The great promise of healthcare technology, at least for clinicians, has always been a view of the “whole picture” of a patient’s health. If we could just collect everything in one place, then better diagnosis and treatment would surely follow.
But in practice, the idea that a clinician can review a complete longitudinal record for every patient is a fallacy. The “whole picture” exists in the EHR as thousands of pages of notes, labs, imaging reports, and medication histories that no human can realistically absorb and act on in a short window of time. Every day, clinicians are asked to do the impossible: spot life-threatening details buried in thousands of lines of chart data while moving at a pace no human can sustain.
The gap between what exists in the record and what a clinician can actually see is where harm occurs. Every clinician ends each shift knowing there was something in a patient’s chart they didn’t have time to review. That awareness is causing moral injury at an unsustainable rate. It’s a symptom of a healthcare system that is drowning in data and starving for insights.
As we enter 2026, the conversation around what healthcare technology can achieve, and more importantly, what it should achieve, must change. Instead of showing clinicians the “whole picture,” AI should augment clinicians by surfacing the “right picture” – the critical details that can materially impact diagnosis, risk, or treatment at the bedside.
In medical school, we’re trained to gather every piece of information we need to evaluate a single patient. I’d spend hours – even days – putting together every piece of the puzzle, because that was the time afforded to me. In practice, that luxury doesn’t exist. In the hospital, clinicians have to juggle multiple patients, each with a mountain of data. Whether it’s the ICU, ED, or clinic, there’s rarely time to fully review every record. According to the American Medical Association, 22.5% of physicians spent more than eight hours on the electronic health record outside of normal work hours in 2024.
As an ICU physician, I often rush to the bedside to save a crashing patient. I have very little time to understand the patient’s complete medical history, yet the charts are getting bigger every day. Meanwhile, emergency departments face growing patient volumes, increasing boarding times, and mounting wait times, making these impossible decisions even more frequent.
In a practice environment like this, clinicians can only rely on our experience and the limited information we can digest and analyze with the time we have. Nearly all patient data goes unseen. As many as 900,000 patient data points on a single critical care bed go to waste every hour, and only 3% of healthcare data is ever used.
These blindspots are why my colleagues and I go home every night worried about what we missed. Every diagnostic error that harms or risks harm to a patient creates moral injury. It’s the root cause of increasing burnout and depression among physicians.
No doctor enters medicine wanting to spend more time analyzing data than caring for patients. Yet, that’s the reality of modern healthcare; it’s a system that asks humans to process an impossible amount of information in an unreasonable amount of time. A 2024 study, for example, found it would take clinicians more time than is available in a single day – 26.7 hours – to deliver recommended care guidelines for an average number of patients per day.
It doesn’t need to be this way. AI can illuminate those blind spots – what I’ve been calling “augmented intelligence” – enhancing clinicians’ ability to use their training and knowledge. It can synthesize an entire patient record and surface the handful of insights that truly matter: a lab that contradicts the working diagnosis, a medication the patient forgot to disclose, an overlooked note buried deep in the EHR.
For clinicians, changing the way we practice medicine isn’t easy. We’ve been inundated with new tech over the past few decades – tools we were promised would be transformative, but delivered little more than distractions. As a result, many of us have grown used to practicing within a system riddled with blind spots.
But we’ve reached a breaking point. The cost of inaction – of not altering our workflows, of hanging on to the ways we’ve grown accustomed to working – is being measured in moral injury and patient lives.
If adapting to a new, AI-augmented way of practicing medicine means fewer missed diagnoses, fewer sleepless nights, and more moments where we can actually be present with our patients, that’s a change worth making.
The grand promise of AI in medicine isn’t automation. It’s not “efficiency” or “optimization” – it’s the restoration of the quality of care people deserve and doctors expect of themselves. AI has already proven valuable in reducing administrative burdens – documenting visits, summarizing notes, improving billing – but little of that meaningfully improves patient care. The real transformation will come when AI moves beyond paperwork and into clinical decision-making, helping physicians see what we might otherwise miss.
When AI can analyze every data point in a chart and bring forward the insights that matter most, it strengthens our ability to save lives. It buffs away the callouses of moral injury that have built up over years of practicing under impossible conditions.
If AI can surface key insights from the flood of patient data, care gets safer and the people delivering it can finally breathe again. The goal of healthcare technology should no longer be to provide clinicians more data. It should be to help us see the data that matters most in moments of care.
No comments:
Post a Comment