Digital Health for Space: Towards Prevention, Training, Empowerment, and Autonomy

Modern terrestrial healthcare systems are increasingly burdened by aging demographics, the growing prevalence of chronic diseases, and the widespread shortage of healthcare professionals. Concepts such as prevention, self-management of health, and decentralized care are widely recognized as crucial. However, their practical implementation remains limited due to fragmented infrastructures, unclear responsibilities, and underdeveloped technological ecosystems.

In remote or underserved regions, disaster and conflict zones, people already face similar limitations to those in space: delayed or irregular access to care, reliance on simplified telemedicine solutions and limited real-time monitoring. Examples include chronically ill patients who are rarely consulted remotely, emergency responders who work with minimal diagnostic tools, or elderly people in long-term care who are not constantly monitored. These terrestrial conditions are similar, to a lesser extent, to the restrictions that apply to space missions.

However, human spaceflight magnifies structural challenges by turning slow-moving terrestrial problems into immediate survival risks. Microgravity accelerates musculoskeletal decline [19]. Radiation exposure increases significantly beyond Earth’s magnetic shielding [22]. Communication may be delayed, intermittent, or completely unavailable. Delays begin at the Lunar Orbital Platform-Gateway (LOP-G) and become more severe on missions to Mars and beyond [13]. In such environments, medical evacuation is impossible and real-time support from Earth can no longer be relied on.

Space exploration demands the development of resilient and autonomous healthcare systems that integrate prevention, individual empowerment, and real-time situational awareness. Human spaceflight environments offer a unique testing ground for developing digital health technologies and models. Once validated under extreme conditions, these solutions can be transferred to terrestrial healthcare systems that urgently require transformation.

This position paper uses space exploration as a strategic lens to investigate how digital health technologies, especially artificial intelligence (AI)-based decision support, continuous multimodal sensing, and adaptive human-computer interaction (HCI), must co-evolve to support individual autonomy and proactive health management. We propose a situational framework of autonomy levels across four mission stages, Earth orbit, lunar vicinity (e.g. Lunar Gateway), Mars, and deep space, and outline the system-level requirements at each stage.

Through this lens, we highlight LOP-G as a concrete near-future scenario, where moderate autonomy, intermittent support, and resource constraints converge. LOP-G illustrates the urgent need for trustworthy, transparent, and user-adaptive digital health ecosystems. Our argument is that the co-development of monitoring technologies, AI, and HCI in space not only provides mission-critical capabilities, but also transferable insights for Earth. Advances made under extreme conditions can accelerate progress in preventive healthcare, education, and self-management, particularly in terrestrial systems challenged by workforce shortages and rising costs.

Future digital health systems in space must dynamically adapt to the conditions of each mission. Key variables include increasing distance from Earth, longer communication delays, and rising demands for autonomous operation beyond low Earth orbit. Figure 1 introduces a structured framework with four progressive levels of autonomy, defined by mission distance and expected communication latency. The levels correspond to specific health management scenarios, operational constraints, and system requirements. Each stage also reflects analogous conditions on Earth, where similar limitations affect the delivery of medical care.

• $\mathrm{■}$

  Level 0 (Real-time Telemedicine) - Earth Orbit (ISS / Low Earth Orbit) Full external medical support available. Clinical tasks are guided by Earth-based experts with minimal onboard autonomy.

  • –

    Communication Delay: ~0.03–0.25 s (one-way), effectively real-time.

  • –

    Return to Earth: ~3-6 h (e.g., via Soyuz or Crew Dragon; emergency concepts such as UC Davis’ “Space Ambulance” aim for ~4 h return capability).

  • –

    System Requirements: Basic health monitoring (e.g., ECG, heart rate), minimal explainable AI support, dependence on ground control.

  • –

    Earth Analogue: Urban hospital settings or intensive care units with constant expert supervision and high-bandwidth infrastructure.

• $\mathrm{■}$

  Level 1 (Conditional Autonomy) - Moon (LOP-G / Lunar Surface)

  Requires moderate onboard autonomy to bridge intermittent connections and resource constraints.

  • –

    Communication Delay: ~1.2–2.6 s (one-way).

  • –

    Return to Earth: ~3-5 d (depending on launch readiness and transfer trajectory).

  • –

    System Requirements: Continuous wearable and behavioral monitoring, hybrid AI, basic digital twin models, asynchronous decision support, adaptive HCI for varying users.

  • –

    Earth Analogue: Remote outposts, submarines, or Antarctic research stations with delayed telemedical support and non-specialist medical personnel.

• $\mathrm{■}$

  Level 2 (High Autonomy) - Mars (Transit / Surface)

  Autonomous health operations must compensate for extended delays and mission-critical isolation.

  • –

    Communication Delay: 4–24 min (one-way), depending on the planetary alignment.

  • –

    Return to Earth: 6-9 month (e.g., minimum-energy transfer), possibly >1 year if waiting for next return window.

  • –

    System Requirements: Comprehensive multimodal monitoring, MGM with digital twins, advanced predictive analytics, decision support, highly contextualized HCI.

  • –

    Earth Analogue: Disaster zones or warzones with disrupted infrastructure, limited staff, and time-critical care needs.

• $\mathrm{■}$

  Level 3 (Full Autonomy) - Beyond (Deep Space Missions)

  Complete onboard autonomy for diagnostics, interventions, and preventive care. No possibility of medical evacuation or remote assistance.

  • –

    Communication Delay: Several hours to no contact, depending on location and antenna pointing (e.g., beyond Mars orbit or into interstellar space).

  • –

    Return to Earth: Not possible (no feasible rescue within human lifetime or mission design).

  • –

    System Requirements: Self-sufficient care: AI-supported diagnosis and intervention, robotic support, transparent models, adaptive HCI to sustain crew trust and self-efficacy.

  • –

    Special Case: The enthralling and to date still visionary approach of hibernation (torpor) for the crew may require continuous AI-based vital control without human oversight [5].

  • –

    Earth Analogue: Isolated elderly in rural regions or long-term care settings without specialist access, especially in crisis or climate-affected regions.

LOP-G represents a realistic near-term scenario to assess both the technological readiness and operational integration of conditional autonomy systems (Level 1). In contrast to the ISS, crew on LOP-G will periodically operate with limited external support. This creates a demand for moderate on-board decision-making capabilities and greater reliance on predictive monitoring and self-managed care. LOP-G therefore provides a crucial test environment for developing health management systems that must function under the more extreme conditions anticipated for Mars missions and beyond.

A structured continuum of situation-specific autonomy demands that health systems identify critical mission contexts, adjust operational modes, and follow predefined procedural models. These systems must remain robust and context-sensitive to enable smooth transitions between supported and fully autonomous care, especially in scenarios where evacuation is not possible and communication with Earth is delayed.

Numerous foundational technologies for space-relevant digital health have already been demonstrated aboard the International Space Station (ISS). Wearable biosensors, including smart garments such as ESA’s MagIC-Space vest, have allowed unobtrusive monitoring of ECG, respiration, temperature, and movement during sleep without compromising astronaut comfort or mobility [8]. Radiation dosimetry has also achieved reached a high level of maturity. For example, ESA’s active dosimeter continuously provides real-time data on the absorbed dose and the equivalent doses within the spacecraft [22]. Robotic systems such as Robonaut 2 have been used to explore remotely operated diagnostic procedures, including ultrasound imaging and simulated injections [20]. However, these systems remain largely disconnected. They are developed independently and tested under idealized conditions, either within the resource-rich environment of the ISS or in terrestrial contexts such as hospital settings and home care. As a result, their transferability is limited. Technologies designed for short-term clinical monitoring on Earth often perform poorly under the physiological stress, behavioral variability, and environmental constraints encountered in space or in remote terrestrial locations, such as polar research stations, disaster zones, or submarines. Three major categories of integration challenges can be identified:

• $\mathrm{■}$

  Sensor Fusion and Environmental Validity: Wearable, ingestible, and ambient sensors are usually validated in stable terrestrial environments. Their reliability in dynamic and extreme conditions, such as microgravity, partial gravity, or high-radiation zones, remains uncertain [21, 1]. Motion, fluid redistribution, and changes in sensor-skin adhesion may lead to physiological baseline shifts and signal artifacts. These conditions challenge conventional signal processing methods. To ensure reliable measurements throughout different mission phases, systems require redundant sensing modalities, context-aware filtering, and adaptive calibration strategies.

• $\mathrm{■}$

  AI Maturity and Contextualization: On Earth, AI systems increasingly support diagnostics, triage, and predictive analytics. These typically rely on large datasets from well-characterized populations. However, in space or other isolated settings, current AI applications often lack transparency, robustness, and the ability to adapt to mission-specific constraints. Most systems do not integrate knowledge of environmental stressors, resource limitations, or procedural workflows. Future architectures must combine data-driven approaches with structured evidence-based models and formal process representations to support interpretability and simulation-based decision making under uncertainty.

• $\mathrm{■}$

  Human-System Interaction: Usability under mission-relevant stressors, such as cognitive load, restricted mobility due to gloves or microgravity, and time-critical scenarios, has rarely been systematically studied. Many current HCI approaches follow engineering-driven designs and are not developed in collaboration with end users [17, 3]. Research shows that a lack of transparency, limited adaptability, and insufficient participatory design reduce user trust and adherence to AI-supported systems [4]. These risks are especially relevant for long-duration missions and future space tourism, where users may have limited medical training.

These open challenges require a shift in research and development strategies. Future digital health systems must be designed as cohesive and adaptive ecosystems. Rather than building isolated technical modules, developers must integrate robust sensor platforms, interpretable AI, and interaction systems that support transparent, interpretable, and user-adaptive HCI. Health autonomy should scale based on mission constraints while remaining transparent and understandable for all crew members, regardless of background or role.

Technological readiness should be evaluated not only by the existence of functional prototypes but also by integration maturity, adaptability to different mission phases, and usability under operational stress. Platforms such as the LOP-G, with intermittent communication and evolving autonomy requirements, offer a unique opportunity to test integrated systems under controlled yet realistic spaceflight conditions.

Digital health in space must be developed as an integrated framework. Progress in three key areas is required to support autonomous health operations during human spaceflight: continuous multimodal monitoring, structured AI, and adaptive HCI. Figure 2 presents this perspective as a modular system framework. HCI serves as a design principle that organizes the integration of enabling technologies such as multimodal sensing, interactive user interfaces, and the medical service hub with operational health contexts such as telemedical support, autonomous crew care, and smart health systems. HCI provides the conceptual basis for adaptive, interpretable, and trustworthy interaction between humans and system components. The framework allows closed-loop configurations that maintain health-related functions under varying levels of connectivity. Although bidirectional communication with Earth-based telemedical services remains possible in some scenarios, it cannot serve as a reliable baseline. The figure illustrates the need for integrated development across all components. Design decisions in sensing, inference, or interaction must reflect the requirements of each health context and support coherence throughout the framework.

To illustrate how the technological ecosystem evolves alongside mission-specific autonomy, Figure 1 also outlines selected milestones across the domains of monitoring, AI, and HCI. At Level 0, early-stage HCI interfaces serve primarily as research tools to observe crew interaction and guide future system development. On the Moon (Level 1), procedural AI guidance and initial digital astronaut models enable asynchronous decision support. Mars missions (Level 2) demand context-aware hybrid models and adaptive interfaces to support time-sensitive crew decisions without real-time supervision. Finally, deep space scenarios (Level 3) require fully integrated self-care ecosystems, including self-learning and self-correcting AI-HCI systems capable of autonomous adaptation and operational robustness in the absence of fallback options.

Autonomous healthcare in space depends not only on technical systems but also on the crew’s preparedness to understand, trust, and act upon health-related data and AI-generated recommendations. As autonomy increases with mission distance, astronauts must assume new roles as informed decision-makers, caregivers, and active participants in managing their health.

To realize the vision of autonomous, user-centered digital health systems in space, several strategic research directions must be pursued. These extend beyond individual technologies and emphasize system integration, operational realism, and ethical deployment under extreme conditions.

The paper outlines a forward-looking vision for digital health in space, centered on prevention, individual empowerment, and adaptive autonomy. We introduced a structured framework of autonomy levels to demonstrate how mission distance and communication constraints shape health system requirements. Within our framework, we emphasized the co-evolution of multimodal monitoring, model-guided and explainable AI, and adaptive HCI as the foundation for context-aware and trustworthy care.

Rather than treating sensing, AI, and HCI as isolated technologies, we argue for their integrated development through user-centered design and iterative validation. Human factors, learning support, and ethical oversight are not secondary features, but essential design principles that allow autonomy without compromising safety or comprehension.

The Lunar Gateway provides a critical near-term platform to evaluate these concepts under realistic operational constraints. Insights gained from timely space missions will inform a stepwise advancement of digital health systems for more complex scenarios, including Mars transit and deep-space habitats.

Beyond their relevance for spaceflight, the digital health technologies and principles discussed here address urgent needs on Earth. Remote diagnostics, self-managed care, and adaptive AI support have the potential to transform healthcare delivery in underserved regions, during crises, and for aging populations. By investing in digital space health, we also invest in the resilience and future-readiness of healthcare on Earth.