Assistive technology: activity, context and the problem of abandonment

People judge assistive devices by what they can do: how fast they read, how far they reach, how loudly they speak. This sixth of ten units argues that capability explains little of whether the technology helps; what decides the outcome is the fit between device, person and setting. It is also a construct the field calls abandonment but has never defined to its own satisfaction.

Hold two questions apart. Does the technology work, and does the person use it, towards their own goals, in their own context? The literature rested on the second and keeps returning the first as the answer.

The device is not the unit of analysis

The field's founding move was to treat technology as one element in a system. Cook and Polgar (2015) set out the Human Activity Assistive Technology model, in which the device sits inside an activity performed by a person in a context. Fuhrer, Jutai, Scherer and DeRuyter (2003) gave the causal version. Outcomes result from interaction among a device type, its users and their environment. That chain runs through procurement and introductory use, then moderating factors, towards shorter-term outcomes and possibly longer-term use. They offered the framework so that device-specific models could be built, which the literature has largely declined.

Lenker and Paquet (2003) reviewed six conceptual models, among them Scherer's Matching Person and Technology model and the International Classification of Functioning, Disability and Health. None had been systematically validated against outcomes. Hersh and Johnson (2008) set out the device classes and show what the framework has to hold. Their electronic travel aid, a hand-held cane that senses obstacles with ultrasound and reports them through vibration, helps only in relation to the route, the surfaces and the traffic.

Abandonment: the first outcome the field measured

The empirical core here is a negative result. Phillips and Zhao (1993) surveyed 227 adults with disabilities on device selection, acquisition, performance and use. They found 29.3 per cent of devices completely abandoned, most often in the first year and after five years. The four significant factors were the user's opinion not being considered in selection, easy procurement, poor performance, and changed need. Two of the four are properties of the selection process, and easy procurement points against a simple theory of scarcity.

Riemer-Reiss (1999) framed discontinuance as a case of Rogers' diffusion of innovations, and found that relative advantage and consumer involvement each significantly predicted it. Jahan, Guitard and Jutai (2024) screened 1,029 articles and retained 27, finding non-use, abandonment, non-compliance and non-adherence used inconsistently, with only 10 of the 27 defining the term used. The central outcome is a construct in search of a definition.

The field modelled the match, then disputed the outcome

The response was to model the match. Scherer (2002) assembled the Matching Person and Technology approach, whose claim is that use follows from the relationship among person, technology and environment rather than from device specification, and whose forms assess the person and the setting before a device is chosen. Ripat and Booth (2005) tested the service side with 18 stakeholders in three focus groups of providers, funders and users, and found three themes: the user is a unique individual, a decision-making process exists, and devices and services are complex.

Scherer and Glueckauf (2005) then disputed the outcome itself, arguing that the impact on activities and participation settles value and that those influences had been studied too little. On that argument continued use is not the dependent variable; participation is, and a device that is used but changes nothing has met no goal.

Provision: where the constraint actually binds

Fit presupposes a device available to be fitted, and for most of the world it is not. Borg, Lindstrom and Larsson (2009) set out the responsibilities created by the Convention on the Rights of Persons with Disabilities, and recorded that in low- and middle-income countries only about 5 to 15 per cent of people who need assistive technology have access to it. The first global estimate, by the World Health Organisation and UNICEF (2022), puts need above 2.5 billion people, a figure set to rise with ageing populations, with nearly a billion of them denied access. de Witte, Steel, Gupta, Ramos and Roentgen (2018) generalised this into a provision framework. It lists affordable products, awareness, advice, eligibility and funding rules, training, follow-up and repair.

High-income countries run the same failure in milder form. Kaye (2000) analysed 10,480 households in the Current Population Survey: 23.9 per cent of people with disabilities had a home computer, against 51.7 per cent of others, and internet use was 9.9 per cent against 38.1 per cent. The gap is about whether the device was there, not about ability to operate it.

Use is social, and the partner is part of the system

Shinohara and Wobbrock (2011) interviewed 20 individuals about assistive technology use in social and professional settings, where devices sometimes marked their users as disabled, and two misperceptions pervaded use: that a device could functionally eliminate a disability, and that its user would be helpless without one. Functional access took priority over feeling self-conscious, so stigma is not a simple cause of abandonment.

Light and McNaughton (2014) pushed the argument into competence itself, revisiting Light's 1989 definition of communicative competence for people who require augmentative and alternative communication and re-examining the linguistic, operational, social and strategic domains, with the psychosocial and environmental factors later added. Competence becomes a property of the interaction, and the communication partner's skill is part of the intervention.

Contested question: is abandonment a failure of fit or a failure of measurement?

The literature disagrees about what abandonment means, and that decides how much the thesis can claim. One position holds that abandonment is a real, predictable failure of fit. Phillips and Zhao (1993) supply the base rate, its predictors dominated by selection and poor performance, and Riemer-Reiss (1999) confirms relative advantage and consumer involvement. Scherer (2002) and de Witte and colleagues (2018) treat the failure as a design target that better matching and provision can reduce.

The opposing position holds that the field is measuring an artefact of its own vocabulary. Jahan, Guitard and Jutai (2024) show the terms used interchangeably, with only 10 of 27 studies defining the term used, so a reported rate may describe a definition or a clinician's decision as readily as a user's experience. Lenker and Paquet (2003) add that the models giving the terms meaning have not been validated. A third position declines to treat non-use as failure: Scherer and Glueckauf (2005) locate value in activities and participation, and Shinohara and Wobbrock (2011) show users weighing functional access against how others read a device.

The two sides rest on evidence of different kinds. The fit position rests on two surveys with measured predictors and two frameworks for matching and provision. The measurement position rests on a review of 27 retained studies, which counts how the field defines its terms rather than measuring any outcome. The participation position rests on an argument about the right outcome and one interview study.

The strongest objection is that the unit's headline number does not survive its own review. If the 29.3 per cent describes a definition rather than a behaviour, the unit cannot state how common abandonment is, and its four predictors then explain a quantity it cannot fix. The fit thesis is clear only once a definition is settled first, which is the move the unit otherwise resists.

What would settle it is a design the literature lacks: one device class, with the operational definition used for non-use, the fit and provision measures recorded before supply, and the participation outcome measured afterwards. Jahan and colleagues (2024) supply the definition, Scherer and Glueckauf (2005) the outcome, and de Witte and colleagues (2018) the provision variables.

What to do while the dispute is open

While the dispute is open, record before supply which outcome is measured and how non-use is defined. Measure participation as well as continued use, and report the provision variables rather than assuming them, among them who pays and whether the instructions assume a reader at home in English.

Where the argument is heading

The named work points towards standardising the instrument before disputing the result: Jahan, Guitard and Jutai (2024) argue for operational definitions tied to conceptual models, de Witte and colleagues (2018) recommend an international standard for provision, and Scherer and Glueckauf (2005) place participation at the end of the chain such a standard would have to measure. The next unit moves to the software the technology must interoperate with, where the same question recurs in the accessibility tree.

The sources this unit rests on

• Phillips, B. and Zhao, H. (1993). Predictors of Assistive Technology Abandonment. Assistive Technology. doi:10.1080/10400435.1993.10132205 Surveyed 227 adults with disabilities, finding 29.3 per cent of devices completely abandoned with predictors led by user opinion in selection, easy procurement, poor performance and changed need. The unit's base rate sits on a measure the field later failed to standardise.
• Riemer-Reiss, M. L. (1999). Applying Rogers' diffusion of innovations theory to assistive technology discontinuance. Journal of Applied Rehabilitation Counseling. doi:10.1891/0047-2220.30.4.16 Applied Rogers' diffusion of innovations in a quantitative correlational study of discontinuance, in which relative advantage and consumer involvement both significantly predicted it. It shows involvement, not capability, carrying the signal.
• Kaye, H. S. (2000). Computer and Internet Use among People with Disabilities. Disability Statistics Report 13. https://eric.ed.gov/?id=ED439579 Analysed 10,480 households in the Current Population Survey, finding home computer access of 23.9 per cent for people with disabilities against 51.7 per cent for others and internet use of 9.9 against 38.1 per cent. The device is often absent before it can be mismatched.
• Scherer, M. J. (2002). Assistive technology: Matching device and consumer for successful rehabilitation. Washington: American Psychological Association. doi:10.1037/10420-000 Assembled the Matching Person and Technology approach and its assessment forms, which hold that use follows from the relationship among person, technology and environment rather than from device specification. It anchors the unit's fit claim.
• Fuhrer, M. J., Jutai, J. W., Scherer, M. J. and DeRuyter, F. (2003). A framework for the conceptual modelling of assistive technology device outcomes. Disability and Rehabilitation. doi:10.1080/09638280310001596207 Proposed a framework in which outcomes result from interaction among device type, user and environment, running through procurement and introductory use to shorter and longer-term outcomes. The field knows what the causal chain contains and has not filled it in.
• Lenker, J. A. and Paquet, V. L. (2003). A Review of Conceptual Models for Assistive Technology Outcomes Research and Practice. Assistive Technology. doi:10.1080/10400435.2003.10131885 Reviewed six conceptual models on outcome measures, predictive characteristics, validation and utility. None was systematically validated against outcomes, so the field's explanatory models remain proposals.
• Hersh, M. A. and Johnson, M. A. (eds) (2008). Assistive Technology for Visually Impaired and Blind People. London: Springer. doi:10.1007/978-1-84628-867-8 Sets out the engineering and design of assistive technology for blind and vision-impaired people by device class, including electronic travel aids, screen readers and speech, text and Braille conversion. It shows capability as one term in a fit that also depends on activity and setting.
• Ripat, J. and Booth, A. (2005). Characteristics of assistive technology service delivery models: stakeholder perspectives and preferences. Disability and Rehabilitation. doi:10.1080/09638280500264535 Ran three focus groups with 18 stakeholders, namely providers, funders and users. Three themes emerged: the user as a unique individual, the existence of a decision-making process, and the complexity of devices and services.
• Scherer, M. J. and Glueckauf, R. (2005). Assessing the benefits of assistive technologies for activities and participation. Rehabilitation Psychology. doi:10.1037/0090-5550.50.2.132 Argued that the value of assistive technology is determined by its impact on activities and participation, and that those influences are insufficiently studied. It reframes low use as a question about goals rather than compliance.
• Borg, J., Lindstrom, A. and Larsson, S. (2009). Assistive technology in developing countries: national and international responsibilities to implement the Convention on the Rights of Persons with Disabilities. The Lancet. doi:10.1016/s0140-6736(09)61872-9 Recorded that only about 5 to 15 per cent of people in low- and middle-income countries who need assistive technology have access to it. Provision, not fit, is the binding constraint for most of the world.
• World Health Organisation and United Nations Children's Fund (2022). Global Report on Assistive Technology. Geneva: World Health Organisation. https://www.who.int/publications/i/item/9789240049451 Reports the first global estimate of need for and access to assistive products, with more than 2.5 billion people needing one and nearly a billion denied access. It supersedes the unit's older access figure.
• Shinohara, K. and Wobbrock, J. O. (2011). In the shadow of misperception: assistive technology use and social interactions. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. doi:10.1145/1978942.1979044 Interviewed 20 individuals on use in social and professional settings; devices sometimes marked users as disabled, and functional access took priority over self-consciousness. Social context bears on use without making stigma a simple cause of abandonment.
• Light, J. and McNaughton, D. (2014). Communicative Competence for Individuals who require Augmentative and Alternative Communication. Augmentative and Alternative Communication. doi:10.3109/07434618.2014.885080 Revisited Light's 1989 definition of communicative competence for people who require augmentative and alternative communication and the four domains it rests on. It treats competence as a property of the interaction, not of the device.
• Cook, A. M. and Polgar, J. M. (2015). Principles of assistive technology. In Assistive Technologies: Principles and Practice. St. Louis: Elsevier. doi:10.1016/b978-0-323-09631-7.00001-6 Sets out the Human Activity Assistive Technology model, placing the device inside an activity performed by a person in a context. It is the frame this unit uses to argue that the device is not the unit of analysis.
• de Witte, L., Steel, E., Gupta, S., Ramos, V. D. and Roentgen, U. (2018). Assistive technology provision: towards an international framework for assuring availability and accessibility of affordable high-quality assistive technology. Disability and Rehabilitation: Assistive Technology. doi:10.1080/17483107.2018.1470264 Set out the elements of an international framework for provision, from affordable products and awareness to advice, eligibility, funding, training, follow-up and repair, and recommended an international standard. It moves the analysis from device to delivery system.
• Jahan, A. M., Guitard, P. and Jutai, J. W. (2024). Assistive devices non-use, abandonment, or non-adherence? Toward standard terminology for assistive devices outcomes. Assistive Technology. doi:10.1080/10400435.2024.2362139 Screened 1,029 articles on failure to use prescribed assistive devices and retained 27; the terms were used inconsistently and only 10 defined the term used. The central outcome is a construct the field has not standardised.