Clinical encounter heterogeneity and methods for resolving in networked EHR data: a study from N3C and RECOVER programs

Abstract

Objective: Clinical encounter data are heterogeneous and vary greatly from institution to institution. These problems of variance affect interpretability and usability of clinical encounter data for analysis. These problems are magnified when multisite electronic health record (EHR) data are networked together. This article presents a novel, generalizable method for resolving encounter heterogeneity for analysis by combining related atomic encounters into composite "macrovisits."

Materials and methods: Encounters were composed of data from 75 partner sites harmonized to a common data model as part of the NIH Researching COVID to Enhance Recovery Initiative, a project of the National Covid Cohort Collaborative. Summary statistics were computed for overall and site-level data to assess issues and identify modifications. Two algorithms were developed to refine atomic encounters into cleaner, analyzable longitudinal clinical visits.

Results: Atomic inpatient encounters data were found to be widely disparate between sites in terms of length-of-stay (LOS) and numbers of OMOP CDM measurements per encounter. After aggregating encounters to macrovisits, LOS and measurement variance decreased. A subsequent algorithm to identify hospitalized macrovisits further reduced data variability.

Discussion: Encounters are a complex and heterogeneous component of EHR data and native data issues are not addressed by existing methods. These types of complex and poorly studied issues contribute to the difficulty of deriving value from EHR data, and these types of foundational, large-scale explorations, and developments are necessary to realize the full potential of modern real-world data.

Conclusion: This article presents method developments to manipulate and resolve EHR encounter data issues in a generalizable way as a foundation for future research and analysis.

Keywords: database; electronic health records; informatics.

Figure 1.

Visual representation of macrovisit aggregation. Overlapping microvisits are merged into a continuous, bundled macrovisit and the earliest visit start date and latest visit end date are assigned to the macrovisit. Microvisits that occur ≥1 calendar day later with no other, overlapping microvisit generate distinct macrovisits.

Figure 2.

Resource density by length of stay. Visual representation of the variation in maximum resource density across macrovisits as LOS changes. Each color corresponds to a maximum resource density bin. The macrovisits with >50 maximum resources roughly follow the expected inpatient LOS distribution, and a majority of 0-day LOS macrovisits have no more than 25 maximum resources. LOS: length-of-stay.

Figure 3.

Microvisit heterogeneity within macrovisits. Visual representation of varied composition of a randomly selected macrovisit for each site. Small colored markers indicate individual microvisits included in the macrovisit. For example, site 37 has a macrovisit consisting of one longitudinal inpatient visit with a variety of 0-day visits spread throughout the stay. Site 8 has a macrovisit consisting of 2 overlapping inpatient stays, again with a variety of 0-day visits over the entire macrovisit.

Figure 4.

Length-of-stay. LOS distributions for inpatient visits, macrovisits, and high-confidence hospitalizations for the subset of sites with most variance between raw data and algorithm results (median = circle, mean = triangle, IQR = line). LOS: length-of-stay.

Figure 5.

Microvisit density. Distribution of the number of component microvisits within macrovisits and high-confidence hospitalizations for the subset of sites with most variance between macrovisit algorithm and high-confidence hospitalization algorithm results (median = circle, mean = triangle, IQR = line).

Figure 6.

Measurement density. Distribution of the number of measurements within inpatient visits, macrovisits, and high-confidence hospitalizations for the subset of sites with most variance between raw data and algorithm results (median = circle, mean = triangle, IQR = line).