The intersection of driving and health: Toward a framework for healthy driving

Document Type

Article

Source of Publication

Journal of Transport and Health

Publication Date

5-1-2026

Abstract

Introduction This paper presents an investigation into health-conscious driving, a dimension that has received limited attention compared to traditional focuses on road safety and eco-efficiency. To address this gap, we introduce the concept of healthy driving as a complementary objective in Intelligent Transportation Systems (ITS), which emphasizes the physiological well-being of drivers and examines how driving behaviors and environmental conditions are associated with physiological load. In an observational on-road case study in Muscat (Oman) involving 33 drivers, our objective is to identify contexts in which physiological indicators cross clinically relevant thresholds or show marked individualized deviations. Such events are treated as associatioal signals of increased physiological strain rather then indicators of immediate or long-term health risks. Methods A dataset of real-world GPS trajectories and physiological parameters from a group of drivers employed by a governmental institution in Muscat, Oman, was collected over a 13-month period. Initially, 58 professional male drivers were enrolled. Based on pre-specified data-quality criteria, only 33 were retained for the study. Despite their uniform application, we acknowledge that these selection criteria may introduce bias toward participants with more complete or higher-quality recording. We used a multi-phase approach combining data exploration (including geospatial mapping), driver profiling (clustering), and predictive modeling (an attention-augmented LSTM) to estimate short-horizon abnormally high physiological states (heart rate, systolic Blood Pressure - BP) from driving behavioral (harsh acceleration and breaking) and road contextual (curvatures) information under group-aware cross-validation. Results Results show that short-horizon abnormal physiological states can be reliably inferred from driving behavior and road infrastructure features. Across cross-validated analyses, performance gains were observed under class-balanced optimization, with consistently strong discrimination for heart rate and systolic blood pressure. The findings suggest that routine exposure to specific road contexts is associated with measurable physiological load. These results are associational and hypothesis-generating. They underscore the need to redefine Intelligent Transportation Systems (ITS) to account for driver health outcomes. They furthermore advocate for the development of health-aware infrastructure, monitoring tools, and adaptive in-vehicle systems. Conclusions Our findings suggest the feasibility of health-aware analytics and outline the HEALD (Holistic Evaluation and Analytics for Livable Driving) framework, a modular roadmap for health-aware monitoring and early warning. HEALD includes a Driving Early Warning Score (DEWS) schema specifying risk banding and alert logic, with operating thresholds to be validated prospectively. The interpretation of the findings should be cautious given the male-only sample, wearable sensor accuracy limits, and the constraints of the field randomization (e.g., unavoidable traffic and weather variation, etc.). To deliver proactive tools for safety and well-being within the HEALD roadmap, our future work will focus on: (i) recruiting broader, sex-balanced, and geographically diverse driver samples; (ii) conducting prospective, preregistered validation of DEWS thresholds; and (iii) evaluating driver coaching, routing, and speed-moderation interventions.

ISSN

2214-1405

Publisher

Elsevier BV

Volume

48

Disciplines

Computer Sciences | Medicine and Health Sciences

Keywords

Intersection (aeronautics) (0.66) | Computer science (0.56) | Artificial intelligence (0.4) | Computer vision (0.32) | Noise (video) (0.29) | Perception (0.29) | Face (sociological concept) (0.28) | Poison control (0.26) | Action (physics) (0.25) | Context (archaeology) (0.25)

Scopus ID

105026581950

Indexed in Scopus

yes

Open Access

no

Link to Full Text

Share

COinS