{
  "abstract": "Healthcare continues to rely primarily on reactive safety—responding after harm occurs—rather than proactively identifying and addressing system weaknesses upstream. The purpose of this commentary is to argue that healthcare must treat near misses and early ‘safety signals’ as essential system-level intelligence and must build the infrastructure, processes and culture to learn from them at scale. We define proactive safety as the systematic identification, analysis and mitigation of risks before harm occurs, an approach widely used in other high-reliability industries but still underdeveloped in healthcare. We use the term ‘safety signal’ to refer to any observation—such as a near miss, workaround, recurrent interruption or unexpected system behaviour—that indicates a latent weakness in the system before patient harm occurs. Unlike traditional incident reports focused on actual harm, safety signals capture early warnings embedded in everyday clinical work. We acknowledge that definitions of a ‘near miss’ vary across healthcare systems and regulatory bodies, and in some systems, early-stage interceptions may already fall under the near-miss umbrella. Our use of ‘safety-signal’ is not meant as a replacement for existing terminology, but a conceptual extension intended to highlight earlier and often unrecognised indications of system vulnerability that are typically not captured or prioritised.",
  "authors": [
    {
      "affiliations": [
        "Institute for Healthcare Improvement, Boston, Massachusetts, USA"
      ],
      "name": "Joshua Clark"
    },
    {
      "affiliations": [
        "Thomas Jefferson University, Philadelphia, Pennsylvania, USA"
      ],
      "name": "Oren Guttman"
    },
    {
      "affiliations": [
        "Erasmus Universiteit Rotterdam Erasmus School of Health Policy and Management, Rotterdam, ZH, Netherlands"
      ],
      "name": "Josje Kok"
    }
  ],
  "title": "Safety signals and near misses: exposing the design failures we can prevent",
  "uid": "0eb0fcab-a571-50ae-96b6-81582bdc67e8"
}