For critically ill patients on breathing machines, a simple step drastically improves their survival chances by almost 10% — from 60% to 70%. It involves programming the machine to deliver enough life-sustaining breaths, but not so much that it damages their lungs by overinflating them. Given that this intervention could prevent more suffering than many wonder drugs, one would expect that there would be zero market for a breathing machine that didn’t make lung-preventive ventilation as easy as possible. But in health care, few things work as expected. Fewer than half of patients, and in some hospitals fewer than 20%, receive this life-saving intervention.
One big reason why is that hospitals purchase technologies without requiring that they communicate with each other. The optimal air flow is based on a straightforward calculation using the height of the patient. Height data, however, resides in the electronic medical record, which typically does not communicate with the ventilator. As a result, physicians must retrieve this information from the medical record, perform the calculation (sometimes on paper), and enter the order. A respiratory therapist then takes the order and types it into the ventilator, often relying on memory.
If the ventilator and medical record communicated with each another, calculating the ideal air flow would be automated and clinicians would only need to verify the correct settings. Instead, they waste time on nonproductive work, filling the gap between these two systems. Because similar gaps exist between dozens of other hospital technologies, and clinicians are asked to perform hundreds of steps each day to deliver evidence-based care, unnecessary errors occur, and providers’ productivity has fallen, even while spending on technology has ballooned.
Health care’s safety and quality challenges are exacerbated by its procurement problem. For years, hospitals have invested in sophisticated devices and IT systems that, on their own, can be awe-inspiring. Yet these technologies rarely share data, let alone leverage it to support better clinical care.
How did we get here? First, the number of devices that work well with others is small. Manufacturers have been slow to embrace interoperability, which would allow health care technologies to share data with one another. In recent years, there has been movement to change that. More companies have pledged to open their data, giving innovators everywhere the chance to mine that data and use it to drive better care. But we are far from where we should be.
Second, despite significant work, health care lacks widespread adoption of interoperability standards that govern formats and elements of data shared between different systems. Without such standards, data cannot be shared and understood among devices. An accelerated effort is needed to create mature standards and expand their adoption by manufacturers. At Johns Hopkins, we are leading development of a report for the National Academy of Medicine that will identify the barriers to widespread interoperability and suggest opportunities to overcome them, such as policies, requirements, standards, and purchase specifications.
Part of the solution must involve hospitals. If they truly want technologies that save lives and boost productivity, they will need to exert their considerable pressure as purchasers, requiring that manufacturers embrace openness and interoperability, and only purchasing devices that support this. Too often, hospitals treat equipment and IT procurement in a siloed way, focusing on price without looking at how those devices will work as part of a larger system. For example, many new hospital beds come with a sophisticated array of sensors that can track such information as whether a patient is at risk of developing a bedsore, based on data about how often they move in bed. Such sensors may be 30% of a bed’s costs. Yet at one of our hospitals, that data is unusable — it’s in a format that our system cannot read.
It’s a similar situation for much of the data that is fed from wireless monitors of patients’ heart rate, blood oxygen levels, blood pressure, and breathing rate: This data doesn’t link to the medical record.
Health care is woefully underengineered. Too often, clinicians mold their work processes around the demands of multiple devices and health IT systems, yet those technologies don’t work together to serve their needs and the needs of patients. Using systems engineering, we can integrate technologies and build hospitals and clinics that ensure consistently safe, high-quality, and efficient care.
The vision of an integrated hospital unit that is much safer and more productive will not be possible without widespread availability of products that share data openly and freely. Just as the U.S. Navy demands that its submarines and ships have interoperable technologies, this change can be driven by those who purchase these technologies. Health care leaders that purchase technologies need to do the same. When health systems insist on interoperable technologies, the market will respond.
It’s unrealistic to think that each hospital should go it alone, exerting its purchasing power to move the marketplace. However, hospitals could work together, writing specifications and functional requirements for the products that they will purchase and refusing to do business with manufacturers that don’t comply. Group purchasing organizations, which help procure products and devices for thousands of hospitals under their umbrella, might also fill that role.
And we should take it one step further: Rather than looking to assemble hospital rooms one product at a time, hospitals should be able to purchase modules, sets of interoperable products that work together to support an aspect of care. This model makes sense, as few if any hospitals have the resources to design and manage all the connections between technologies, or to optimize how the data is used and displayed to support top-quality care. Ultimately, when a hospital is built or renovated, it would have the option to buy modular patient rooms, clinical units or floors — a “hospital in a box,” built to its specifications.
We don’t expect airlines to build their own planes. They buy them from experienced system integrators such as Boeing or Airbus. There’s no reason that hospitals shouldn’t have a similar model. The question is whether health care leaders will have the resolve to require it. The survival of their patients, the financial survival of their organizations, and our ability to reduce health care costs may depend on it.
Peter Pronovost, M.D., Ph.D. is Director of the Armstrong Institute for Patient Safety and Quality and Senior Vice President for Patient Safety and Quality at Johns Hopkins Medicine.
Sezin Palmer is the Mission Area Executive for National Health at the Johns Hopkins University Applied Physics Lab.
Alan Ravitz, Ph.D. is Chief Engineer, National Health Mission Area at the Johns Hopkins University Applied Physics Lab.