Building reliable evidence from real-world data: Needs, methods, cautiousness and recommendations.

National healthcare systems of advanced countries, including Italy, widely agree on the approach whereby public healthcare decisions should be driven by available evidence on effectiveness and safety of therapeutics. It is equally accepted that randomized controlled clinical trials (RCTs), although universally recognised as the most robust "evidence generators", are insufficient for guiding the decision-making process since they are intrinsically unsuited to capture the impact of treatments in routine clinical practice. The complexity of treatments, as well as the demographic and clinical heterogeneity of patients receiving the treatments, and the long period of many treatments, explain the gap between the evidence generated in the controlled, but artificial, setting of RCTs and their current impact in the real world. The so-called pragmatic RCTs, despite guaranteeing greater flexibility compared to conventional trials, are not always able to reduce this gap. This explains the growing interest in the development of methods able to produce evidence on the real-world impact of care pathways (i.e., real-world evidence). Among them, those based on the Electronic Healthcare Records (EHRs), as the databases on the healthcare services of the National Health System provided to beneficiaries, known as Healthcare Utilization Databases (HCU), are becoming established and receiving increasing attention from the scientific community and healthcare decision-makers. We described the research areas in which HCU databases may be particularly useful, jointly with strength, weakness and potential of this approach. It is concluded that HCU data cannot substitute RCTs but they can usefully complement RCT data for adequately supporting healthcare decision-makers.