Because hospitals were ‘flying blind,’ specialist societies hurried to develop registry tools to help identify patterns in care
When the pandemic struck, clinicians struggled to cope with treating a virus nobody had ever seen before. Medical societies responded by creating or amending data registries to gather and disseminate COVID information as quickly as they could. “In the early stages of the pandemic, we were flying blind,” said Clifford Ko, M.D., director of the division of research and optimal patient care at the American College of Surgeons, which rapidly built a registry tool to share with hospitals. Ko and other society executives recently discussed lessons learned from the scramble to gather COVID data.
The ACS has a long history of developing and maintaining clinical data registries for trauma, cancer, surgical quality improvement for adult and pediatric hospitals, and bariatric surgery.
Ko, who also serves as vice chair and professor of surgery and health services at the University of California, Los Angeles, was speaking during a webinar put on by the Council of Medical Specialty Societies. He said that at the start of the pandemic there were reports of extremely poor surgical outcomes. The mortality rates were from three to five times higher in elective surgery in patients who were known to have COVID or who were not known to have COVID and showed symptoms post-operatively.
“The mortality rates were sky-high and that really scared a lot of folks around the world,” Ko said. Also, there was so much that was unknown. “We were hearing from many of our hospitals they did not know how they were doing with COVID,” he added. They didn’t know what was helpful or not in terms of treatment, and who to test and who not to test, and didn’t have outcomes data.
“We had multiple requests for help with data collection to help them see how they were doing,” Ko said. “To be honest, in the beginning the college’s thinking was that there were going to be a lot of efforts in this area, and that we should stay focused on surgery and rely on others, but hospitals were telling us they needed data now, so we very quickly did two things: We put COVID variables into our current registries and we launched a basic COVID registry for hospitals. We heard these requests in the last week of March, and by second or third week of April we launched this registry.”
The ACS COVID-19 Registry was developed with the input of several experts who are currently treating COVID-19 patients. Included data variables cover demographics, severity predictors, admission information, hospitalization information, therapies used, and discharge information. So far, the registry has data on 10,000 patients from 30 states. The hospital registrars who typically enter data into surgical registries have volunteered to enter this data.
Looking back, Ko said that there are some basic questions specialty societies need to ask about their COVID registries. Are the data good enough or enough for clinical care information and epidemiological information? Is it adequate for research investigations? Is it helping clinicians improve? Are we making gains in the fight against COVID? “I am not sure we are making huge gains, particularly if we agree that the pandemic is ongoing and worsening in many areas,” he said. “We have to ask ourselves if we could be doing something better. First we have to ask ourselves why are we collecting the data and what are the questions we are trying to answer now and what are potential questions we will need to answer in the future?”
ASCO’s COVID Registry
Elizabeth Garrett-Mayer, Ph.D., is director of the Division of Biostatistics and Research Data Governance, Center for Research and Analytics (CENTRA) at the American Society of Clinical Oncology (ASCO). She said that in order to better understand the impact on oncology practices and their patients, ASCO launched a registry of community oncology practices on April 10. “We now have 41 practices in 23 states,” she said, and 175 patients have data submitted. ASCO’s objectives are to analyze distributions of symptoms and severity of COVID-19 among people with cancer; examine the impact of COVID-19 on cancer treatment and outcomes; and document adaptations of cancer care to the pandemic.
“We are very aware that this is dynamic.” Garrett-Mayer stressed. “We recognize that the way we set this up, we would not be able to foresee things coming down the road. We formed a steering group to help us figure out how to move forward and make changes to the data collected.”
“We have tried to be transparent as possible about how we are collecting data,” Garrett-Mayer added. “We are having conversations with the National Cancer Institute and other groups on their registries for cancer patients, and we plan to release reports publicly periodically starting in September. We plan to make the entire data set available to other researchers next year.”
Society of Critical Care Medicine’s VIRUS Registry
Also speaking on the panel was Greg Martin, M.D., professor of pulmonary and critical care at Emory University School of Medicine and president-elect of the Society of Critical Care Medicine (SCCM).
SCCM’s COVID registry is called VIRUS, short for Viral Infection and Respiratory Illness Universal Study. VIRUS is a cross-sectional, observational study and registry of adult and pediatric patients who were admitted to a hospital with COVID-19 confirmed disease or high clinical suspicion. “We have 170 sites from the U.S., but also in the Middle East and Asia, and we are bringing on more sites all the time,” Martin said. There are currently approximately 15,000 patients in the registry. “We are working to facilitate more and more electronic data capture.”
Martin said the primary aims were to create a real time COVID-19 registry of current ICU and hospital care patterns to allow evaluation of safety and observational effectiveness of COVID-19 practices, and to determine practice variations across different hospitals.
Their hypothesis is that the registry will be essential for near-real time observational comparative effectiveness studies to learn effective treatment strategies and/or provide meaningful hypotheses for clinical trials.
Another goal, he said, is to develop a prototype of an EHR-based automated collection of standardized data elements for testing in two community and two academic hospitals equipped with a standard EHR. Before the pandemic struck, Martin added, SCCM had been working on creating a Critical Care Data Dictionary (C2D2), to harmonize critical data definitions for rapid deployment and data collection. The hypothesis is that EHR-based automated collection will be feasible, valid and reliable (compared to manual data entry) with lower workload of data collection.
The speakers mentioned the challenges involved in spinning up registries during a crisis — and having to do much of the work manually — rather than having the electronic data capture tools in place between cycles of emergency. (Two of the registries use the open source REDCap survey tool developed at Vanderbilt University.) They also touched on the potential benefits (and challenges) of integrating data from different societies’ registries.
Regenstrief and Indiana University utilized Indiana’s statewide health information exchange data to automate a previously burdensome process.
For most healthcare organizations, reporting and finding all the necessary information about a new case of a reportable disease can be a burdensome process that could take up to two weeks to fully process, and sometimes it can get lost in the shuffle.
It begins with the laboratory faxing a report to the health department. Then the health department follows up with the clinic to get a more detailed report. If the detailed report does not come back right away, an administrator would have to follow up with the clinic again.
With more cases piling up over the course of a week or two, the health department could be forced to close the case and move on if they did not receive the detailed report from the clinic.
Because of this manual process, Brian Dixon, PhD, director of public health informatics at the Regenstrief Institute, and his team of experts knew they needed to automate disease tracking and surveillance in the state of Indiana.
With sexually transmitted infection (STI) cases on the rise from around 2015 through 2019, a high volume of cases came through the Indiana health department. In return, administrators spent a lot of time calling clinics to get additional information.
“Could we automate this process better?” Dixon posited. “We've done it for labs. Let's try to do it for clinics, but in a way that's going to minimize provider burden.”
Regenstrief had been in this space before, doing a lot of work to automate laboratory-based reporting from labs to the public health department for public health reporting purposes.
“Our public health partners said, ‘The laboratory stuff is working great, but we still have this problem where we're not getting a lot of information from providers.’” Dixon said in an interview with EHRIntelligence.
Because a high volume of STI cases were coming through the laboratory, lab workers noticed providers weren’t sending additional information, such as the symptoms or if the patient was treated.
“That's really where a lot of this was started,” Dixon explained. “We were in communication with the health department. They identified this as a need, particularly for STI's, where they wanted to get additional information to support their activity.”
As a non-profit support community-based organization, Regenstrief works as the connecting organization. In this instance, the organization used its connection to the Indiana Health Information Exchange (IHIE) to link to researchers at Indiana University to utilize IHIE’s data to develop the EHR-implemented clinical decision tool.
To begin the implementation process, Regenstrief utilized Documents for Doctors (Docs for Docs), which has been utilized by IHIE in the past. Docs for Docs is similar to a traditional HIE use case where documents are generated and then exchanged.
“This is the platform that the doctors use to get the laboratory results, the radiology reports, and other consults from specialists, routed back to them at the primary care office,” Dixon explained. “It’s similar to direct messaging where it’s a virtual inbox hosted by IHIE.”
Because this system was already in place, researchers used the platform to integrate IHIE’s patient data into the PDFs, which were previously done by hand and oftentimes not filled out correctly. Once data populates into the PDF, the sheet gets sent to the provider, where the provider can review and sign off on it before sending it to public health.
“We would deliver that right at the same time we were delivering the laboratory results,” Dixon said. “It kind of came to them in parallel through Docs for Docs or the EHR platform. We took advantage of these existing connections and integrations that have already been done by the health information exchange to make the implementation process pretty easy.”
Dixon noted it made life easier for the clinics, which now just had to give IHIE permission to activate the system.
“We could just drop it right into where they were receiving documents from IHIE, and away they went,” he explained. “It's similar to getting an automated reminder saying, ‘This needs to be reported the public health,’ rather than getting that phone call from public health days later.”
The clinical decision tool cannot be optimized for COVID-19 until there is a vaccine or treatment. But, once there is a vaccine, the team will aim to optimize it for COVID-19 when patients are getting outpatient treatment.
On the other hand, COVID-19 has made Regenstrief and their partners think of other technological advances throughout the pandemic.
“COVID-19 has really shifted our focus to looking at how we can apply analytics, machine learning, and advanced techniques, to identify the diseases earlier in the process,” Dixon said.
“But on the analytics front, they’re doing more community-wide dashboards for diseases of focus. The community really came around the health system during COVID-19 to work together to address this major threat to health.”
Since COVID-19 has the nation’s attention, Regenstrief and its partners plan to try to keep the same level of attention and direct it towards other major health challenges in their community. If they can spend the same energy on obesity or diabetes, for example, as they did on COVID-19, then the possibilities are endless, Dixon said.
“So that's what we're thinking about for the future,” Dixon concluded. “How we can develop these technologies that can push information out, not just to the frontline physicians, but also to the health system administrators, to the two community-based organizations and to the public itself, to keep them informed about healthcare in their community.”
SSM Health will now provide its local EMS personnel with real-time access to EHR patient data.
SSM Health, a non-profit health system made up with 11,000 providers in Wisconsin, Missouri, Oklahoma, and Illinois, announced a connection to the ESO Health Data Exchange, which enables interoperability between emergency medical services (EMS) personnel and EHR patient data.
Interoperability and patient data sharing between first responders and hospitals is not always timely. However, this connection allows SSM Health to integrate patient data into its connected EMS documentation systems.
“This technology is the future of healthcare, setting the standard for integration between the pre-hospital and hospital environments,” Kacey Kronenfeld, MD, SSM Health EMS program medical director, said in a statement.
“It will allow us to streamline the way we communicate with our EMS partners, and thus, provide a higher level of exceptional care to our patients.”
The EMS record connects directly to the SSM Health’s EHR, and in real-time, the health system’s clinical outcome data is made available to EMS personnel on a secure portal. This data includes patient demographics, billing, vitals and procedures with results, emergency room physician notes, lab results, and imaging results.
Real-time access to this data enhances care coordination in-and-out of SSM Health facilities and leads to improved patient outcomes.
Efficient and timely access is especially crucial during a health crisis, such as COVID-19, Kronefeld said.
SSM Health is the first health system in Wisconsin to connect to ESO Health Data Exchange, and Fitch-Rona EMS is one of the first EMS agencies to attach its documentation system to the health system.
“First responders often treat patients and never learn about their outcomes,” Patrick Anderson, chief of Fitch-Rona EMS, said in a statement. “We will have access to that health information and our caregivers can receive feedback on their efforts, which will allow us to better serve our community.”
Leaders at SSM Health said they are aiming to add as much as 50 EMS agencies to the ESO Health Data Exchange collaborative.
Health organizations and health information exchanges (HIEs) across the country are aiming to improve interoperability and allow easier EHR data access to EMS workers.
Within the last year, the American Heart Association launched a collaborative initiative to improve EHR data access for EMT personnel.
To make this happen, American Heart Association partnered with an emergency health IT company, RapidSOS, to connect EMS personnel with real-time patient data.
“Heart disease and stroke are the first and fifth leading causes of death in the United States. If emergency medical responders had access to a patient’s medical information when arriving on-scene, this could dramatically change the way in which care is delivered and tailored to the person’s medical needs,” Michael Kurz, MD, chair of the American Heart Association’s Systems of Care Subcommittee, said at the time of the announcement.
“Delays in recognition and treatment of time-sensitive conditions like heart attacks, strokes and cardiac arrest can be the difference between life and death,” Kurz continued. “RapidSOS helps us close the gap between emergency medical response and patients, resulting in better and more efficient emergency care.”
AHA said they want to simplify the communication between individuals who call 911 and responders. Instead of having the individual communicate her personal and clinical information over the phone in a stressful situation, the collaboration enables public safety personnel to access critical health data, such as patient medical history, medications, allergies, and emergency contacts, for a more timely, streamlined response.
Also in 2019, California’s statewide HIE, Manifest MedEx, received a $4.9 million grant from the Emergency Medical Services Authority (EMSA) to promote interoperability with EMS personnel.
This grant connects six California EMS agencies, 13 EMS providers, and 16 hospitals across the state, to enable interoperability.
In 2017, ONC published its SAFR guide to help EMS services link bidirectional health data exchange with HIEs.
A lack of HIE connection to EMS providers is why ONC put an emphasis on a better longitudinal patient health record. This disconnect had a negative impact on patient care between the time a patient is cared for in an emergency situation and the healthcare facility.
Only 22 percent of prescribers located within a high opioid prescribing area had PDMP EHR integration that allowed efficient prescribing oversight.
Hospitals located in areas with high rates of opioid prescribing were less likely to have prescription drug monitoring programs (PDMPs) integrated into the EHR, which limits the impact of PDMPs in these vital areas of the country, according to research published in JAMA Network Open.
A PDMP that is implemented into the EHR can decrease the strenuous process of checking the PDMP, which as a result, can reduce the cognitive workload and clinician burden.
However, EHR-integrated PDMPs are not particularly widespread.
PDMPs are state-run electronic databases that create a record for the patient so that healthcare providers can access the PDMP to understand a patient’s full prescription drug history.
A PDMP aims to enhance prescribing rates, mitigate the risk of opioid abuse or fraud, and target treatment to at-risk patients.
With opioid overdose rates quadrupling between 1999 and 2015, the opioid epidemic has shown no signs of stopping. However, an easy to use PDMP could help clinicians slow down the epidemic.
While study authors noted increased awareness surrounding PDMP usability, there have been limited studies about PDMP adoption and EHR integration.
Additionally, hospitals located in counties with high opioid prescribing rates may benefit more from PDMP EHR integration, due to the providers being in contact with more people who are prescribed opioids.
Researchers evaluated the number of hospitals located in counties with high opioid prescribing rates, compared to other hospitals closer to the mean opioid prescribing rate, and the level of PDMP EHR integration at those hospitals.
Researchers analyzed data from over 3,500 hospitals that took the American Hospital Association Annual Survey and IT Supplement for 2018. Using this data, researchers wanted to answer three PDMP-related questions: the ability to e-prescribe, the ability to conduct a PDMP query within the EHR, and PDMP EHR integration.
Along with those three questions, researchers gauged hospital data, such as size, rurality, region, system membership, and teaching status to identify whether or not each hospital was located in a high opioid prescribing county based on 2017 prescribing rates.
Of the 3512 hospitals, 639 were located within a county with a high opioid prescribing rate.
However, the 2,873 hospitals in other counties were more likely to be able to e-prescribe than the hospitals located within a high opioid prescribing county. Only 57.3 percent of those high prescribing hospitals could e-prescribe, compared to 65.2 percent of other hospitals.
Only 10.8 percent of hospitals located within counties with high prescribing rates had PDMP EHR integration, compared to 15.1 percent of other hospitals.
Finally, while EHR-based PDMP queries were lower across the board, only 22.1 percent of hospitals with high prescribing rates had that capability, compared to 30.3 percent of other hospitals.
While over half of hospitals with both high and normal prescribing rates had the ability to e-prescribe, the majority of hospitals were unable to integrate PDMP data into the EHR, and thus were unable to query from within the EHR. If a PDMP is not implemented directly into the workflow, it is not unusual for a query to take up to seven minutes to process, putting a major strain on clinician workflow.
Hospitals located in counties that needed these capabilities the most, were unable to fully maximize the benefits of an EHR-integrated PDMP.
“Physicians and policy makers are actively considering new strategies to manage the opioid epidemic,” concluded the study authors.
“Focusing efforts on hospitals in areas with high rates of opioid prescribing to build PDMP integration into their EHR may be associated with improved PDMP effectiveness and reduced prescriber administrative burden.”