Healthcare Musings | Posted By Matthew Sappern
One might argue “poor comparison, the access to and accuracy of this data is not on par to that of the US.” However, multiple studies suggest that in the US, at least 50% of adverse events in childbirth are preventable with the vast majority of those rooting back to a clinician’s “failure to recognize and delayed treatment of clinical warning signs.” If that sounds like legalese, it is. That term is familiar to any medical malpractice attorney. These data are certainly contributing factors to The Joint Commission’s recently introduced performance measure PC-06, which tracks unexpected complications in childbirth.
The rate of avoidable complications is not an indictment of clinicians – far from it. The causes of adverse events in childbirth are many and impacted by a broad array of social determinants. As well, the last decade has seen climbing rates of extreme obesity, diabetes and hypertension in laboring moms. Moreover, in the face of rising patient complexity, hospitals must deal with decreasing availability of experienced nurses. Electronic medical records consume more nursing attention than ever before, leeching patient-facing time. Nursing shortages complicate an already volatile and difficult to forecast staffing model, especially as the most experienced baby-boomer nurses retire. In response, technology companies have done virtually nothing by way of clinical innovation, even as other sectors such as financial services, retail, and entertainment have made technologies such as AI part of the common vernacular.
I know from firsthand experience that some of you think more education alone will solve this “failure to recognize” problem, but in and of itself, it won’t. Education has been ongoing for years and safety programs have ramped up in hospitals throughout the U.S., and still we see unacceptable rates of adverse events. Studies have shown that lack of knowledge is not a common problem in preventable incidents; rather poor recognition of the degree of illness, lack of communication and failure to seek help are the biggest contributors. Denial, complacency, wishful thinking are real issues. These factors all contribute to delayed intervention. Legacy approaches and legacy technology will continue to deliver legacy results.
Addressing the many social determinants that impact outcomes will take many years. In the short term, let’s turn our attention to defined, manageable and intuitive approaches to start chipping away at the number of adverse events in childbirth. As stated above, failure to recognize the significance of clinical warning signs is pervasive in maternal mortality. The same kinds of human lapses are present in severe maternal morbidity and birth-related brain injury cases. Frankly, humans are poorly suited to ongoing, split second assessments of large and dynamic data sets. Counting on ANY number of humans to look at the same clinical data, reach the same conclusions and take the same actions is naïve at best. Fortunately for moms and babies, there are technologies that are well-suited to consistent, unbiased assessments. These automated early warning software systems for obstetrics are designed to enhance clinical efficiency, timely intervention and standardization of care, and the underlying technology has played a critical part in reducing adverse events at a number of US health systems. At a glance or by automated notification, clinicians can quickly see when patient conditions are worsening. The AI-driven technology is never tired, biased or unavailable when helping a colleague across the hall. It continuously analyzes clinical data, and communicates abnormalities in a timely and quantitative fashion. Rather than attempting to discover potential issues amidst a sea of data, clinicians can spend their time evaluating the issue that has been discovered by the AI technology. Using their experience in combination with training, clinicians can then determine the appropriate path forward.
Establishing an overwatch
This sort of AI technology offers the additional benefit of being able to support a centralized hub that providing “overwatch” over multiple beds, departments, or hospitals. For example, a single clinician in front of a screen can monitor all the labors across an entire group of affiliated hospitals and their labor and delivery units, and be alerted only when specified parameters are breached. Rather than increasing nursing staff or hiring patient safety clinicians at every hospital a large health system can monitor cases efficiently and economically from one central location. This overwatch concept can be extended to remote or unaffiliated hospitals as well, to offer centralized monitoring and advisory services. Clinical leverage at a time when clinical staffing is quite expensive.
Teaching tool
Another value AI Labor & Delivery (L&D) technology offers is as a teaching tool for clinicians – especially those new to the hospital, or new to the L&D floor. Once an alert has been generated, they can compare their own evaluations to the system’s recommendations, helping them build confidence in their own decision-making abilities as they learn the hospital’s protocols and processes. Even experienced clinicians can sharpen their skills, especially around less common or more confusing occurrences.
Finally, AI technology can be a real asset for nurses, residents or other clinicians when speaking with attending physicians who are not immediately present about issues they see. Being able to say, “Both I and the system see the following pattern” and communicating objective data leads to a more clinically relevant discussion.
Taking down the crisis
Addressing this crisis requires short, medium and long term strategies. We need to take advantage of existing modern solutions that incorporate today’s technology now, even while we figure out other alternatives and solutions. This tide will take a long time to turn, let’s get started.
About the Author
Matthew Sappern is CEO of PeriGen, an innovator of perinatal early warning systems. Previously, he served as senior vice president of client sales at the EMR company Allscripts.
