JMIR Publications

ABSTRACT

Background:

A novel disease poses special challenges for informatics solutions: biomedical informatics relies for the most part on structured data; structured data require a pre-existing data/knowledge model; but novel diseases do not have pre-existing knowledge models. In an emergent epidemic, language processing could allow rapid conversion of unstructured text to a novel knowledge model. But, although this has often been suggested, there has never before been an opportunity to actually test that claim in real time. The current pandemic presents such an opportunity.

Objective:

To evaluate the added value of information from clinical text in the response to emergent diseases.

Methods:

We explored the effect of long-term treatment by calcium channel blockers on the outcome of COVID infection in patients with high blood pressure during in-patient hospital stay, using two sources of information: data available strictly from structured Electronic Health Records and data available through structured Electronic Health Records and text mining.

Results:

In this multicenter study involving 39 hospitals, text mining increased statistical power sufficiently to change a negative result for an adjusted hazard ratio to a positive one. Compared to the baseline structured data, it increased the number of patients available for inclusion in the study by 2.95 times, the amount of available information on medications by 7.2 times, and additional phenotypic information by 11.9 times.

Conclusions:

Calcium channel blockers are associated with decreased in-hospital mortality in patients with COVID-19 infections. This was found because a natural language processing pipeline could be adapted quickly to the domain of the novel disease and still perform well enough to extract useful information. When that information is used to supplement existing structured data, a sample size can be increased enough to see treatment effects that were not previously statistically detectable.