Health care organizations of all sizes have reacted to the Covid-19 pandemic by rapidly adopting technology — such as telehealth and virtual patient monitoring — that was sparsely used before the spring onset of cases in the U.S. Other advanced IT, such as analytics, machine learning and clinical decision support, has also equipped providers with improved protocols for treating patients with more success.
Information technology has played an important role in helping health care organizations cope with the pandemic, and lessons learned earlier in the year will assist clinicians in delivering treatment and addressing spikes in cases in the coming months.
Organizations such as Northwell Health, a New York-based delivery system that operates 23 hospitals in and around New York City, was faced with heavy Covid-19 caseloads in March and April, now are offering insights into the ways health care IT has aided operations and care delivery, including tools to predict Covid-19 caseloads, machine learning to identify improving patients, and electronic health records systems that enabled Covid-19 treatment standardization and optimization.
In a presentation delivered by Dr. Michael Oppenheim, Northwell’s Chief Medical Information Officer, he explained how Northwell Health launched special documentation and tools for severity-based longitudinal monitoring of both inpatients and outpatients.
Northwell used predictive modeling based on machine learning to identify patients who showed signs of improving to the point where they might be ready for discharge, thus enabling better planning related to bed availability. Conversely, it was also able to use machine learning tools to monitor clinical indicators and identify patients whose health was in decline or likely to decline, enabling teams to conduct goal-of-care conversations with other clinicians and the patient’s family.
Based on its experience with predictive tools, Northwell developed a Survival Calculator that it is letting any interested organization use to assist clinicians in assessing their patients.
Predictive analytics also power another Northwell tool that is publicly available to help predict Covid-19 caseloads based on web search volumes, from patients who self-enter information on symptoms, for example. These kinds of capabilities will help provider organizations manage surges and declines in cases with more precision.
Health care organizations have “been able to apply many functionalities in health care information technology to address Covid,” said Leigh Burchell, vice president for government affairs and public policy at Allscripts, a health information technology company.
“Clinical decision support and predictive analytics have been fundamental in supporting treatment,” she said. “The pandemic use case has stress-tested emerging forms of health IT, and it has been effective in numerous ways.”
Being able to forecast trends in patient loads also helps organizations move clinicians to facilities where they are most needed. For example, Northwell had to work around reduced staff availability during the height of the pandemic — at one point this past spring, several of Northwell’s hospitals saw 20 percent of their registered nurses on leave, either because of their own illness or exposure, or because of family illness concerns.
Information systems also helped Northwell in moving patients among its facilities to balance patient loads so that particular facilities weren’t handling too high of a caseload. Its electronic health records system helped facilitate these transfers by enabling patient information to be available at facilities to which patients were transferred. These transitions were made possible through the use of operational dashboards to facilitate cross-location monitoring for patient load. Northwell also implemented a Covid-specific data mart to enable emergency operations at both the facility and central levels to understand supply and demand for services, equipment and care space.
Clinical decision support capabilities of electronic health records systems enable organizations to standardize treatment protocols for patients with a variety of conditions, such as infection with the Covid-19 virus. Health care organizations tweaked how they treated patients over time through experimentation with approaches — for example, physicians early on found that placing the sickest coronavirus patients on their stomachs helps increase the amount of oxygen that’s getting to their lungs. Research at Northwell and other organizations found that putting ventilated patients on their stomachs for about 16 hours a day was most beneficial for patients.
“Our health IT solutions were critical in terms of disseminating the very dynamic knowledge base around the disease, with frequent updates to treatment protocols and clinical trial participation criteria,” Dr. Oppenheim said. “We were able to get creative on-the-fly around tracking patients as we load-balanced across sites and employed predictive analytics both for clinical parameters and resource needs, which were critical functions to our success.”
Other organizations have highlighted the contributions that technology has played in learning about the coronavirus and its impact on patients. A recent article in the Wall Street Journal reported on a virtual panel about Covid-19 research detailing how data storage, cloud computing and artificial intelligence served as important tools in Covid-19 research, informing investigators about the virus.
For example, multidisciplinary teams versed in medical imaging analysis and machine learning played important roles in sifting through large datasets, said Jayashree Kalpathy-Cramer, scientific director at the Center for Clinical Data Science, which is part of Massachusetts General Hospital and Brigham and Women’s Hospital in Boston. Large amounts of data storage were crucial to enable access to data and for building complex AI models.
Work by public health agencies is continuing to facilitate better data exchange with health care delivery organizations, said Allscripts’ Burchell. Conversations between software developers, providers and state and federal public health entities are ongoing to better communicate information on case volumes and ICU capacity, which is critical information for epidemiological surveillance. And building on that improved communication will be crucial as vaccines are approved for use and then distributed across the country.
“Tracking and reporting (for vaccine distribution) will likely be a smoother process,” in part because information systems will be adapted to help report records for patients who receive the two-phase injections, she predicts. Health care information technology companies like Allscripts are collaboratively working with federal agencies to prepare their systems for tracking Covid vaccines administered to the U.S. population next year.