What sepsis management teaches us about healthcare tech

Technology might just empower healthcare providers to think outside of the hospital setting and go back earlier in medical histories, perhaps even visualizing timelines through clinical data.
6 January 2022

A dummy is used to familiarize members of the medical staff with the new Medisana Temi robot. (Photo by GIL COHEN-MAGEN / AFP)

The US healthcare system is widely seen as dysfunctional — due to high costs that don’t always reflect high quality, the now antiquated bundling of insurance with conventional employment, payment issues that conjure up existential dread and stifle innovation, the social determinants of health and resulting inequities, and multiple other aspects that would require a stack of books to adequately explain!

So, is technology a major part of the solution, or a distraction from systemic issues?

Addressing these fundamental questions, Dr. Robert H. Shmerling wrote that the US healthcare system sometimes loses its focus on preventive care by emphasizing technology along with disease and specialty care.

Incentives can factor into this lopsided approach. He explained, “Doctors practicing in specialties where technology abounds (think anesthesiology, cardiology, or surgery) typically have far higher incomes than those in primary care.”

Previously in TechHQ, I observed that the high capital requirements of healthcare tech, especially involving new AI, can act as a hindrance to market development, but digital transformation is nevertheless improving operational visibility and process compliance. This can mitigate critical health conditions such as sepsis, which originates outside of the hospital in nearly 87% of U.S. cases but requires urgent treatment regardless of its origins. Using IBM technology, AHMC Healthcare brought sepsis management compliance up from 10 percent to 100 percent in some of its facilities.

A recent JAMA viewpoint by Angus and Bindman (2021) elaborated on the issue of how sepsis, which is an extreme, life-threatening response in a bodily systemic process, can itself be treated through the optimization of healthcare operational processes. This includes technology.

A prompt and accurate diagnosis for this high variability condition, which often has nonspecific symptoms, can significantly improve outcomes but is inherently challenging. Each patient’s risk status must be accurately assessed — with technology being the most accurate, reliable barometer — to avoid potentially fatal neglect or unnecessary treatment that could lead to antimicrobial resistance.

This type of digital transformation can positively redirect attention and typically results in some kind of cultural change, as well.

Technology might just empower US healthcare providers to think outside of the hospital setting and go back earlier in medical histories, perhaps even visualizing timelines through clinical data.

Mahmoud el-Komy (R), a mechatronics engineer, tests a self-funded prototype for the CIRA-03 remote-controlled robot to assist physicians in running tests on suspected COVID-19 patients. (Photo by Khaled DESOUKI / AFP)

Technology innovates healthcare

While some tech innovators are accustomed to generating or receiving engagement/retention-oriented mobile push notifications that can be rightly dismissed as nuisances, these are the types of alerts that really matter. The authors identified another relevant technological approach in the form of predictive analytics (which also, incidentally, has relevance to more widely familiar but ultimately less important technological areas such as marketing).

As they explained, machine learning algorithms can be used “to comb across electronic health record data (or higher-fidelity information such as continuous physiologic monitoring data streams) in the hope of detecting unrecognized patterns of clinical data that are predictive of sepsis yet occur before the patient meets established clinical criteria.”

Technology might also empower healthcare providers to think outside of the hospital setting and go back earlier in medical histories, perhaps even visualizing timelines through clinical data.

Angus and Bindman wrote: “With the plethora of newer technologies that facilitate less-invasive monitoring and the engagement of patients in care processes, the time seems ripe to study whether it is possible to diagnose and intervene on at-risk patients even before the signs of sepsis are fully manifest.”

Even setting aside technological implementations, they noted that the strategies of sepsis diagnostics haven’t been adequately evaluated in large-scale randomized clinical trials. Ultimately, the complexity of sepsis diagnostics and management signals a lesson that is applicable throughout healthcare.

New technologies can be powerful but they can’t necessarily preempt, or rectify, systemic industry problems or a lack of scientific knowledge. At least, not without massive, sustained effort and perhaps the infusion of capital from new players that are able and willing to design from a blank page.