BACKGROUND
Since the creation of the Problem Oriented Medical Record, the building of problem lists has been the focus of many researches. To this day, this issue is not well resolved, and building an appropriate contextualized problem list is still a challenge.
OBJECTIVE
This paper presents the process of building a shared multi-purpose common problem list at the University Hospitals of Geneva, a consortium of all public hospitals and 30 outpatient clinics of the state of Geneva. This list aims at bridging the gap between clinicians’ language expressed in free text and secondary usages requiring structured information.
METHODS
The strategy focuses on the needs of clinicians by building a list of uniquely identified expressions to support their daily activities. In a second stage, these expressions are connected to additional information, building a complex graph of information. A list of 45,946 expressions manually extracted from clinical documents has been manually curated and encoded in multiple semantic dimensions, such as ICD-10, ICPC-2, SNOMED-CT or dimensions dictated by specific usages, such as identifying expressions specific to a domain, a gender, or an intervention. The list has been progressively deployed for clinicians with an iterative process of quality control, maintenance and improvements, including addition of new expressions, or dimensions for specific needs. The problem management of the electronic health record allowed to measure and correct the encoding based on real-world usage.
RESULTS
The list was deployed in production in January 2017 and was regularly updated and deployed in new divisions of the hospital. In 4 years, 684,102 problems were created using the list. The proportion of free text entries reduced progressively from 37.47% (8,321/22,206) in December 2017 to 18.38% (4,547/24,738) in December 2020.
In the last version of the list, over 14 dimensions were mapped to expressions, among them 5 international classifications and 8 other classifications for specific usages. The list became a central axis in the EHR, being used for many different purposes linked to care such as surgical planning or emergency wards, or in research, for various predictions using machine learning techniques.
CONCLUSIONS
This work breaks with common approaches primarily by focusing on real clinicians’ language when expressing patient’s problems and secondly by mapping whatever is required, including controlled vocabularies to answer specific needs. This approach improves the quality of the expression of patients’ problems, while allowing to build as many structured dimensions as needed to convey semantics according to specific contexts. The method is shown to be scalable, sustainable and efficient at hiding the complexity of semantics or the burden of constraint structured problem list entry for clinicians. Ongoing work is analyzing the impact of this approach at influencing how clinicians express patient’s problems.