Appendix: Summary of Best Papers Selected for the 2021 Edition of the IMIA Yearbook,
Clinical Research Informatics Section
Bahmani A, Alavi A, Buergel T, Upadhyayula S, Wang Q, Ananthakrishnan SK, Alavi A,
Celis D, Gillespie D, Young G, Xing Z, Nguyen MHH, Haque A, Mathur A, Payne J, Mazaheri
G, Li JK, Kotipalli P, Liao L, Bhasin R, Cha K, Rolnik B, Celli A, Dagan-Rosenfeld
O, Higgs E, Zhou W, Berry CL, Van Winkle KG, Contrepois K, Ray U, Bettinger K, Datta
S, Li X, Snyder MP
A scalable, secure, and interoperable platform for deep data-driven health management
Nat Commun 2021 Oct 1;12(1):5757
This paper presents a major effort in building a secured and scalable platform to
gather big biomedical data from different sources (including genomics, EHRs, wearable
sensors). The authors focus on technical aspects of building such a platform: security
(local data storage, defense against reverse engineering, mobile app security, anonymization),
scalability (authentication, messaging, machine learning, infrastructure based on
the open-source tool Terraform) and analysis (data preprocessing, feature extraction).
Although interoperability issues are not really mentioned, several APIs for data collection
are described. The main features of the platform are data visualization, monitoring
and alerts, as well as feature prediction with logistic regression (84 features).
The platform can be used at patient-level or at cohort-level and has been used for
the detection of pre-symptomatic COVID-19 cases, and for biological characterization
of insulin-resistance heterogeneity. In conclusion, this very large-scale platform
for biomedical data offers guarantees in terms of security, scalability, data preprocessing
and provides features for visualization, monitoring and analysis.
Cheng AC, Duda SN, Taylor R, Delacqua F, Lewis AA, Bosler T, Johnson KB, Harris PA
REDCap on FHIR: Clinical Data Interoperability Services
J Biomed Inform 2021 Sep;121:103871
This paper describes the development and evaluation of the REDCap Clinical Data Interoperability
Services (CDIS) module that provides seamless data exchange between the REDCap research
Electronic Data Capture (EDC) system and any EHR system with a FHIR API without project-by-project
involvement from Health Information Technology staff. An iterative process has been
used to design all aspects of the CDIS module (access control, authentication, variable
selection, and mapping) in such a way that end users could easily set up and use the
module in 2 use cases. In the “Clinical Data Pull” (CDP) mode the CDIS automatically
pulls EHR data into user-defined REDCap fields. In the “Clinical Data Mart (CDM)”
mode, the CDIS collects all specified data for a patient over a given time. Beyond
the stakeholders group initially involved including Vanderbilt University Medical
Center (VUMC) health IT and EPIC EHR teams, other healthcare organizations and EHR
vendors have been associated through the REDCap consortium. As of Nov 2020, since
its release (1st CDP project live@VUMC launched Q3 2018, REDCap released on Epic App
Orchard Q1 2019), 82 projects are running at VUMC (55 CDP, 27 CDM) with 19.5 M data
points transferred. With a large scale adoption in REDCap consortium sites (26 implementations
in other institutions with EPIC EHRs / 47 ongoing implementations in institutions
with EPIC (n=26) or Cerner (n=9) EHRs), the REDCap Clinical Data and Interoperability
Services (CDIS) are key contributions to the integration of care and research activities.
Thanks to the CDIS modules, leveraging the FHIR standard to use of EHR as electronic
source for clinical research, the researchers can self-service the setup of real time
and direct data extraction from the EHR reducing the need for manual transcription
and flat file uploads and improving the accuracy and efficiency of EHR data collection.
Pedrera-Jiménez M, García-Barrio N, Cruz-Rojo J, Terriza-Torres AI, López-Jiménez
EA, Calvo-Boyero F, Jiménez-Cerezo MJ, Blanco-Martínez AJ, Roig-Domínguez G, Cruz-Bermúdez
JL, Bernal-Sobrino JL, Serrano-Balazote P, Muñoz-Carrero A
Obtaining EHR derived datasets for COVID-19 research within a short time: a flexible
methodology based on Detailed Clinical Models
J Biomed Inform 2021 Mar;115:103697
Responding to the urgent need for health data insights during the COVID-19 pandemic,
and utilizing this as a use case for a generalizable methodology, Pedrera-Jiménez
et al report on the use of Detailed Clinical Models (DCMs) as a formalized representation
for the structure and semantics of research data sets. They propose this as a data
transformation pathway to generate datasets rapidly and accurately from EHRs for secondary
use, without loss of meaning or error, allowing for frequent changes in specification,
and being easy to validate.
The authors took as their use case the need to rapidly generate a research data set
conforming to the International Severe Acute Respiratory and emerging Infection Consortium
(ISARIC-WHO) COVID-19 specification. Instead of the classical approach of authoring
this data set in an electronic case report form (eCRF), the authors modelled this
research data set as a portfolio of Detailed Clinical Models: EHR archetypes conforming
to the ISO 13606 standard. These archetypes each expressed a specific data structure
pattern which was a profiled subset of the generic 13,606 EHR interoperability reference
model, and incorporated semantic constraints (e.g,. value sets) drawn from SNOMED-CT
or LOINC, as appropriate for clinical or laboratory concepts. These DCMs were used
as the data extraction mapping target from the EHR system at the Hospital Universitario
12 de Octubre in Madrid. The extraction included data on 4,489 patients hospitalised
with COVID-19 over a six-month period during 2020. The flexibility and agility of
this method was demonstrated through the ability to revise the data set specification
easily by modifying or adding further archetypes. The authors discuss the future potential
of this method to also utilise HL7 FHIR resources as alternative DCMs, through a forthcoming
ISO Technical Specification on “Guidelines for implementation of HL7/FHIR based on
ISO 13940 and ISO 13606”. There is a growing need and opportunity for the systematic
and interoperable reuse of routinely collected (real-world) EHR data for research,
through the centralised or federated querying of standardised data sets. This research,
although mono-centric and exemplified through COVID-19, is included as a best paper
because the methodology is generalisable to any area of clinical research for which
there is relevant real-world data.
