Patient safety has improved in recent years, yet preventable harm remains a persistent challenge across health systems. Harm from unsafe care continues to affect a substantial proportion of patients and remains a significant contributor to the global burden of disease. Increasing complexity of care and changing health needs, particularly in ageing populations, have intensified the urgency of coordinated action. Emerging capabilities, including predictive analytics, artificial intelligence and digital engagement tools, offer new opportunities to anticipate and prevent adverse events. Implementation, however, must be carefully managed to avoid unintended consequences. Healthcare organisations rarely achieve sustained progress alone. Strategic partnerships involving providers, patients, families, policymakers and technology developers become increasingly more essential to building comprehensive plans that reduce harm and strengthen safety culture across diverse care settings.
Real-Time Measurement and Predictive Feedback
Patient safety measurement has traditionally relied on voluntary incident reporting after harm has occurred. Such approaches undercount events and are often delayed. Expanding use of electronic health record data has enabled more timely detection of adverse events and identification of risks that may lead to future harm. Within an international community of senior health leaders, three focus areas were identified to advance safety, including provision of real-time measurement feedback through predictive analytics.
Must Read:SAFER Guidance Updates Patient Identification Practices
Automated trigger systems embedded within electronic records have detected substantially more adverse events than voluntary reporting. In one comparison, automated detection identified thousands of adverse events compared with a much smaller number captured through standard reporting. Reductions in inpatient all-cause harm and specific harms were reported following implementation. Automated approaches also reduced the time required for record review compared with manual retrospective processes. Artificial intelligence tools were incorporated to reduce false positives and mitigate alert fatigue.
Implementation requires clear communication of goals and reinforcement through dashboards and feedback on progress. Leaders must foster a positive safety culture to sustain change. Barriers include reliance on incident reporting systems, legal and reputational concerns associated with detecting additional events and limited clinical resources to respond. Sharing safety data at management or quality team levels rather than exclusively with front-line clinicians has been identified as a practical strategy. Legal protections linked to Patient Safety Organisations can facilitate secure data sharing. Trials providing patients and caregivers with access to individual safety dashboards demonstrated lower readmissions and mortality, strong acceptance and no increase in fear responses, indicating potential for shared data to support prevention.
Patient and Family Engagement in Safety
Patients and families increasingly expect active participation in care decisions and safety initiatives. Individuals bring knowledge of their conditions and experiences that can contribute to harm detection, advocacy and co-development of solutions. Digital tools are enabling more self-directed safety approaches, while organisational mechanisms such as Patient and Family Advisory Councils provide structured avenues for engagement.
Across regions including North America, Europe, Australia and Asia, advisory councils have been associated with improvements in clinical outcomes and patient experience. In one initiative focused on sepsis, collaboration with a family advisor supported establishment of a triage screening process in an emergency department, followed by a reduction in risk-adjusted mortality over two years. In Singapore, the SingHealth Patient Advocacy Network was created as a self-driven network of patients and caregivers representing the collective voice of patients. Members have participated in quality improvement, process design, curriculum development and facility projects and supported development of a Plain English Glossary to enhance communication between patients and clinicians. The network has expanded to include dozens of advocates involved in more than 100 projects.
Engagement required outreach to clinical teams, pilot projects to demonstrate value and participation of patient advocates in training programmes to build trust. Ongoing measurement of impact is planned using recognised indicators related to patient and family engagement and patient-reported experience metrics. A toolkit has been developed to share insights with other organisations seeking to strengthen patient advocacy and safety culture. Partnerships with regional and international bodies have facilitated dissemination of engagement models.
Leveraging Technology for Safer Care
Technology plays an expanding role in preventing or limiting patient safety events while strengthening workforce capacity. Medication errors and adverse drug events remain common sources of harm, involving multiple professionals across prescribing, verification, dispensing and administration. Traditional rule-based clinical decision support systems have been associated with alert fatigue and limited ability to detect unanticipated error types.
At Sheba Medical Center, a medication safety initiative implemented a machine learning–based monitoring platform designed to reduce unnecessary alerts and improve clinical relevance. The platform was validated across millions of lives and demonstrated high accuracy with low false-positive rates. Compared with a legacy rule-based system, substantially fewer alerts were generated while clinical relevance and prescribing changes increased. In additional evaluations across academic centres, most reviewed alerts were confirmed as accurate and clinically valid, and many would not have been produced by existing systems. Estimated cost savings associated with prevented adverse events were reported in outpatient settings.
Implementation depends on workflow redesign, attention to human factors and clear communication to ensure adoption without overreliance on technology. Data completeness and quality directly influence performance of machine learning systems. Financial pressures and rising healthcare costs create barriers to investment, although evidence links inpatient harm to negative financial outcomes for hospitals and payers. Analyses of artificial intelligence use in clinical workflows have reported reductions in waiting and triage times in radiology contexts. Leaders must also consider equity implications, as disparities in access to advanced technologies and high-quality data may widen gaps in diagnosis and outcomes.
Reducing preventable harm requires coordinated strategies that combine real-time measurement, patient and family engagement and responsible use of technology. International examples demonstrate that partnerships across organisations and sectors can generate measurable improvements in safety and efficiency. Effective implementation depends on communication, culture change, workflow redesign and attention to data governance and equity. Scaling successful approaches across institutions and countries will require collaboration among health systems and policymakers, alongside continued generation of real-world evidence to support sustained progress in patient safety.
Source: Health Affairs Scholar
Image Credit: iStock
