NEW YORK – As the COVID-19 pandemic marks its one-year anniversary, laboratory industry stakeholders have significantly changed the way they perform testing and manage care for patients.
Two of the key shifts throughout the industry are the transition to remote and at-home testing and the accelerated move toward digitization of data and machine learning, a panel of lab industry members said on Thursday during a session on trends and topics at the American Clinical Laboratory Association's annual meeting, held virtually this year
Due to stay-at-home orders during the pandemic, as well as fewer people visiting doctors' offices for routine care, testing options for customers to use at home has seen a significant uptake in the past year, and laboratories have adjusted their offerings to take advantage.
Dorothy Adcock, the CMO and senior VP at Laboratory Corporation of America, said the company has had direct-to-consumer offerings since before the pandemic with its Pixel at-home testing kits and its program called LabCorp At Home, where physicians can order a test and have it sent to a patient's home. The patient then self-collects the sample and sends it back to the lab for testing. Although it has been used for routine SARS-CoV-2 testing, Adcock said it is also ideal for pre-operative testing for patients who didn't want to go to a doctor's office but needed to be cleared before a procedure.
LabCorp is continuing to expand its at-home testing capabilities, Adcock said, emphasizing that at-home testing will "stay with us for a while" and is going to be a "paradigm shift."
Mohamed Salama, CMO at Mayo Clinic Laboratories, echoed Adcock's sentiments, saying the impacts from the pandemic are "going to affect every aspect of our lab." Salama also said he's expecting a lot more home collection kits and different collection modalities.
One of those modalities Mayo Clinic is using to facilitate remote testing is validating its tests for use with dried blood spot collection or self-capillary collection devices to make blood-based testing easier without a patient having to see a physician or phlebotomist in person. Mayo Clinic has used card-based dried blood spot testing to do antibody testing for its 60,000 internal employees.
At-home blood collection has sparked Mayo Clinic's interest in areas like transplants and chronic lymphocytic leukemia (CLL) disease, where physicians can order tests, patients can collect blood samples at home, and then follow-up care can be provided through telehealth. The "primary drivers for these dramatic changes are primarily going to be economic conditions, as well as customer demands and marketing," he said.
Biodesix CFO Robin Harper Cowie said her company has also seen an increase in mobile phlebotomy, which it had used to some degree prior to the pandemic but has now become a critical feature to get testing samples collected without patients physically being in the presence of a doctor or a nurse.
Tied into at-home testing and remote care is the acceleration of telemedicine, the stakeholders emphasized. Telemedicine has been around for years, but the pandemic has clarified just how useful it can be. Although telehealth is "not necessarily new," said Laura Housman, the head of access, outcomes, and population health at Exact Sciences, its use has more than doubled in the past year. Exact Sciences happened to launch its online portal for clinicians to order its Cologuard test in February 2020, and Housman said that it was particularly helpful as a way for patients to gain access to screening without being in person.
Salama noted that the Mayo Clinic had ramped up its telehealth platform since the pandemic and said the industry as a whole has adopted digital health tools that would normally take decades to reach widespread acceptance. The pandemic is "forever reshaping the traditional care delivery models" and shifting both care and testing to be more decentralized, he said.
Cowie also shared Salama's thoughts, saying that she has started to see more telehealth or care via Zoom as an option, particularly in rural areas. The pandemic "accelerated this shift to the norm" as hospitals and doctors rely more on telehealth, something that might have taken 10 years to get to this point without the pandemic.
However, there are some barriers to full telehealth adoption, Cowie said. Basic necessities, like nursing staff and access to electronic medical records, aren't always in place at certain facilities. "Technology is and will continue to be an issue," she continued, especially as technology inequities across offices and hospitals become clearer.
Salama also noted one example of laboratory-specific telemedicine advancements: telepathology. The Mayo Clinic moved quickly during the pandemic to using telepathology for certain cases, such as signing off on autopsies, and once the system was set up it can be deployed in other areas, he said. Digital pathology, along with the digitization of data, can enable artificial intelligence that could revolutionize the practice of image-based pathology, he added.
Adcock also mentioned the "great advance" for digital pathology, saying that it is more objective and that laboratories that have implemented it find more efficiency and can do things more rapidly than before. Digital pathology firms have seen significant interest from investors as well, with companies like Paige and Proscia raising millions in funding rounds. Paige, in particular, has invested in artificial intelligence-based digital pathology products, such as a prostate cancer detection system.
Digitization of data to provide healthcare interventions was another major focus of the panel, with an emphasis on using machine learning and artificial intelligence. Modeling has been a key component of the SARS-CoV-2 pandemic, as national and state agencies use models to monitor and forecast the spread of the virus, Salama said. Using models can also help in implementing certain restrictions, such as business and school closures, by knowing where outbreaks are occurring and where further spread could be discovered.
Both Salama and Cowie said they have used machine learning to create models to help predict a patient's response to COVID-19. Salama said one of the COVID-19 innovation groups at the clinic developed a machine learning model to provide mortality predictions. Primarily using routine lab measures, such as complete blood cell counts, and clinical covariants, the model can provide a prediction within 72 hours of a PCR test result, he said. Once the prediction is available, clinicians can decide if the patient should be admitted to the hospital, monitored, or transferred to another facility to provide more acute care.
Cowie said Biodesix has been using machine learning and artificial intelligence as part of its new test discovery platform for a long time and as a result was able to leverage its partnerships to expand into other areas of research. One of those areas was developing a machine learning algorithm to assess which COVID-19 patients are likely to need intervention. She also said the company is exploring what other clinical questions the platform can address with input from electronic medical records rather than just wet labs.
One challenge of this quest to digitize data and optimize machine learning is actually getting enough data to train the machine in the first place, Salama said. It's "all about the amount of data" a laboratory can feed to the machine and getting that data can be difficult. Although the Mayo Clinic reached out to other labs and hospitals to try to get more data for its platform, doing so proved difficult due to different regulations and concerns around confidentiality, Salama said.
Beyond the pandemic, laboratory data can and has been used to drive population health measures and help with intervention efforts for chronic diseases, Salama said. It's a "matter of us prioritizing it." One way to do that is by combining lab data with results from electrocardiograms and applying artificial intelligence to see which people in the population are likely going to develop cardiovascular disease or experience heart attacks, he said. Data can also be used to determine certain conditions without a patient having to undergo a biopsy or invasive procedure, he continued.
An essential component that laboratory leaders must focus on, however, is making sure hospitals and other facilities know how to use and protect the data – such as maintaining it correctly and putting it where it needs to go, Salama said. For example, a lot of flow cytometry data doesn't always make it into the laboratory information system. It sounds simple, but if the data isn't in the system, it can't be utilized.
Digitizing data can also identify gaps in care for underserved populations. Housman said Exact Sciences has been developing technologies for data sharing that have kicked into overdrive as a result of the pandemic. It's been particularly helpful for payors to identify care gaps, she added. Once those gaps are discovered, they can be addressed through campaigns, protocols, and techniques such as marketing to reach underserved populations and ensure they're receiving needed care. Housman added that Exact Sciences has updated its modeling to include race and ethnicity and take note of communities that may need intervention.
Adcock said Labcorp has also been using what it calls "insight analytics" to look at gaps in care, which has allowed the laboratory to play a bigger part in patient care. Labcorp utilizes its data for community antibiograms – reports about how the susceptibility of different pathogens to different antibiotics – which are generally made available to hospitals. But the program can also determine that information by zip code so the company can notify pharmacists and providers in specific zip codes to let them know if there is an issue with antibiotic resistance near them.
Because of data and analytics, "we can provide real community services," she said.