USF Camélias, ACES Gaia (Grande Porto VII - ARS Norte), Vila Nova de Gaia, Portugal
Mariana Lobo - MEDCIDS - Department of Community Medicine, Information and Health Decision Sciences, Faculty of Medicine, University of Porto, Porto, Portugal
Paulo Santos - MEDCIDS - Department of Community Medicine, Information and Health Decision Sciences, Faculty of Medicine, University of Porto, Porto, Portugal
Alberto Freitas - MEDCIDS - Department of Community Medicine, Information and Health Decision Sciences, Faculty of Medicine, University of Porto, Porto, Portugal

Background  Diabetes mellitus (DM) is a worldwide public health priority. The increasing prevalence and the budget constraints force to have effective healthcare, especially at the primary healthcare (PHC) level. We aim to assess primary care efficiency considering the best use of human resources to produce optimal diabetes care in terms of prevention quality indicators (PQIs) rates across national ACES (health centre groupings). Methods  We conducted a two-stage data envelopment analysis (DEA) to assess the technical efficiency of ۵۴ Portuguese primary care health centre groupings for the ۲۰۱۶-۲۰۱۷ biennium. In the first stage, efficiency scores were obtained through five output-oriented DEA models under vector return to scale (VRS) assumption, using three input variables representing key primary care human resources and one output representing each one of the five PQIs related to diabetes. In the second stage, Tobit regression models were estimated to assess the determinants of primary care efficiency in diabetes care.Results  A total of ۱۳ ACES reached the efficiency frontier. Better managing human resources could reduce PQI rates by ۵۲.۳% in ۲۰۱۶ and ۴۹.۱% in ۲۰۱۷. Higher proportion of patients under ۶۵ years old and better controlled with a hemoglobin A۱c (HbA۱c) ≤۶.۵% were associated with better efficiency in diabetes care, whereas higher prevalence of DM and unemployment worsened hospitalizations rates by diabetes short-term complications and lower-extremity amputation. Conclusion  Inefficiency in DM care was found in most of the primary care settings which can substantially improve the avoidable hospitalization rates by DM using their current level human resources. These findings help to improve diabetes care by targeting human resources at primary care level, which should be integrated into performance assessments considering broader and integrated scopes.

Diabetes Mellitus, Health Policy, Primary Care, Preventable Admission, Efficiency, Data Envelopment Analysis