Working in partnership with Epic, through which it will be delivered, the Cleveland Clinic has released a new COVID-19 risk prediction model drawing on both patient entered data and clinical information.
Researchers at the Cleveland Clinic developed and tested the model using clinical data from more than 11,000 of its patients. The Clinic used retrospective patient data from patients tested for COVID-19 at locations in Northeast Ohio and Florida. Data scientists then transformed the data into a statistical risk-prediction model.
When in use by a provider, the tool integrates both medical records drawn from Epic and patient information collected by the patient in MyChart (Epic’s patient portal). As part of the process, patients complete a short self-assessment in MyChart addressing symptoms they might have and their potential exposure to COVID-19.
The predictive model pulls this information into the mix with clinical and demographic data already in the chart to calculate their risk score. When patients receive a high-risk score, they are advised to receive a test and their care team members can be notified.
This news comes as part of a larger landscape in which data science expertise and AI tools will be as important to taming ongoing COVID-19 issues as direct clinical intervention was during the initial wave of the pandemic. You may remember that John highlighted a new AI Model for a patient’s oxygen needs as another example of what’s being worked on.
It’s little surprise to learn that in the Dell Technologies 2020 Digital Transformation Index, 74% of healthcare respondents were confident that using their organization’s data and AI tools will help speed up efforts to develop medicines and technologies to fight COVID-19.
That being said, while there’s no reason to doubt the efforts of a methodical and well-resourced organization like the Cleveland Clinic, providers still need to take care when attempting to use health IT to tame a still-emerging pandemic. The rush to track and manage COVID-19 data has taken a toll of its own already.
For example, a survey released in July found that some healthcare providers feel that the rush to tackle COVID-19 may actually have worsened existing problems with their infrastructure and workflow.
The survey, which was sponsored by secure document transfer solutions vendor Biscom, found that 90% of respondents felt that COVID-19 related tech efforts had highlighted existing problems in their infrastructure.
Meanwhile, 75% of respondents agreed that COVID-19-related workarounds put in place had increased security and privacy risks related to patient data. Also, 43% said that given how quickly their organization brought on new tools, it was unlikely that the tools had been vetted as much as they would have been otherwise.
None of this is to say that the Cleveland Clinic’s model isn’t sound. However, it’s worth noting that given the frantic pace at which providers have shored up their infrastructure during the pandemic, there might be leakages in their data management process that they haven’t discovered as of yet. It might be wise for health systems to consider the soundness of their current dataset before they commit to making predictive models like this a backbone of their COVID-19 efforts.