Freedom to Choose: Understanding Input Modality Preferences of People with Upper-body Motor Impairments for Activities of Daily Living
Franklin Mingzhe Li, Michael Xieyang Liu, Yang Zhang, Patrick Carrington · 2022 · Proceedings of the 24th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '22) · doi:10.1145/3517428.3544814
Summary
This paper investigates how people with upper-body motor impairments currently perform Activities of Daily Living (ADLs) and what input modalities they would prefer for technology-assisted independence. ADLs encompass basic tasks like bathing, dressing, grooming, toileting, and eating, as well as instrumental ADLs (IADLs) like shopping, cooking, managing finances, and driving. The researchers conducted semi-structured interviews with 12 participants who had various upper-body motor impairments including spinal cord injuries (4), cerebral palsy (4), muscular dystrophy (1), primary lateral sclerosis (1), stroke (1), and arthrogryposis multiplex congenita (1). The study presented participants with nine categories of input modalities documented in prior research — touch, voice, eye-based, head-movement, face/mouth-based, hand/arm gestures, brain-computer interface, biometric, and automatic recognition — and asked about preferences for each across 18 ADL categories. This comprehensive approach examined not just computing tasks but the full spectrum of daily living activities, many of which still require personal care assistants (PCAs) or family member help. The study addressed three research questions: current practices and challenges, how various input modalities could benefit people during ADLs, and how multimodal input could be leveraged to support ADL performance.
Key findings
A striking disparity emerged between computing-based ADLs and physical ADLs: only 10% of participants needed help with phone use or managing finances, but over 90% required PCA assistance for toileting, dressing, and cooking. About 67.2% of participants required PCAs for each ADL on average. Touch and voice were the most preferred individual input modalities across all tasks, with touch preferred for bathing (10 participants) and voice for cooking (10) and driving (10). Nearly all participants (11/12) wanted to change their existing input methods for ADLs. The study identified five key factors influencing input modality choice: usability (reliability, precision, learnability), efficiency (response time, physical effort, simplicity), consequences (mess prevention, safety, confidentiality), personalization (physical ability, familiarity, independence), and context (environmental noise, reachability, social acceptability). For multimodal input, participants valued it for two distinct purposes: redundancy (having backup methods for reliability, accommodating progressive conditions like muscular dystrophy, and coping with inaccessible environments) and variability (using different inputs for sub-tasks within a larger task to enable natural mappings, prevent false activation, and add input dimensions). Participants also highlighted that multimodal input could enhance self-confidence in social interactions and expressed desire to contribute to household tasks when family members are busy.
Relevance
This research has significant implications for smart home design, assistive technology development, and IoT accessibility. The finding that existing input systems mostly lock users into fixed single modalities — and that these mappings are nearly impossible for users to modify themselves — points to a fundamental design failure in current assistive technology. For practitioners designing accessible smart home systems, the study provides a detailed task-by-task breakdown of which input modalities users prefer and why, offering a practical design reference. The emphasis on multimodal input as both redundancy and variability challenges the common approach of designing single-input assistive solutions. The paper also highlights often-overlooked dimensions of ADL technology design: social acceptability (voice control in public is embarrassing), safety (people with spinal cord injuries cannot feel burns from hot water), privacy concerns during intimate care tasks, and the financial burden of PCA reliance. The discussion of collaborative ADL experiences — where technology and PCAs work together rather than technology fully replacing human assistance — offers a more nuanced and realistic vision for assistive technology than full automation. The recommendation to categorize ADLs by social context (private/public, privacy sensitivity) when designing multimodal systems is particularly actionable.
Tags: motor impairment · multimodal input · activities of daily living · assistive technology · personal care assistants · independent living · input methods · spinal cord injury · cerebral palsy · smart home