Search

BackgroundAn established and well-known method for usability assessment of various human-computer interaction technologies is called heuristic evaluation (HE). HE has been adopted for evaluations in a wide variety of specialized contexts and with objectives that go beyond usability. A set of heuristics to evaluate how health information technologies (HITs) incorporate features that enable effective patient use of person-generated health data (PGHD) is needed in an era where there is a growing demand and variety of PGHD-enabled technologies in health care and where a number of remote patient-monitoring technologies do not yet enable patient use of PGHD. Such a set of heuristics would improve the likelihood of positive effects from patients’ use of PGHD and lower the risk of negative effects. ObjectiveThis study aims to describe the development of a set of heuristics for the design and evaluation of how well remote patient therapeutic technologies enable patients to use PGHD (PGHD enablement). We used the case of Kinect-based stroke rehabilitation systems (K-SRS) in this study. MethodsThe development of a set of heuristics to enable better use of PGHD was primarily guided by the R3C methodology. Closer inspection of the methodology reveals that neither its development nor its application to a case study were described in detail. Thus, where relevant, each step was grounded through best practice activities in the literature and by using Nielsen’s heuristics as a basis for determining the new set of heuristics. As such, this study builds on the R3C methodology, and the implementation of a mixed process is intended to result in a robust and credible set of heuristics. ResultsA total of 8 new heuristics for PGHD enablement in K-SRS were created. A systematic and detailed process was applied in each step of heuristic development, which bridged the gaps described earlier. It is hoped that this would aid future developers of specialized heuristics, who could apply the detailed process of heuristic development for other domains of technology, and additionally for the case of PGHD enablement for other health conditions. The R3C methodology was also augmented through the use of qualitative studies with target users and domain experts, and it is intended to result in a robust and credible set of heuristics, before validation and refinement. ConclusionsThis study is the first to develop a new set of specialized heuristics to evaluate how HITs incorporate features that enable effective patient use of PGHD, with K-SRS as a key case study. In addition, it is the first to describe how the identification of initial HIT features and concepts to enable PGHD could lead to the development of a specialized set of heuristics.

BackgroundPerson-generated health data (PGHD) are health data that people generate, record, and analyze for themselves. Although the health benefits of PGHD use have been reported, there is no systematic way for patients to measure and report the health effects they experience from using their PGHD. Patient-reported outcome measures (PROMs) allow patients to systematically self-report their outcomes of a health care service. They generate first-hand evidence of the impact of health care services and are able to reflect the real-world diversity of actual patients and management approaches. Therefore, this paper argues that a PROM of utilizing PGHD, or PROM-PGHD, is necessary to help build evidence-based practice in clinical work with PGHD. ObjectiveThis paper aims to describe a method for developing PROMs for people who are using PGHD in conjunction with their clinical care—PROM-PGHD, and the method is illustrated through a case study. MethodsThe five-step qualitative item review (QIR) method was augmented to guide the development of a PROM-PGHD. However, using QIR as a guide to develop a PROM-PGHD requires additional socio-technical consideration of the PGHD and the health technologies from which they are produced. Therefore, the QIR method is augmented for developing a PROM-PGHD, resulting in the PROM-PGHD development method. ResultsA worked example was used to illustrate how the PROM-PGHD development method may be used systematically to develop PROMs applicable across a range of PGHD technology types used in relation to various health conditions. ConclusionsThis paper describes and illustrates a method for developing a PROM-PGHD, which may be applied to many different cases of health conditions and technology categories. When applied to other cases of health conditions and technology categories, the method could have broad relevance for evidence-based practice in clinical work with PGHD.

Background: Collecting information on adverse events following immunization from as many sources as possible is critical for promptly identifying potential safety concerns and taking appropriate actions. Febrile convulsions are recognized as an important potential reaction to vaccination in children aged <6 years., Objective: The primary aim of this study was to evaluate the performance of natural language processing techniques and machine learning (ML) models for the rapid detection of febrile convulsion presentations in emergency departments (EDs), especially with respect to the minimum training data requirements to obtain optimum model performance. In addition, we examined the deployment requirements for a ML model to perform real-time monitoring of ED triage notes., Methods: We developed a pattern matching approach as a baseline and evaluated ML models for the classification of febrile convulsions in ED triage notes to determine both their training requirements and their effectiveness in detecting febrile convulsions. We measured their performance during training and then compared the deployed models' result on new incoming ED data., Results: Although the best standard neural networks had acceptable performance and were low-resource models, transformer-based models outperformed them substantially, justifying their ongoing deployment., Conclusions: Using natural language processing, particularly with the use of large language models, offers significant advantages in syndromic surveillance. Large language models make highly effective classifiers, and their text generation capacity can be used to enhance the quality and diversity of training data., (©Sedigh Khademi, Christopher Palmer, Muhammad Javed, Gerardo Luis Dimaguila, Hazel Clothier, Jim Buttery, Jim Black. Originally published in JMIR AI (https://ai.jmir.org), 30.08.2024.)