New metrics and forecasting models are key to understanding and anticipating the challenges of the COVID-19 Pandemic.
The COVID-19 pandemic has fundamentally disrupted U.S. health organizations. Hospitals have faced drug and equipment shortages and have set up new intensive care units overnight. Care plans have evolved out of necessity, and the carefully designed patient flow systems in hospitals have improved.
In an article published today in NEJM Catalyst Innovations in Nursing ServicesBeth Israel Lahey Health (BILH) executives and clinical researchers suggest using complexity science – an area that deals with understanding dynamic, unpredictable systems such as the human brain, the economy, or the climate – to identify strategies on health organizations better able to respond to the ongoing pandemic and anticipating future health care challenges.
“COVID-19 has been a painful reminder that healthcare – both as an industry and as a set of complex organizations – has slowly evolved over time, as have the metrics and models that we use to determine quality, safety and future needs He said lead author Jennifer Stevens, MD, MS, director of the Center for the Science of Health Care at Beth Israel Deaconess Medical Center (BIDMC), part of Beth Israel Lahey Health. “The Principles of Complexity Science provides three strategies that healthcare leaders can use to manage operations during the COVID-19 pandemic: incorporating diverse perspectives into leadership teams, being open to new metrics, and creating forecasting tools that reflect complex healthcare systems.”
Bring different thinkers into one dashboard design
Stevens and colleagues find that in a complex system like healthcare, individuals are often unable to see the “big picture” and recommend adding additional, potentially unexpected, voices to leadership teams. For example, while epidemiologists naturally led healthcare management teams in the context of COVID-19, Stevens and colleagues suggest patient and / or community representatives, doctors from disciplines that may be more tangentially related to the current crisis, or employees from clinics and operations to be included The communities are disproportionately affected by the pandemic. “Voices from these different perspectives can broaden the perspective and enable healthcare leaders to learn more about the complex system and implement strategies that anticipate future needs,” said Stevens.
Identify metrics that triangulate different sides of a complex healthcare system
Identifying a wider range of relevant metrics can also broaden the executive’s view of the complex system. For example, patient demographics was a largely unreported variable at the start of the pandemic – before doctors had a clear picture of how closely race, ethnicity, and other socio-economic factors were linked to the risk of developing and dying from COVID-19.
“A better understanding of the different effects COVID-19 has on color communities could have helped healthcare leaders better anticipate the influx of patients into clinics, as well as the impact on clinical staff and personal protective equipment needs,” said Stevens. “Healthcare executives need to be open to new metrics and pay attention to undervalued variables. Otherwise, we may be so focused on a set of metrics that we overlook the importance of more meaningful data.”
As an example of a metric with a shift in value, Stevens and colleagues described how the importance of the number of patients with COVID-19 in BIDMC’s intensive care unit changed as spring turned into summer 2020. In the spring, patients with COVID-19 presented to the hospital with severe respiratory symptoms and required immediate intensive care. The increasing number of patients admitted with COVID-19 meant the hospital would need additional beds and ventilators in the intensive care unit, personal protective equipment and other health care resources.
However, as soon as the state of Massachusetts and the hospital decided it was safe to resume urgent and elective procedures earlier this summer, all patients were tested for COVID-19 on admission. The universal tests found infected but asymptomatic patients who had been hospitalized for non-COVID-19 reasons.
“All of a sudden, we were admitting asymptomatic patients with COVID-19 who did not need the same resources and care as those who were seriously ill with the novel coronavirus,” said Stevens. “So the importance of the number of COVID-19 patients at BIDMC really shifted in the early months of the pandemic.”
Build forecasting tools that mirror the complex systems of healthcare
The team also suggests that forecasting tools really need to reflect the complex realities of the COVID-19 pandemic. To build a model that does this, the team used machine learning to pull relevant data from each of the 13 hospitals and three business units of the Beth Israel Lahey Health System, rather than depending on published reports from China or Italy. Next, they added publicly available local cell phone data to the model, revealing how much people moved and interacted with other people. Taken together, these data sources – reflecting both the shift in local health policies and the shift in new norms of social behavior over the course of the pandemic – contributed to a model that can provide timely and locally relevant predictions.
“Our model uses the principles of complexity to guide hospital management and provides weekly updates to a group of healthcare executives on how and when a new wave of infections may emerge,” said Stevens. “Models need to reflect the changing health and policy landscape – to take into account the complexity of the pandemic itself – so that any health organization can use it meaningfully.
“Healthcare is facing one of its greatest challenges, partly because our familiar metrics and dashboards, designed to address the problems of a complex system that couldn’t see the ‘big’ coming,” said Dr. Kevin Tabb. President and CEO of Beth Israel Lahey Health. “To adapt to the new realities that COVID-19 has brought to the fore, healthcare leaders need to develop new models that reflect the true complexity we face, bring in new voices, and remain flexible and curious about our metrics . We are still in the middle of this earthquake with many aftershocks ahead of us. ”
Reference: “Earthquakes in Healthcare: Lessons Learned from Complex Adaptive Systems to Develop Covid-19 Responsive Actions and Models: New Metrics and Predictive Models Are Key to Understanding and Anticipating the Challenges of the Covid-19 Pandemic” by Jennifer P. Stevens , MD, MS, Ashley O’Donoghue, Dr. med. Steven Horng, MMSc, Manu Tandon, MBA, MPA and Dr. Kevin Tabb, October 23, 2020, New England Journal of Medicine Catalyst.
DOI: 10.1056 / CAT.20.0505
Other authors include Ashley O’Donoghue, PhD, and Steven Horng, MD, MMSc, from BIDMC, and Manu Tandon, MBA, MPA, from BILH.
Stevens is supported by the Agency for Research and Quality in Healthcare Grant number K08HS024288. The content is solely the responsibility of the authors and does not represent official AHRQ views. Horng announces support from Philips Healthcare. The other authors do not report any conflicts of interest or disclosures relevant to this work.
