Intelligent visualization and exploration of time-oriented data of multiple patients
Section snippets
Introduction: intelligent visualization of time-oriented data of multiple patients
A key task facing clinicians and medical researchers is the analysis of time-stamped, longitudinal medical records, particularly, records of multiple patients. This capability is necessary to support, for example, quality assessment tasks, analysis of clinical trials, and the discovery of new clinical knowledge. Although the task of accessing patient data has been solved mostly through the increasing use of electronic medical record (EMR) systems, there still remains the task of intelligent
Combining domain knowledge, temporal abstraction and information visualization in medical informatics
The use of a domain knowledge base can both support an in-depth analysis of longitudinal patient records and simplify and facilitate the data exploration process, since the user can explore only high-level concepts based on complex temporal patterns (or, in general, on any abstract concepts) previously defined in a domain-specific knowledge base and detected in the patients’ data. In addition, it has been demonstrated that visual representation can often communicate information much more
Desiderata for effective exploration of time-oriented data for multiple patients
Our study of the problem of effective and usable visualization and exploration of raw clinical data and, especially, of derived meaningful abstractions from these data, revealed the following set of desiderata for the intelligent interface and exploration operators supporting the task of exploration of time-oriented data for multiple patients.
- 1.
Evaluation of the functionality and usability of the KNAVE-II system for exploration of longitudinal data of individual patients [2] demonstrated the
Methods
The specific methods that we used are presented below in six subsections, each in the context of its relevant desideratum as presented in Section 3.
Example of a clinical scenario
In this section, we present an example of an exploration clinical scenario for application of the VISITORS system and an analysis using a TAC. The example, which relates to a retrospective database of bone-marrow transplantation (BMT) patients (see evaluation Section 6), comprises an investigation of the bone-marrow recovery characteristics of patients, who are either young [<20 years] or old [>70 years], following an autologous BMT procedure.
- (1)
In a previous study [4] we introduced the graphical
Research questions
We designed and developed the VISITORS system according to the desiderata listed in Section 3, and envision it as potentially useful for two types of users: clinicians and medical informaticians. We also envision that the system can be used to answer different clinically motivated questions. We conducted an evaluation of the system with the aim of answering the following four research questions:
- 1.
Overall functionality and usability: are the interactive exploration operators of the VISITORS system
Results
This section summarizes the evaluation results in terms of the four research questions defined in Section 6.1.
Contributions and advantages
The major contribution of the VISITORS system is the provision of a comprehensive environment for intelligent, i.e., knowledge-based, investigation of time-oriented data for multiple patients, including the specification and retrieval, visualization, exploration, and analysis of the time and value associations among both raw and abstract clinical concepts. Based on the results of the current study, the VISITORS system might be described as an “intelligent equalizer” for data interpretation and
Acknowledgments
This research was supported by Deutsche Telekom Labs at Ben-Gurion University of the Negev and the Israel Ministry of Defense, BGU award No. #89357628-01. We thank all the clinicians and medical informaticians who contributed their time to the evaluation. We thank Ms. Efrat German for her work on the Tempura system, and Mr. Ido Hacham and Mr. Shahar Albia for their work on the Multi-TOQ system.
References (62)
A framework for knowledge-based temporal abstraction
Artif Intell Med
(1997)- et al.
Distributed, intelligent, interactive visualization and exploration of time-oriented clinical data and their abstractions
Artif Intell Med
(2006) - et al.
Evaluation of an architecture for intelligent query and exploration of time-oriented clinical data
Artif Intell Med
(2008) - et al.
Utilizing temporal data abstraction for data validation and therapy planning for artificially ventilated newborn infants
Artif Intell Med
(1996) - et al.
Abstraction on clinical data sequences: an object-oriented data model and a query language based on the event calculus
Artif Intell Med
(1999) Information visualization and its application to medicine
Artif Intell Med
(2001)Visualization and interactive analysis of blood parameters with InfoZoom
Artif Intell Med
(2001)Information visualisation in clinical Odontology: multidimensional analysis and interactive data exploration
Artif Intell Med
(2001)- et al.
Data mining on temporal data: a visual approach and its clinical application to hemodialysis
J Vis Lang Comput
(2003) - et al.
Formalizing visual interaction with historical databases
Inform Systs;
(2002)
MQuery: a visual query language for multimedia, timeline and simulation data
J Visual Lang Comput
Visualizing queries on databases of temporal histories: new metaphors and their evaluation
Data Knowl Eng
Solving multi-granularity temporal constraint networks
Artif Intell
Iconic SQL: practical issues in the querying of database through structured iconic expressions
J Visual Lang Comput
A distributed temporal-abstraction mediation architecture for medical databases
Artif Intell Med
Knowledge-based temporal abstraction in clinical domains
Artif Intell Med
User interfaces and visualization
Graphical displays to improve guideline-based therapy of hypertension
Multi-level temporal abstraction for medical scenario construction
Int J Adapt Control
An algorithm for the machine calculation of complex Fourier series
Math Comput
Abstraction and representation of repeated patterns in high-frequency data
A logic-based calculus of events
New Gener Comput
LifeLines: visualizing personal histories
Aligning temporal data by sentinel events: discovering patterns in electronic health records
Visual methods for analyzing time-oriented data
IEEE Trans Visual Comput Graph
Temporal extensions to the relational model and SQL
HSQL: A historical query language
The TSQL2 temporal query language
Querying temporal clinical databases with different time granularities: the GCH-OSQL language
Cited by (91)
Computing and analyzing decision boundaries from shortest path maps
2023, Computers and Graphics (Pergamon)An overview on state-of-the-art electrocardiogram signal processing methods: Traditional to AI-based approaches
2023, Expert Systems with ApplicationsGuidelines and workflow models
2023, Clinical Decision Support and beyond: Progress and Opportunities in Knowledge-Enhanced Health and HealthcareVisualization of frequent temporal patterns in single or two populations
2022, Journal of Biomedical InformaticsIncreasing and standardizing quality of care using computerized guidelines for clinical decision support
2022, Personalized Health Systems for Cardiovascular DiseaseImplementation and evaluation of a multivariate abstraction-based, interval-based dynamic time-warping method as a similarity measure for longitudinal medical records
2021, Journal of Biomedical InformaticsCitation Excerpt :The KBTA method has been shown to produce useful abstractions that are similar to those produced by domain experts in multiple clinical domains, such as oncology, diabetes and children’s growth assessment [47–48], and that are useful for multiple tasks in other domains, such as for detecting malware threats in the cyber security domain [39–40,42] or for visualizing potential malware-effect patterns [41]. KBTA-generated temporal abstractions have also been successfully used to visualize and explore the longitudinal records of individual patients’ data [50,30], and to visualize, explore, and find associations among large numbers of multiple patients’ data [21,22–23]. The same temporal abstractions have been demonstrated to be the basis for automatically generating, in an accurate and highly complete manner, summaries in free text of the multivariate data of longitudinal medical records.