Healthcare Data Sharing Connects the Dots for COVID-19 and Beyond
Original Article by Jessica Kent, www.healthitanalytics.com
July 31, 2020
The COVID-19 pandemic has pushed healthcare to engage in data sharing and research partnerships, facilitating new strategies that many hope will endure after the crisis subsides.
When one considers the healthcare industry, it’s easy to picture the entire ecosystem existing as one massive, cohesive machine – a system seamlessly working to achieve the goals of quality care, lower costs, and improved outcomes.
If one were to take a closer look, however, they would see that the industry is a little more complicated than that.
Healthcare often operates as an entity with many different moving parts. And while those parts do share a common, overall aim, they typically function in isolation.
Research institutions, provider organizations, pharmaceutical companies, and other healthcare bodies usually conduct their efforts on disparate digital systems, making it difficult to access relevant health data and draw actionable conclusions. Add regulatory and privacy constraints to the mix, and healthcare data sharing becomes an incredibly challenging task.
In the context of the COVID-19 pandemic, these flaws have only become more evident.
“Because we don’t have a centralized healthcare system in this country, we also don’t have centralized healthcare data,” Melissa Haendel, PhD, associate professor of medical informatics and clinical epidemiology in the Oregon Health and Science University School of Medicine, told HealthITAnalytics.
“A lot of the efforts to analyze data across the nation rely on a variety of reporting structures that summarize the data in different ways and for different purposes. In times of emergency such as this, it becomes even more apparent that this lack of standardization and process for sharing data are huge barriers to understanding the full epidemiological picture and to finding potential therapies.”
With the entire country working to track, control, and treat the virus – and limited means to exchange helpful information – it’s clear that the industry needs data sharing solutions to improve care now and in the future.
As the pandemic continues on, leaders from all areas of healthcare are partnering to better understand COVID-19 and further facilitate data sharing going forward.
Setting the stage for increased data sharing, access
When coronavirus first hit the US, researchers and provider organizations moved quickly to discover everything they possibly could about the disease: which populations it would impact the most, what measures were needed to reduce the spread, and how the virus would present itself in different individuals.
However, the novel nature of the virus, coupled with its rapid spread, presented healthcare leaders with an overwhelming, unprecedented task.
“COVID-19 is a new disease that reached pandemic proportions in a very, very short period of time. Because of this, we needed data integration in a way that we never have before,” said Jomol Mathew, PhD, chief of biomedical informatics at the University of Wisconsin School of Medicine and Public Health.
“We needed to bring in data from different hospitals and different geographical regions, to get an estimate of the prevalence itself, and also to learn about evolving symptoms and the course of the disease.”
The pandemic highlighted the significant need for accurate, real-time data access, both in the near-term and the long-term, Mathew explained.
“Data access and sharing is required to understand the effect of comorbidities, optimization of treatment in various comorbidity settings, and even for public health surveillance. Data exchange is necessary for meaningful research because if we want to collaborate with entities outside of a particular institution, we can easily leverage their expertise if we’re able to share data more freely,” she said.
For those organizations that did have the means to access necessary information and conduct meaningful research, data standardization became a prevalent issue.
“Everyone is interested in quickly capturing data relevant to their own scientific questions on COVID-19 exposures, behaviors, impacts, symptoms, et cetera. And as people launch these efforts, there aren’t a lot of off-the-shelf measures that people can just pull for these new kinds of activities. So many people are inventing their own,” said Lis Nielsen, PhD, director of the Division of Behavioral and Social Research at the National Institute on Aging (NIA).
“And the challenge then becomes, if you have a lot of small studies using different measures, how can you possibly compare the results and know whether we’re coming to common conclusions?”
To ensure that data is harmonized and effectively shared, NIA is working with NIH colleagues to coordinate and catalog the large number of pandemic-specific questionnaires, modules, and scales about COVID-19 experiences. The goal is to discourage the development of one-off survey instruments and facilitate seamless data sharing.
“We have a lot of tools and a lot of collaborative efforts at NIH to make sure that data harmonization is occurring throughout this pandemic. We want to make sure that it is the expectation that all of this data will be shared in an accessible format as soon as possible,” said Kenneth Santora, PhD, director of the Division of Extramural Activities at NIA.
One of these NIH-funded efforts, the National COVID Cohort Collaborative (N3C), is aiming to turn EHR data from hundreds of thousands of coronavirus patients into effective treatments and predictive analytics tools. The project will support the analysis of EHRs on a new, secure database.
“The N3C was founded on the principle that because no one institution has that much data – especially some of the states that have fortunately lower incidence rate for COVID – if we could aggregate and harmonize our data, then we could develop tools like machine learning models and sophisticated statistics,” said Haendel, who serves as the lead investigator of N3C.
“By doing this, we could have enough data to reveal patterns that wouldn’t otherwise be revealed within a single institution.”
New methods, more timely decision-making
In the midst of a global health crisis, it’s not enough for healthcare organizations to form partnerships and develop innovative data-sharing strategies – they must use these resources to make more informed care decisions.
With COVID-19 shining a harsh spotlight on healthcare’s biggest issues, leaders are using the data they have to get ahead of some of the pandemic’s worst side effects, most notably health disparities.
“If you’re collecting data on individuals affected by COVID-19 – particularly on those most affected – it may be very difficult to reach certain subpopulations,” Nielsen said.
“Finding creative ways to overcome those challenges is important if you want to get a full picture of the pandemic as it’s unfolding. And people are creatively thinking in those areas and realizing that it’s an essential part of understanding what’s happening.”
At NIA, researchers are getting creative in reducing gaps in coronavirus care. The institution is addressing COVID-19 testing in underserved communities with the Rapid Diagnostic Accelerator for Underrepresented Populations (RADxUP) initiative. The project seeks to improve COVID-19 testing technology, capacity, and accessibility for vulnerable individuals.
“The initiative is bringing forward these very important questions that investigators always need to ask when conducting clinical research. Hopefully, this initiative will foster that practice in other and future research initiatives.”
Nielsen echoed these hopes.
“We expect that RADxUP will foster continued use of common data sets around the pandemic, and create a model about how this can be done well. The initiative can serve as a data resource that people can use for years going forward,” she said.
New methods of data sharing and collaboration could also accelerate the use of machine learning and other advanced clinical decision support tools, Mathew noted. These technologies require large amounts of data to produce accurate results, and in the whirlwind that is COVID-19, many organizations lack the resources required to train and deploy analytics models.
“Some hospitals may have a very small number of patients, and in the case of a disease like COVID the outcomes are much more imminent. They need to provide patients with the best treatment options very quickly, so being able to share data and learn from other people’s experiences becomes more of a necessity than a nice-to-have,” Mathew said.
“Even in our own hospital setting, I started a COVID registry and we are now working with imaging experts so that they can bring in imaging data. We’re also talking to experts to see how we can use that data to conduct more machine learning and predictive analytics.”
The N3C project is also working to provide a solution to this issue. Researchers are offering access to a secure, cloud-based database containing records of patients who had COVID-19 or were suspected to be infected. Individuals granted access to the database will be able to run algorithms on this patient dataset without seeing actual patient records, leading to important discoveries about the virus.
“The first steps are really focused on validating what others have found before, such as, is there an association between COVID-19, diabetes, and morbidity? What’s really great is that it’s a combination of clinicians, researchers, informaticists, and machine learning experts all working together to define those things. The degree of collaboration that we’re seeing is unprecedented,” Haendel said.
“The goal is to create predictive algorithms that can help inform clinical care. For example, deciding when to intubate somebody or when to use a particular drug. A lot of the investigations are focused on revealing predictive indicators, candidate drugs, or other co-occurring conditions that either exacerbate or protect against the disease.”
Developing strategies that will stand the test of time
In healthcare, the silver lining of the pandemic has been the wave of innovation and collaboration that soon followed. For this brief moment in time, it’s seemed the industry has started operating as a unified whole, and not as many different moving pieces.
The challenge will lie in extending this brief moment beyond COVID-19 and into the future.
“There has been so much advancement in data sharing because of the COVID-19 pandemic. There are so many tools that have been developed,” Santora said.
“We’ve seen a fundamental shift, and moving forward, we need data harmonization across the board because of these concerted efforts. We’ve seen this with vaccines and how rapidly they have been developed. And that can only happen with communication and data sharing in order to meet this common goal.”
Research partnerships, which have played a major role in the pandemic, will greatly benefit the industry going forward. Patients could also have a part in improved data sharing and decision-making, Mathew said.
“On the research side, there’s a lot of emphasis now on team science, and much more collaboration between data science and clinical and molecular research. The pandemic has instilled dependency and a spirit of team science amongst all players. And some people are very interested in day-to-day learning from the data. That’s a trend that I hope and think will continue,” she stated.
More patient-centric data sharing would also benefit the close-knit ecosystem, she added.
“We can integrate it with environmental data, exposures, prior vaccinations, et cetera, and then augment it with the hospital data to do more predictive modeling. In the event that that person gets infected, we are in a much better position to provide the appropriate level of treatment to them.”
COVID-19 has helped the industry overcome data privacy barriers as well – something that will improve research, outcomes, and costs beyond the current pandemic.
“We’re learning a lot about every step of this process, and how to overcome the regulatory constraints of sharing data in a secure way. The data in the N3C enclave has probably some of the highest security for clinical data in the nation, but at the same time, we’re able to allow people to access it and work on it. It’s a very exciting time from that perspective,” Haendel said.
A pilot project within the organization also looking to generate synthetic data, Haendel said, which researchers may be able to share more broadly across stakeholders.
“That would allow a lot more people to have their hands on very real looking data, without the security constraints,” Haendel said. “We hope that project will be successful in validating some of the approaches that are being taken in the initiative. If it’s successful, that would be a long-lasting way in which to better improve access to clinical data.”
Despite all the progress that’s been made throughout the pandemic, the only way to create real, lasting change in the healthcare industry is to continually promote data sharing and collaboration as often as possible, Nielsen said.
“We do what we can to make data resources available publicly to investigators throughout the country, and throughout the world. But there are always new people to get the word out to, and people who maybe have traditionally worked in fields where it was more focused on one laboratory or research group,” she explained.
“There’s a constant educational component to this that we have to be perpetuating as new investigators come up through the pipeline. The more outreach there is about it, the better. And the more discussion of it in the media and the scientific press the better as well.”
When the country thinks of moving past the current situation, most people tend to think about the behaviors and practices they’ll be able to leave behind. For individuals in the healthcare industry, it will be about the practices they should continue long after the pandemic has subsided.
“The COVID-19 pandemic will help foster a new generation of scientists in that mentality of always sharing data and training new people to use common data across all disciplines, in order to better understand what research project you happen to be working on,” Santora said.
“Going forward, those best practices are going to be incorporated and people will understand the importance of sharing data in a timely manner.”
