COVID-19 hospital network shares key lessons to improve care
Learning from real-time data from 40 Michigan hospitals reduced wide variation in everything from antibiotic use to clot prevention, and yielded findings that could be used worldwide.
More than 120,000 times in the past 15 months, COVID-19 has made Michiganders sick enough to need hospital care, as three huge waves of coronavirus infections swept across the state.
And even as doctors, nurses and therapists worked to save their lives, those patients had no idea that they were getting better care thanks to patients who might have been in the same beds just weeks before.
Nor did they know that the lessons learned from their own hospital care and post-hospital experience might go on to help future COVID patients in Michigan and across the globe.
But that’s what has happened, thanks to a statewide effort called the Mi-COVID19 initiative.
Since April 2020, the 40-hospital network has made it possible to collect, analyze and share data from thousands of COVID patients, and apply the findings almost immediately in hospitals across the state.
The knowledge produced by MiCOVID-19 has already reduced variations in COVID-19 care in Michigan. That contributed to a lower risk of death for hospitalized patients in the state, especially this past winter and spring, according to findings presented at a recent public webinar.
But the potential impact goes further.
Through publications of key findings in medical journals, such as a new paper about the best strategy for preventing blood clots in hospitalized COVID-19 patients, and another about the overuse of “just in case” antibiotics soon after a COVID-19 admission, Michigan’s effort has informed hospital care elsewhere.
Findings from post-hospital surveys of hospitalized patients led to a widely read and shared paper documenting the physical, mental and financial impacts of “long COVID.” That meant better awareness of the needs of people who had survived a hospitalization for COVID-19.
Most important: the study found that 10% of those who survived an ICU stay would go on to die within 60 days of leaving the hospital, signaling an urgent need for more intensive follow-up care.
There are still hundreds of patients in Michigan’s hospitals, far lower than the peak of more than 4,000 for weeks at a time twice in 2021. The Mi-COVID19 team continues to use data from the registry to create new knowledge and share it worldwide.
The team just launched a new COVID-19 inpatient risk-calculation tool based on patterns seen in the registry. Doctors can use to predict the risk of mortality for any hospitalized COVID patient by entering just a few key parameters. They describe the effort to create and validate the tool in a new preprint.
Looking back at the beginning
Scott Flanders, M.D., medical director of Mi-COVID19, recalled the origins of the effort, less than two weeks after the state’s first confirmed COVID-19 patient was admitted to University Hospital at Michigan Medicine, the University of Michigan’s academic medical center.
“It started with the recognition of an opportunity to build on our existing infrastructure to create a registry, leverage our strengths, share best practices, and act in the best interest of all our patients,” he said. “What we learned in the first peak in southeast Michigan’s hospitals was shared statewide before the next two, much more widespread, peaks. We learned, we disseminated what we knew, and we did better. But by no means are we done.”
An early win: early data show that nausea and vomiting were key COVID-19 onset symptoms. Knowing this, when coronavirus testing wasn’t widely available, helped inform care decisions in the first months and beyond.
The registry for COVID-19 patients actually used a framework for studying sepsis patients that was just about to launch in early 2020, under the leadership of critical care specialist Hallie Prescott, M.D., M.Sc., of Michigan Medicine and the VA Ann Arbor Healthcare System.
Blue Cross Blue Shield of Michigan had provided the funding to develop and launch the sepsis registry as part of the Hospital Medicine Safety Consortium. HMS is one of the collaborative quality initiatives, or CQIs, that BCBSM has funded for years as part of an effort to improve care across the state.
The sepsis framework became a springboard for rapid-turnaround COVID-19 data collection. At each of the 40 hospitals, staff who normally do detailed analysis of patient records for several CQIs pivoted to focus on the records of COVID-19 patients.
They fed the data to the coordinating center at U-M, where a team crunched the numbers and shared findings back to hospital teams through emails, webinars and other means. The team also surveyed patients two months after they had left the hospital, yielding critical follow-up data not available through medical records.
Reducing variations in care
“One of our key achievements has been to document, and work to reduce, variability between hospitals, which is something that CQIs have done for other types of care,” Flanders said. “In this case, variability reflected uncertainty, where providers simply didn’t know best practices because the virus was so new. That helped us figure out what we needed to address with answers from the data.”
For instance, at the start of the pandemic, most COVID-19 patients left the participating hospitals with no follow-up primary care appointment booked on their calendar. Now, many more do.
Some hospitals gave “just in case” doses of antibiotics to almost all patients, others to a minority. Few turned out to actually need them, and Mi-COVID19 data showed a single blood test could determine who could benefit most. Education efforts since that discovery have reduced unneeded prescribing and the risk that patients will suffer side effects.
Attention to clot prevention and treatment yielded the paper that just came out last week, led by former U-M hospitalist Valerie Vaughn, M.D., M.Sc., and Michigan Medicine cardiologist Geoffrey Barnes, M.D., M.Sc. It shows the lower risk of in-hospital and post-hospital death among patients who received consistent anticoagulation medications focused on preventing a dangerous type of clot called venous thromboembolism, or VTE.
Based on these data, Michigan hospitals have changed their care, under the leadership of the Michigan Anticoagulation Quality Improvement Initiative, another CQI.
“Back in spring of 2020, we did a lot of stuff that was wrong – it was our best guess at the time but it was wrong, because we hadn’t done the science yet,” said Theodore “Jack” Iwashyna, M.D., Ph.D. at the recent webinar. “Our inpatient science progressed at an unprecedented rate because we would find ways to ask questions and answer them.”
Iwashyna is a critical care specialist at Michigan Medicine and the VAAAHS.
“I’ve never seen science move as quickly in a collaborative fashion as I have in the last year and a half,” said Robert Hyzy, M.D., the medical director of the Michigan Medicine’s Critical Care Medical Unit and co-leader of the Mi-COVID19 webinar series. “This effort has borne fruit.”
Moving forward
Now, with the third and hopefully final wave of COVID-19 cases almost over in Michigan, the Mi-COVID19 registry has stopped gathering new data. But the HMS sepsis effort is now going strong, and about half of the patients it is tracking have sepsis related to COVID-19 – a sign of just how serious the disease continues to be, especially for unvaccinated people.
From the beginning, the Mi-COVID19 effort has shone light on the disparities laid bare by the pandemic, with help from funding from the Michigan Public Health Institute. That work will continue.
Besides documenting the disproportionate toll on the Black community in Detroit, the registry helped show the relationship between the social vulnerability in a particular area of Michigan, and the odds that patients would need a ventilator or intensive care.
Another key lesson was the sheer lack of data in hospitals’ electronic health records about patients’ race, ethnicity, living circumstances, or other factors known as social determinants of health. Most hospitals didn’t even have a place in their digital record-keeping systems to note that information.
Changing this for future will take partnership between clinicians, hospital leaders, information technology professionals and EHR companies, the registry team said.
Iwashyna said he hopes that the lessons learned about health disparities, the importance of transferring patients between hospitals appropriately, and the power of reducing regulatory burdens such as extending the licenses of nurses and respiratory therapists in an emergency, will be carried forward.
Also important is making sure COVID-19 survivors get appropriate follow up care, just as health systems have done for survivors of heart attacks or strokes. Michigan Medicine’s new post-COVID clinics for adults and children, and other clinics launched by hospitals across the state, aim to do that.
Mi-COVID19 team members are also focusing on the long-term mental health effects of COVID-19 on patients and health providers alike. They’re also looking at patterns of how many patients were sent home from the hospital with prescriptions for drugs that can affect the brain and behavior in risky ways, such as opioids and benzodiazepines.
“I think this rate of change in clinical care is unprecedented in my time as a practicing physician,” said Flanders. “Through an immense amount of work by many people, we significantly improved outcomes for patients in Michigan and provided evidence that could affect care far beyond our borders.”
Flanders, Prescott, Iwashyna, Barnes and lead author of the post-hospital paper Vineet Chopra, M.D., M.Sc., are members of the U-M Institute for Healthcare Policy and Innovation. Vaughn, now at the University of Utah, is a former IHPI member. Prescott, Iwashyna and Chopra are members of the VA Center for Clinical Management Research.