Fluid Independence: Powered Wheelchair Users' Perspectives on Autonomy, Care, and Assistive Technology
Md Tanzil Shahria, Nayan Banik, Md Samiul Haque Sunny, Mohammad H Rahman · 2026 · Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI '26) · doi:10.1145/3772318.3791867
Summary
This CHI 2026 paper reports a large qualitative study of 55 powered-wheelchair users (PWUs) across ten U.S. states, using 45-minute semi-structured interviews and Braun-and-Clarke thematic analysis to examine how independence is experienced, negotiated, and mediated by assistive technology, caregiving relationships, and structural systems. Participants spanned spinal cord injury, quadriplegia, multiple sclerosis, spinal muscular atrophy, cerebral palsy, and other mobility-affecting conditions, and their caregiving arrangements ranged from one to twenty-four hours of daily assistance (mean 8.7 hours). The study is grounded theoretically in critiques of static AT models (HAAT, SETT, PEO), Sen's Capability Approach, and disability-studies framings of relational autonomy, extending them with an explicit focus on 'ask-points' — the discrete everyday moments when a PWU must request help (pick up an object, open a door, transfer, dress, shower, eat). The authors argue that cutting even one or two high-frequency ask-points delivers disproportionate autonomy gains. They synthesize findings into an ecosystem systems model (user agency ↔ caregiver ecology ↔ technology affordances ↔ structural context) and propose a Trajectory- and Ask-Point-Centered Co-Design Framework plus an 'Assistive-Tech Commons' — a user-led, community-governed directory that indexes solutions by task, aggregates DIY modifications, and packages trial evidence into insurer-ready dossiers.
Key findings
Ask-points concentrated in reaching/picking up objects (87%), food preparation (67%), personal hygiene (62%), opening doors (53%), dressing (40%), transfers (33%), and household chores (25%). Adoption clustered on inexpensive mainstream tools — voice assistants on phones (24%) and smart speakers (18%), reachers/grabbers (18%) — while high-impact advanced technology was rare: only 3 participants used a Kinova JACO robotic arm despite transformative reports ('before the JACO, it was very hard for me to be alone at any time. Now I can be left alone for several hours'). Dependency carried heavy affective weight: 47% reported privacy invasion, 36% feeling like a burden, 27% loneliness, 13% emotional safety concerns, 7% infantilization. Decision-making was polycentric — OTs/PTs consulted by 73%, family by 58%, peer PWU networks by 33%, social-media reviews by 20%, crowdfunding by 9% — and family pragmatism often overrode clinical recommendations. Five barrier clusters emerged: cost and insurance denial (42 participants; the JACO arm costs ~$60k with no insurance coverage), caregiver shortages (20), training and knowledge gaps (15), design and reliability issues (12), and policy/accessibility friction (10). Enablers were peer networks, responsive clinicians who co-designed rather than prescribed, and community-driven DIY modifications (3D-printed custom add-ons). The central conceptual finding: independence is not a binary measurable in 'unsupervised hours' but a fluid, contextually negotiated state that shifts across daily trajectories.
Relevance
For accessibility practitioners, policymakers, and AT designers, this paper reframes what 'success' means for assistive technology: not device capability in isolation but how well a device compresses high-frequency ask-points, fits the household and caregiving ecology, and preserves emotional dignity and privacy. Concrete takeaways: (1) measure impact via ask-point reduction and caregiver-minutes saved, evaluated under peak-dependence conditions (morning and evening routines) rather than routine scenarios; (2) treat emotional dignity — privacy relief, autonomy during intimate tasks, reduced infantilization — as a first-order design constraint, not a soft consideration; (3) build insurer-ready evidence packs alongside prototypes, because ~30% of devices are abandoned and cost/coverage is the #1 barrier; (4) integrate peer-network and DIY maker knowledge into procurement pathways via community-governed infrastructure like the proposed Assistive-Tech Commons. Limitations: U.S. sample only, powered-wheelchair users specifically (not manual wheelchair or other mobility-aid users), interview-based with recall bias, no longitudinal or cross-cultural comparison. The framework is aspirational rather than deployed; operationalizing the Commons remains future work.
Tags: powered wheelchair · assistive technology · participatory design · independence · caregiving · technology abandonment · disability justice · relational autonomy · qualitative research