In the fall of 2020, while American universities operated fully online, used a hybrid approach or offered in-person classes, the University of Miami was able to offer a choice of in-person or hybrid learning, safely welcoming students back to campus.
And at the Miller School of Medicine, despite facing international supply chain gaps and staffing shortages, scientists/pathologists and researchers met the litmus test for opening: keeping students safe by providing high volume, fast-turnaround COVID-19 testing.
Yi Zhou, M.D., Ph.D., an associate professor and director of flow cytometry, was tasked with managing the student testing process. Thanks to early decisions made by him and other UM laboratory leaders to automate and create new testing processes, by late fall Miami was able to test every student weekly.
Chief among these colleagues was Sion L. Williams, Ph.D., a research assistant professor, director of the Onogenomics Core Facility and co-director of the CFAR Laboratory Core. Together, their teams accomplished unprecedented testing volumes in a student-friendly manner at a low cost and with staff safety secured.
“As with the many other essential departments at the UHealth, the efforts of the faculty and staff of the Department of Pathology & Laboratory Medicine in the response to COVID have been amazing,” said Merce Jorda, M.D., Ph.D., chair of the Department of Pathology & Laboratory Medicine. “It is an honor to work with such a dedicated group to support our mission to the hospital, clinics, and school. The incredible work performed by Dr. Zhou, Dr. Williams, and other colleagues and staff will continue to be the key to our success.”
A New Order of Magnitude
The call to transform laboratories designed to process a handful of samples a week into ones capable of handling up to 3,000 a week required a sharp departure from business as usual. He said the most challenging initial hurdle was finding qualified staff to handle this volume.
“Everyone was seeking to hire the same people," he said. "There weren’t enough trained molecular technicians to go around.”
Supply chains were threatened by skyrocketing global demand for all manner of laboratory products, with plastics -- needed for test tubes and automated equipment parts -- particularly scarce. Another variable to be reckoned with was the cost of this transformation enterprise.
Dr. Zhou and his colleagues went into intense planning mode to tackle these challenges. Finding a practicable way to automate the process was the first idea to gain steam, since it would solve the turnaround time, volume and staffing problems.
“The first thing we did at UHealth Tower was set up RNA tests using PCR assays, which used existing diagnostic hardware," he said. "This platform had relatively low throughout, but supplied the needs of the hospital."
Then, to ramp up volume to cover student testing, Dr. Zhou and Dr. Williams brought in a more open platform for the PCRs.
“We started with the CDC assays but then switched to the PerkinElmer platform, which enabled robotic handling of nasal swab samples end to end – from handling of the collection tubes, to RNA extraction, to setting up the reactions.”
Obtaining Key Equipment
Obtaining the PerkinElmer equipment was a major victory since states and large conglomerates were vying for the same resource.
“Fortunately, UM has a good relationship with this company,” Dr. Williams said. “The hospital department of human genetics helps PerkinElmer develop the DNA sequencing protocols for those robots, so we have a longstanding relationship with them.”
Then Dr. Zhou made a procedural change that had another crucial impact on the lab’s success. Instead of putting the live virus sample through a chemical robotic process, he designed a system for heating and killing the virus before extracting and analyzing the RNA. The homemade assays he made were about a third of of the cost of purchased assays, and protected the staff from infection. Further, by using a molecular buffer that only required one tube transfer to the robot, he was able to cut the manual RNA extraction step.
Dr. Zhou says these changes conferred multiple benefits.
“In this kind of very high volume screening test, we needed to minimize errors" he said. "Each time you transfer the sample, there is a risk that it could be mislabeled or misplaced. By streamlining and simplifying the tasks involved, this automation increased speed and decreased the requirement for people with an advanced assay analysis skill set."
The collective impact of these changes was to speed up the full cycle testing time from six hours to 19 minutes.
“That enabled us to collect a specimen that day and literally report a result at midnight,” Dr. Zhou said.
Roadmap to 275,000 tests
The university set up testing sites around campus for collecting nasal swab samples, and a team of two technicians per shift, working 24/7, five to six days a week, processed them. If anyone tested positive, that student or staff member was isolated to prevent an outbreak on the campus.
Testing capacity topped out at 3,000 tests a day, but generally the average was about 1,500 to 2,000 a day.
“When we started in April, we had a very small footprint,” Dr. Zhou said. “We gradually increased it to about 600 tests a day by August, which met the needs of the hospital. The university, however, needed over 10,000 tests a week, and we met that goal by December.”
Dr. Williams and Dr. Zhou had successfully reached across the system, forming a collaboration between the hospital lab, the pathology department and the Sylvester Comprehensive Cancer Center's genomic core, which had experience in large-scale testing. By June 1, more than 275,000 tests had been performed.
While Dr. Zhou says cost-savings was tertiary to safety and volume goals, it was a concern at every step.
“Simplification of the task naturally drove savings,” he said. “Going from originally collecting saliva to just collecting a nasal swab and then using a buffer suitable for molecular testing alone saved substantial costs and made the operation very convenient. We didn’t have to rely on a supplier to provide a special assay/device. Also skipping steps between assay and RNA extraction meant significant savings right there.”
Dr. Zhou predicts that for the fall 2021 semester, the prominence of the more virulent Delta variant may require more frequent testing, maybe as often as twice a week.
“The lab is preparing for that, but it will depend on how many of the students on campus are vaccinated,” he said.
For now, the plan is for students to continue to test once a week and update their status on their MyUMHealth chart. They will be required to mask indoors and encouraged to mask outdoors in large group settings.
Watching the Variants
Dr. Williams, Dr. Zhou and their colleagues are closely following new variants to prepare for testing changes as the school year progresses. To detect these in a cost-effective manner, the pathologists do a quick QPCR detection on samples to see which ones have the most common variant. At one time it was the UK, and currently it is the Delta. Then they sequence those that don’t match up with Delta to see what they are.
“To check the variants, we use a device that does 100 samples at once, and then we do the RNA extractions on the small percentage of samples that are not the Delta variant,” Dr. Williams said. “That nice clean sample of RNA comes over to my lab, and then Dr. Anthony Griswold, a Research Assistant Professor of Genomics and Bioinformatics, does the DNA sequencing to identify what strain and lineage it is.”
The Delta variant now comprises 90% of active cases, but new concerns include the emergence of the Latin American variant, and in recent weeks, the Lambda variant, which originated in Peru and Argentina. The latest variant to arise is the B1621, emerging in Florida and New York.
“We are trying to get ahead of the virus by looking in that unknown 10% group to see if there anything expanding anything accelerating and growing,” Dr. Williams said. “At the moment, there's no actionable data that you get from sequencing – all patients are treated the same. That might need to change in the future, which is why we are closely tracking them.”
He says a worst-case scenario is high rates of Delta continuing and then Lambda accelerating before Delta drops considerably.
“Then we’ve got a sort of double peak, which will be truly challenging,” he said.