In 1960, British physician Lord Taylor published an essay in The British Medical Journal titled “Hospitals of the Future,” in in which he predicted that the hospital of the future would need to adapt its infrastructure to changing medical needs and organize care around patients.
Since then, discussions about the hospital of the future have tended to focus on a wide array of technologies, such as intelligent patient rooms, automated records, telemedicine, robots, and remote surgery.
Don’t get me wrong: these technologies matter. However, when it comes to the hospital of the future, Taylor’s fundamental observation remains critical: The hospital of the future will be defined by how well those solutions are coordinated around the patient.
A FRAGMENTED REALITY
Many hospitals today operate with infrastructure and technology that have been built over decades. The technologies that keep the hospital operational, from clinical records, staffing platforms, and financial systems to supply applications, were purchased at different times to solve different problems, creating fragmented systems.
In my career as an interventional cardiologist and now as a technologist, I have seen hospital leaders wrestle daily with questions that seem routine—even mundane—but are critical to patient care. For example: Which beds are available and adequately staffed? Which patients could be discharged? Which operating rooms are likely to run late? Are the supplies for tomorrow’s procedures in stock? Has a payer approved the next stage of care? Does this patient need the MRI performed during their inpatient stay, or can it be safely scheduled as an outpatient?
In this fragmented reality, the challenge is integrating new technologies with the existing systems in a way that fits into existing workflows and allows hospitals to function as a coordinated institution.
The problem is rarely a lack of data. More often, the answer must be pieced together from multiple disconnected systems. The last decade has seen an explosion of data. Today, healthcare accounts for as much as one-third of the world’s total digital data, and hospitals have made considerable progress in digitizing and exchanging clinical information.
THE DATA VERSUS INSIGHTS IMBALANCE
This abundance of data is not currently translating into a corresponding surge of meaningful insights. Although 71% of hospitals said necessary clinical information was routinely available at the point of care, only 42% said clinicians routinely used it when treating patients, according to federal health IT data.
We’ve become adept at capturing data, but we have work to do translating it into actionable next steps.
Accomplishing this requires uniting multiple systems, including the electronic health record system, imaging queues, and laboratory systems, to name a few. Today, employees often must glean insights from disconnected sources, and even manual processes, to create an operating picture. For example, a nurse manager may believe a unit is adequately staffed for the day based on current census but may be unaware that additional patients are expected in the emergency department and that morning discharges are now occurring in the evening. A surgical scheduler may see an open operating room without knowing whether the required implant will arrive on time.
Coupled with rising patient volumes and more complex care cases, the status quo is untenable.
A HOLISTIC VIEW
Hospitals cannot replace decades of infrastructure overnight. But artificial intelligence can help close this gap—many hospitals are freeing up capacity to care for more people in their community, reducing costs, and delivering more timely care with this technology. The key is scaling its use.
AI is innately strong at spotting patterns in data, making it ideally suited to help hospital leaders derive meaningful, actionable insights from various data sources to address the complex challenges hospitals face in balancing patient demand, staffing, and hospital resources.
However, to fully realize Taylor’s vision, AI will need to help hospitals move from prediction to coordination. When an operating room schedule changes, staffing, supplies, patient preparation, and recovery capacity should adjust with it. When emergency demand rises, the hospital should be able to identify likely admissions, locate staffed beds, identify delayed discharges, and resolve routine obstacles before they become capacity problems.
The goal should be to create an operational layer that understands how a decision in one part of the institution affects another, all while keeping humans in control.
THE IMPORTANCE OF HUMAN OVERSIGHT
The goal is closed-loop execution: Identify a problem, act within the existing workflow, measure the impact, and provide the best alternative path when things change. Hospitals should never wonder whether an intervention reduced waiting, released capacity, or prevented delays.
AI should therefore be measured by more than model accuracy. There’s ideally just one indicator of success: Does this technology help elevate the quality of care efficiently?
That requires measuring how quickly recommendations become actions, how many manual handoffs remain, and whether outcomes improve. With human oversight, traceable data, and audit records, operational intelligence can reduce coordination work and help hospitals respond earlier.
Taylor concluded his 1960 report with this advice: “Put the patient in the center of the picture first, last, and all the time.”
Hospitals have spent decades building systems that record what happened. The hospital of the future will be defined by how well it coordinates what happens next.
Taha Kass-Hout, MD, is the global chief science and technology officer at GE HealthCare.
