A hospital’s patient safety initiatives are only as effective as the tools used to track and analyze incidents. Despite significant progress over the past two decades following the 2005 Patient Safety and Quality Improvement Act, safety reporting in many organizations still barely scratches the surface of meaningful data collection.
Near-misses and minor harm events often go undetected, removing valuable learning opportunities for clinicians and decision-makers alike. Much like the automotive industry’s investment in blind spot detection, healthcare should prioritize technology that alerts leaders to unseen risks, ensuring stronger and more effective safety reporting.
The hidden risks of incomplete reporting
Underreporting continues to be a significant challenge for patient safety. By one estimate, hospitals reported just 14% of patient harm events experienced by Medicare beneficiaries. Part of the reason safety events go unreported is that the process for logging a safety event is highly burdensome. When an adverse event occurs – such as a fall, burn, infection or medication error, or even a “near-miss” event – hospital staff must fill out lengthy forms manually, which is time-consuming, takes away from frontline patient care and leads to data inconsistencies.
This means that hospitals lack sufficient data to gain meaningful insights that can help them enhance patient safety and care quality. With only a fragmented view of patient safety, they lack visibility into the root causes and trends that impact care quality.
This safety data gap creates a cycle of persistent risk. Hospitals and health systems need a more simplified, comprehensive and intuitive reporting system that gives them visibility into those safety blind spots without burdening clinicians with cumbersome manual tasks.
How AI can transform safety event reporting
Artificial intelligence (AI) can fundamentally reshape event reporting and transform how we capture and evaluate safety incidents. For example, leveraging AI tools can improve the speed, accuracy and ease with which hospitals and health systems document harm and near-miss events, allowing staff to generate more thorough reports without sacrificing valuable time spent caring for patients.
When generative AI tools are applied to unstructured data such as voice notes or narrative descriptions of an event, they can automatically populate an incident report with consistency and precision. Automating incident reporting using AI tools not only reduces manual tasks for frontline staff, it encourages clinicians to report more incidents. Staff don’t have to worry about losing valuable hours filling out reports because AI has streamlined the process for them. If a patient almost receives the wrong dose of medication, staff are more likely to report it. These “near-misses” can give leaders a more complete picture of patient safety beyond incidents involving actual harm.
AI tools can also improve data consistency and quality to eliminate subjective interpretation and reduce bias in manual reporting.
As an example, if a patient falls out of bed and experiences an injury, clinicians, nurses and other staff may have different evaluations of the seriousness of the incident. AI tools do not have this bias. They will categorize the severity of an event strictly by medical definitions, giving hospital leaders a more accurate picture of what actually happened.
How AI tools make incident data actionable
For hospital and health system leaders, AI doesn’t just streamline the flow of data – it makes that data actionable. With automated analysis for large volumes of text-heavy reports, AI can surface key patterns and present a concise overview of narrative insights, providing hospital leaders with a holistic view of patient safety to inform decision-making at scale.
For instance, if AI tools highlight that a hospital’s medical-surgical unit is experiencing more falls than is typical, hospital leaders could determine that intervention measures are needed in that unit. If there are more incidents occurring at certain times of the day – preventable incidents such as diagnostic errors, delayed treatment or poor communication – these may indicate broader systemic factors that may need to be addressed by giving staff working those shifts additional training on treatment protocols.
AI transforms incident data with the ability to analyze large volumes of reports, surface patterns and highlight problem areas. These insights allow hospital leaders to not only identify trends but also anticipate risks and take more targeted steps to improve patient safety.
Eliminating patient safety blind spots for a safer future
As hospitals and health systems care for an increasingly complex patient population, with higher rates of chronic conditions, they should explore every opportunity to streamline manual processes.
Every incident that goes unreported because a clinician was too busy to fill out a lengthy form is a missed opportunity to understand where those invisible risks are. Reporting tools powered by generative AI can eliminate blind spots in safety reporting, shed light on invisible risks and enable hospitals to take meaningful steps toward building a safer future for patients.
There are valid concerns that as healthcare organizations continue to scale their use of AI tools, there will be unintended consequences. It’s important for healthcare organizations to stay informed of AI’s rapid developments. Companies partnering with hospitals and health systems must stay vigilant to build a clear understanding of the potential risks of AI innovation and ensure responsible implementation and use.
By keeping a human in the loop for continuous evaluation and transparency, organizations can work to mitigate key risks and enable AI to serve as a tool to enhance clinical judgment, rather than replace it.
As we’ve learned from other industries, blind spots are dangerous, but they can be detected and mitigated with the right tools in place. For healthcare, it’s time to embrace technology innovation and put AI-powered safety reporting in the driver’s seat.
Timothy McDonald, MD, JD, is the chief patient safety and risk officer for RLDatix and a professor of Law at Loyola University – Chicago. Tim is a physician-attorney who has assisted hundreds of hospitals and health systems in implementing a principled approach to unexpected events. He is dedicated to communicating honestly to patients and families, providing peer support within the healthcare team and using software technology to learn and improve following patient harm events, including identifying opportunities to reduce disparities in healthcare. His federally funded research has focused on improving the quality of care while mitigating medical liability issues and establishing teaching methodologies for all levels and professions in healthcare and law.
He has published dozens of articles in high-impact peer-reviewed journals such as Health Affairs, Health Services Research, and the New England Journal of Medicine. His published work has been cited by the President’s Council of Advisors on Science and Technology Patient Safety Report and CMS’s recently published Patient Safety Structural Measures. He is a featured TEDx speaker for his talk on “Healing After Harm in Healthcare.”
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
