In 2000, the Institute of Medicine (IOM) published a landmark report titled To Err is Human: Building a Safer Health System. At the time, it was estimated that roughly 98,000 people died in any given year from medical errors in hospitals – asserting that more people died annually from medication errors than workplace injuries.
What it did was shine a light on a common misconception: that healthcare delivery is a perfected artform. After all, the stakes are as high as it gets; there is no room for error. Yet there is. Because, as the title suggests, to err is human. The healthcare system isn’t some ethereal, impenetrable and omniscient being. It’s populated with humans. Humans with an indescribably heavy responsibility: the wellbeing of others.
The report doesn’t claim that people are the problem in healthcare, but rather, as pointed out astutely in an NIH summary of the report, “that good people are working in bad systems that need to be made safer.”
It’s been nearly 25 years since this report elevated the stark realities and complexities of healthcare delivery into the spotlight, and much has changed since then: the advent of EHRs, telemedicine and remote monitoring, personalized medicine and genomics, robotic surgeries and minimally invasive techniques, and automation tools that aid in patient management.
However, systems are still inefficient, and it remains a major problem.
It seems things continue to spiral. I have referred to this data before, but it bears repeating: a recent study conducted by Johns Hopkins quantified the impact of misdiagnosis, finding it as the cause for up to 795,000 deaths and serious medical harms.
But it doesn’t have to stay that way. In the same manner that vaccines tamed smallpox in the late 18th century and medical imaging revolutionized diagnostic possibilities, we’re at the forefront of another breakthrough – AI – transforming how physicians communicate, diagnose and manage patient care. And it does this not by working in isolation, but by addressing the health system as a whole.
Systems of Inefficiency: The Symptoms and the Diagnosis
The problem of fragmented care predates the groundbreaking 1999 IOM report. As early as the mid-19th century, Florence Nightingale’s work (Notes on Matters Affecting the Health, Efficiency and Hospital Administration of the British Army) during the Crimean War showed what happens when there are not standardized processes and protocols.
Even today, many health systems still deal with interoperability challenges – especially from referring practices, un-integrated systems and data normalization issues that all contribute to the current state of disjointed patient data. This results in delays and inefficiencies, as information from different departments or providers may not be readily accessible or synchronized. For example, imagine a patient admitted for surgery who undergoes multiple diagnostic tests. The results from the radiology department might reside in one system, while lab results are documented elsewhere. The lack of interoperability forces clinicians to chase down information, resulting in delayed decision-making and/or redundant testing. This is undoubtedly a reality faced by health systems; a Becker’s survey found that 55% of organizations “rely on more than 50 separate point solutions to manage their healthcare operations.”
From the patient’s perspective, this has created the broken experience they face today. Picture a scenario where a patient must repeatedly share the same medical history with different specialists because their information isn’t accessible to different teams. This not only leads to frustration, but also risks errors in communication–one specialist may overlook a critical piece of the patient’s history that another is well aware of, resulting in delayed or inappropriate care.
Fragmentation is not merely a symptom of inefficiency but its root cause. Poor information-sharing practices manifest in adverse events, delays in treatment and a general decline in care quality. For instance, one study found that care transitions between providers or settings exacerbated by poor communication saw 20% of patients experience an adverse event within three weeks of discharge, with communication failures being a primary contributor.
These inefficiencies emphasize the urgent need to address this core issue to improve patient care. Fragmentation isn’t just an operational flaw; it’s the diagnosis that explains why health systems can struggle to deliver consistent, high-quality care.
AI’s Role in Unifying Healthcare
While AI is not the end-all be-all solution for healthcare, it offers powerful tools to streamline operations across various specialties and facilities including:
- Data Integration and Interoperability: AI can consolidate data from various sources
- Workflow Automation: By automating routine tasks, providers are freed up to focus on patient care and clinical decision making
- Communication and Coordination: AI alerts relevant clinicians in real-time, ensuring critical information is relayed promptly for timely intervention.
Clinically validated studies have shown that post-AI implementation yields greater efficiency within clinical workflows, including:
Addressing Fragmentation: AI as a Solution
It’s time to push for better diagnostic practices and systemic changes. This is not about incremental improvements but about fundamentally overhauling the healthcare landscape. We need to build a better system where technology and human-centered care converge to save lives and improve outcomes.
As we’ve seen throughout history, healthcare’s evolution is marked by transformative breakthroughs, and AI represents the cusp of another one. But this transformation must be thoughtful, rooted in the understanding that healthcare is, and always will be, a profoundly human endeavor.
The seamless integration of AI not only empowers individual healthcare practitioners but orchestrates a harmonious, efficient and patient-centric healthcare ecosystem. This orchestration goes beyond following established protocols, which are part and parcel of today’s fragmented healthcare reality; it dynamically adapts to the evolving clinical landscape, ensuring that no patient is lost to follow-up and that every clinical signal is acted upon promptly and effectively.
Just as the IOM report nearly 25 years ago challenged us to rethink healthcare safety, today we stand at another crossroads. By leveraging AI wisely, we can help alleviate some of the pressures of fragmentation and build a system that truly “does no harm” – a system that works for everyone.
Achieving this unity is the challenge facing health systems today. Everything else mentioned in this article is ancillary to that. Through a unified, effective communication and integrated systems can we take another step towards adhering to the principle of “do no harm” – a principle that, in the end, helps us not just be better clinicians, but better human beings.
