on modelling assistive technology systems – part i

Technology and Disability 20 (2008) 193–215 193IOS Press

On modelling assistive technology systems –Part I: Modelling framework

Marion A. Hersha,∗ and Michael A. JohnsonbaDepartment of Electronics and Electrical Engineering, University of Glasgow, Glasgow, UKbDepartment of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK

Abstract. There has been increasing recognition of the importance and benefits to society of social inclusion and the fullparticipation of disabled people. Many countries have also enacted legislation aimed at removing discrimination against disabledpeople. Removing barriers to full participation by disabled people will require the development of new assistive technologysystems and improved information and distribution systems for existing assistive technologies.This will require an effective and ongoing dialogue between the disabled end-user community, social services, the clinicalrehabilitation services, and the professional engineering disciplines involved in the development, provision, assessment, andongoing support for assistive technology. To support this dialogue there is a need for common terminology, concepts, anddefinitions, embedded within a single, unified model framework. This paper presents the Comprehensive Assistive Technology(CAT) model, which was designed to meet this need.

Keywords: Assistive technology, design for all, modelling framework, technology analysis, end-user studies

1. Introduction

The work reported in this paper arose from a needto find a succinct yet comprehensive framework to de-scribe and analyse the engineering design and the useof assistive technology systems and applications. Oneof the main motivations was the need, arising out ofthe authors’ work on new communication devices fordeafblind people [18,19], for a well-defined frameworkto investigate and formulate end-user and engineeringrequirements and to structure the development of newassistive devices. The second motivation was pedagog-ical, arising from the need to provide an appropriatestructure for the presentation of educational and archivematerial to students and researchers [16,17]. However,it soon became apparent that the methods and modelsfor assistive technology definition and analysis are notwell developed. There are a number of different defini-tions of assistive technology, with the differences deter-

∗Corresponding author. Fax: +44 141 330 6004; E-mail: [email protected].

mined, at least in part, by the model of disability used.There are two main approaches to understanding dis-ability: the medical and social models. The approachtaken here to modelling assistive technology is basedon the social model of disability. The background to,and the implications of the social model of disabilityand the definitions of assistive technologyare presentedin Section 2 of the paper.

Reviews of the concepts and frameworks that supportmodelling of assistive technology are notably scarce.To fill this gap, Section 3 opens with an overview ofassistive technology modelling approaches, includingthe objectives and validity of the different models. Thisreview is followed by three detailed subsections thatexamine three important modelling methods. Section 3closes with a discussion of the requirements for a newmodelling framework that draws together and extendsthe best features of these existing approaches.

The main contribution of this paper is the presenta-tion of a new modelling framework that is given in Sec-tion 4. It is claimed that the new Comprehensive Assis-tive Technology (CAT) model provides a simple, effec-tive, and unified modelling framework to support the

ISSN 1055-4181/08/$17.00 2008 – IOS Press and the authors. All rights reserved

194 M.A. Hersh and M.A. Johnson / On modelling assistive technology systems – Part I: Modelling framework

ongoing dialogue that occurs between end-users, med-ical practitioners, clinical rehabilitation professionals,social support staff, and the engineering community.This unified modelling framework and terminology willalso enable the common structure of assistive technol-ogy systems to be understood by different professionaldomains and provide a pedagogical basis for educa-tional and training courses in the assistive technologyfield. Conclusions are presented in Section 5.

2. Social model of disability and its implications

This section presents a brief discussion of the socialmodel of disability and its implications for modellingassistive technology. Section 2.2 considers definitionsof the domain of assistive technology.

2.1. Social model of disability

Since assistive technology is generally provided fordisabled people, obtaining an understanding of disabil-ity is the first stage in setting up a modelling frameworkfor assistive technology, with different understandingsof disability resulting in different approaches to mod-elling assistive technology. There are two main ap-proaches: the medical model and the social model.

The medical model originates with the World HealthOrganisation (WHO) and therefore is the perspective ofprofessionals who are largely medical and other healthcare practitioners, whereas the social model originateswith an organisation of disabled people, the Union ofPhysically Impaired Against Segregation (UPIAS) andtherefore gives the perspective of disabled activists andcampaigners. The original version of the medical mod-el (sometimes referred to as the ICIDH model) us-es the classification of “impairment”, “disability” and“handicap” developed by the World Health Organisa-tion (WHO) in 1980 [25]. The WHO defined “impair-ment” as “any loss or abnormality of psychological,physical, or anatomical structure or function”. A “dis-ability” then occurs when the impairment prevents aperson from being able to “perform an activity in themanner or within the range considered normal for a hu-man being”. Hence, a “handicap” results when the per-son with a disability is unable to fulfil their normal rolein society and the community at large. Thus, the medi-cal model views disability as residing in the individualand focuses on the person’s impairment(s) as the causeof disadvantage leading to the approaches of occupa-tional therapy and clinical rehabilitation. The updated

version, the International Classification of FunctioningDisability and Health (ICF) [27] considers disablementto result from the interaction between an individual’shealth and contextual factors. However, it is still theindividual’s condition rather than external factors thatis the main driver of the classification.

The social model was first developed by the Union ofthe Physically Impaired Against Segregation [24] andthen modified by the Disabled Peoples International(DPI) [3]. The model is based on the two concepts ofimpairment and disability. “Impairment” is defined asthe functional limitation caused by physical, sensory,or mental impairments. “Disability” is then definedas the loss or reduction of opportunities to take partin the normal life of the community on an equal levelwith others due to physical, environmental, or socialbarriers.

The social model of disability thus emphasizes thephysical and social barriers experienced by disabledpeople [28] rather than their impairments and considersthe problem to be in society rather than with the dis-abled person. The focus is therefore on social barriersrather than the individual. The social model [29]:

– Recognises the creation of disability through theinteraction of structural and attitudinal variablesand the political processes that oppress disabledpeople and deny them civil rights.

– Recognises the voice and opinions of disabled peo-ple and aims to increase the control that disabledpeople and their organisations have over powerand information.

In terms of the social model, engineers and designershave the following two areas of responsibility:

(1) Design for all; that is designing and constructingdevices and environments to be accessible anduseable by as wide a range of the population aspossible, including disabled people.

(2) Design of assistive technology systems, for ex-ample, the design of devices to overcome ex-isting environmental and social barriers therebyextending the opportunities and options open todisabled people.

2.2. Spectrum of assistive technology activities

Although the terms ‘assistive technology’ and ‘reha-bilitation engineering’ are frequently used in the litera-ture and there is a generalised understanding of both ofthem, there has been little coordinated work on defini-tions. The medical model is better known and has had a

M.A. Hersh and M.A. Johnson / On modelling assistive technology systems – Part I: Modelling framework 195

greater influence on practitioners than the social model,probably due to the much greater size and resources ofWHO compared to the DPI.

The term ‘rehabilitation’ refers to the restoration ‘tonormal life by training and therapy’ of someone whois or has been ill, disabled (or in prison). Rehabilita-tion technology can then be considered as the technol-ogy, devices, processes, systems, or services requiredto support this process. One of the commonly useddefinitions of rehabilitation technology expresses thisconcept as follows [9]:

(1) Any tool for remediation or rehabilitation ratherthan being part of the person’s daily life andfunctional activities,

(2) Those technologies associated with acute carerehabilitation process, and,

(3) The segment of assistive technology that is de-signed specifically to rehabilitate an individualfrom their present set of limitations due to somedisabling condition, permanent or otherwise.

Thus rehabilitation technology is used to act on andmodify a disabled person in order to overcome their‘limitations’ rather than to modify their environmentto overcome barriers they might experience. Thus, theunderlying philosophy of rehabilitation technology isbased on the medical model of disability. This is notthe case for assistive technology, the underlying phi-losophy of which can be based on the medical or socialmodels, a combination of them, or even a totally differ-ent approach. Although the term ‘assistive technology’is frequently used and generally understood, there isno universally accepted definition. In the UK, [30] aconsultation meeting of the King’s Fund in March 2001agreed the definition ‘Assistive Technology (AT) is anyproduct or service designed to enable independence fordisabled and older people’ to replace the previouslyused term ‘disability equipment’. This definition ac-knowledges the overlap between assistive technologydesigned specifically for older people and mainstreamtechnology that has been made available to disabledpeople through design for all, but does not provide ameans of distinguishing between them. It has the ad-vantage of being very broad-based, but the disadvan-tage of lacking specificity and therefore not makingclear what is and is not included. However, anotherslightly later UK definition due to the Audit Commis-sion considers ‘assistive technology’ to be ‘any item,piece of equipment, product or system that is used to in-crease, maintain or improve the functional capabilitiesand independence of people with cognitive, physical orcommunication difficulties’ [31].

In the USA, the Technology-Related Assistance forIndividuals with Disabilities Act 1988, the AssistiveTechnology Act 1998 and the Access Board’s Elec-tronics and Information and Technology AccessibilityStandards all contain the following formal legal defi-nition of ‘assistive or adaptive technology’: ‘products,devices or equipment, whether acquired commercial-ly, modified or customized, that are used to maintain,increase or improve the functional capabilities of indi-viduals with disabilities’ [32]. The terminology of thisdefinition has clearly been influenced by the medicalmodel of disability. The focus is on rehabilitation andspecifying the assistive technologies that are eligiblefor funding rather than a wider range of applications.

In the European Union, ‘Assistive Technology refersto products, devices or equipment that are used to main-tain, increase or improve the functional capabilities ofpeople with disabilities. Assistive Technology can helpto compensate functional limitations and enable peoplewith disabilities to participate in the activities of dailylife, including employment and training.’ [33]

Although assistive technology is sometimes still de-fined in terms of a specific list of technologies, devices,or equipment related to particular applications areas,there is increasing recognition that a broad-based ap-proach to definition is preferable. The derivation of adefinition for assistive technology raises a number ofissues, including the following:

– The purpose for which the particular definition isintended. Thus the US definition focuses on thetypes of ‘assistive technology’ that provide ‘reha-bilitation’ and are eligible for funding, rather than,for instance, examining how ‘assistive technolo-gy’ can be used to overcomethe infrastructural andother barriers that disabled people currently expe-rience. It is clearly desirable that funding shouldbe available for this much wider range of ‘assistivetechnology’.

– Whether components of a larger system, device, orservice, such as audio announcements and Braillemarkings in lifts or vibrating indicators at trafficlights, should be considered assistive technologyor design for all applications.

– Whether it is useful to make distinctions betweenassistive technology and design for all applica-tions.

– Whether defining assistive technology as beingused solely by disabled (and elderly) people couldlead to discrimination and therefore a more gener-ic approach is required. In this context the follow-ing definition has been proposed [34]: ‘Assistive


technology consists of a procedure(s), process(es),equipment(s), material, activity(ies) or system(s)that allows an individual or a group of people toperform a task that they would be otherwise beunable to perform or increases the ease and safetywith which the task can be performed.’ However,because definitions of this type do not refer to dis-abled and elderly people, they are too all embrac-ing to be particularly useful. In particular, theycover almost all equipment, procedures and pro-cesses, most of which are used to enable people tocarry out activities either more easily and safelyor which otherwise would be impossible. Thus,unless there is a classification that, for instance in-volves overcoming the barriers that would be oth-erwise be experienced by disabled and/or elderlypeople in carrying out activities easily and safe-ly (or at all), tools such as ordinary knitting nee-dles or a knitting machine would be covered. It isnot possible to knit a jumper without using eitherknitting needles or a knitting machine. Howev-er, classifying them as assistive technology seemsproblematical. This is also contrary to the ap-proach taken by the journal of the Association forthe Advancement of Assistive Technology in Eu-rope (AAATE), Disability and Society. Althoughit does not have a definition of assistive technol-ogy, it considers the end users of ‘assistive tech-nology devices and services’ to be persons withdisabilities and their family members’ [35].

Examination of some of the definitions of assistivetechnology [9,30–35] shows that it has the followingfeatures:

– It is a generic or umbrella term.– It covers technologies, equipment, devices, appa-

ratus, services, systems, and processes used by dis-abled or elderly people to increase their indepen-dence and participation in society and/or enablethem to carry out activities that would be difficult,dangerous, or impossible otherwise.

– Many of the definitions seem to have been influ-enced by the medical model of disability and there-fore talk about assistive technology being requiredto overcome limitations due to impairments and toincrease, maintain or improve functional capaci-ties and ameliorate the problems faced by disabledpeople. However, there is no reason why defini-tions should not be formulated in terms of assistivetechnology being required to overcome the social,infrastructural, and other barriers that reduce the

independence of disabled people and their abilityto carry out activities and participate in society.

– Many of the definitions of assistive technolo-gy exclude environmental modifications, such asdropped curbs. However, there could be benefitsin including environmental modifications in defi-nitions of assistive technology.

– There is not currently a clear definition of the dis-tinction between mainstream products designedusing a design for all approach and assistive tech-nology. For instance, an AAATE position pa-per [36] considers that assistive technology anddesign for all should be considered as ‘componentsof a continuum whose primary aim is to exploit theindividual’s capabilities at their fullest’ and thatthey ‘should be looked at as part of the same do-main of knowledge, rather than. . . two distinctdomains’. However, there is still a need for a dis-tinction between assistive technology and designfor all approaches, while recognising that as assis-tive technology is more widely adopted it may be-come part of a design for all approach. Therefore,for instance, dropped curbs are now widespread inmany countries and should therefore be considered‘design for all’ rather than ‘assistive technology’.

This discussion has led the authors to propose thefollowing definition:

‘Assistive technology is a generic or umbrella termthat covers technologies, equipment, devices, ap-paratus, services, systems, processes and environ-mental modifications used by disabled and/or el-derly people to overcome the social, infrastructuraland other barriers to independence, full participa-tion in society and carrying out activities safely andeasily.’

The specifications of use by disabled and/or elderlypeople and use to overcome different types of barriersare required to distinguish between technology used bydisabled people as assistive technologyand mainstreamtechnology used by disabled people, as well as betweenassistive technology and design for all. While it is use-ful to distinguish what is currently considered assistivetechnology from design for all technology, it shouldbe recognised that they comprise the two ends of thesame continuum and that over time what is currentlytermed assistive technology may become a design forall component of a mainstream product.

However, it is also useful to distinguish the assistivetechnology end of the continuum by specifying that itis used (largely) by disabled and/or elderly people to


Biomedical Engineering

Rehabilitation Engineering

ASSISTIVE TECHNOLOGY

Accessibility in all societal interactions

The Wider Environment

The

Com

munit

y

The H

om

e

Person

Fig. 1. The domain of assistive technology.

overcome barriers. Whereas design for all technologyis used by the wider population and the design is suchthat there are no longer any barriers to the activity ofinterest rather than a specific technology being requiredto overcome them.

Assistive technology (see Fig. 1) is used in a so-cial, cultural, political, economic, and environmentalcontext. This context may facilitate the developmentand use of assistive technology, pose barriers and con-straints, or be neutral. Although the infrastructure forthe provision of assistive technology is more developedin the industrialised countries, assistive technology isrequired in the so-called developing countries as welland should be available worldwide. Users and poten-tial users of assistive technology also vary greatly intheir characteristics, interests, skills, values, and im-pairments. In addition, assistive technology is requiredfor a wide range of different types of tasks and applica-tions. A model for assistive technology should attemptto capture all these aspects of the domain.

3. The modelling response

3.1. Overview

A review of the literature reveals that only a lim-ited number of researchers have studied the develop-ment of assistive technology and that there are threemain approaches to disability and assistive technologymodelling, as listed below.

(1) Classification methodologies(2) System modelling methods(3) Assistive technology outcomes modelling

There are one or two key developments in each ofthese methodological areas, making a modelling re-view particularly straightforward. This situation is cap-tured in Fig. 2, which shows the three methodologicalbranches and a small number of named methods in eachbranch.

Classification approaches

Classification approaches are taxonomic systems fordefining the domain categories in the disability and/orassistive technology fields rather than being modellingmethodsper se. There are currently two main clas-sification systems available: the International Classi-fication of Functioning, Disability and Health, denot-ed ICF, and the International Standards Organisationstandard ISO 9999: 2002 Technical Aids for Personswith Disabilities, Classification and Terminology. Bothclassification systems have had some influence on stan-dardising terminology definitions, data collection re-quirements and setting standards in the field.

The ICF is one of a group of international clas-sifications developed by the World Health Organisa-tion and has been accepted as a United Nations so-cial classification, making it relevant to consider theimpact of this classification system on modelling as-sistive technology. The ICF is reviewed in Section3.2.1. ISO 9999: 2002 is product and application ori-entated, but does not give a basis for understandingthe engineering structure and societal application ofassistive technology systems. The issue of classifica-tion is also relevant to the development of searchabledatabases of assistive technology products, devices,and services, including the Disabled Living Foun-dation (http://www.dlf-data.org.uk/)and ABLEDATA(www.abledata.com) databases.

System modelling approachesThe assistive technologyoutcomes and service deliv-

ery modelling approach has received greater attentionthan assistive technology system modelling, thoughfurther work is still needed in this area. The main orindeed sole significant model in the system modellingcategory is the Human Activities Assistive Technology(HAAT) model due to Cook and Hussey [6], which isdiscussed in Section 3.2.2.

Assistive technology outcomes modelling approach-es

One of the reasons for the better development ofthe literature on outcomes modelling is probably theconnections to the field of health related quality of lifeassessment, which has over a thousand citations each


Classification Approaches

Systems Modelling

ASSISTIVE TECHNOLOGY

WHO- ICF -2 ISO Standards

Instruments and databases

Outcomes Modelling

HAAT Model

Service Delivery Models

MPT Framework Quality of Life Indices

Fig. 2. Approaches to modelling the assistive technology domain.

year [12]. In addition, quality of life measurements areoften used in clinical trials [4,5]. However, a definitivequality of life definition has not yet been obtained. Forinstance, a review of 87 studies from diverse literaturefound 44 different definitions [20]. In addition, thequality of life approaches often have limitations anda tendency to focus on medical issues, which is notsurprising since medicine is the domain of origin ofmany of these approaches.

Quality of life approaches are but one tool in themuch broader task of modelling assistive technolo-gy outcomes and service delivery processes. Thereis a basic model framework established by Cook andHussey [6] and amongst others the Matching Person toTechnology framework due to Fuhreret al. [13]. Al-though these models do not model assistive technologyper se, they are reviewed in Section 3.4.3, due to theneed of a modelling framework for assistive technologywhich is able to combine system and outcomes mod-elling. Section 3.2 presents a detailed model review,with one subsection devoted to each of the three mainmodel areas presented in Fig. 2.

3.2. Model review

The model review considers the three main mod-elling approaches and presents a brief discussion of themain methods and their uses for each area. The sectionconcludes with a discussion and comparative analysisof the main methods in the three different areas.

3.2.1. World Health Organisation: InternationalClassification of Functioning, Disability andHealth(ICF)

The first version of ICF was published in 1980, andthe revised version was approved for publication inMay, 2001 [27]. It is considered to be complementaryto the WHO’s International Classification of DiseasesTenth Revision (ICD -10) [26]. Its aims are to providea scientific basis and common language to describe

health and health-related states, to facilitate the com-parison of temporal data across countries, health caredisciplines and services, and to provide a systematiccoding scheme for health information systems.

As shown in Fig. 3, the ICF classification is a hierar-chical scheme based on a two part taxonomic structure,with each part having two components. The classifica-tion continues through the definition of constructs bythe use of classifiers and then by domains and categoriesat different levels.

Part 1: Functioning and Disability: this containsumbrella terms for body functions, structures, activitiesand participation; and impairments, activity limitationsand participation restrictions respectively:

(a) Body functionsand Structures: respectivelyphysiological and psychological functions; andanatomical parts of the body i.e. organs, limbsand their components . The factors at the con-structs/qualifiers level are changes in body func-tions and structures

(b) Activities and Participation: respectively ‘theexecution of a task or action by an individual’and ‘involvement in a life situation’. However,the difference is ambiguous. There are nine clas-sifications covering activities, such as commu-nication and mobility and participation as in re-lationships and community, social and civic life.The two classifiers are performance and capac-ity or the highest probable level of functioningrelative to the statistical ‘norm’ for humans.

Part 2: Contextual Factors

(a) Environmental Factors: contains five classifi-cations which cover all aspects of the externalor extrinsic world that form the context of anindividual’s life, and as such, have an impacton that person’s functioning. This is illustrat-ed in Fig. 4 in tree diagram format. At theconstructs/qualifiers level, facilitators and bar-riers are the factors in a person’s environment


Part 1 Functioning and Disability

Changes in body function

Part 2 Contextual Factors

Body Functions

and Structures

Activities and

Participation

Environmental

Factors

Changes in body structure

Facilitator - Barrier

ICF

Performance Capacity

Personal

Factors

Item Levels

Item Levels

Item Levels

Item Levels

Item Levels

Fig. 3. Overall structure of the ICF (after [27]).

Table 1Example of ICF code structure: ‘d4503.3’

Part Part 1 Functioning and disabilityComponent D Activities and Participation1st Level/ Chapter Level d4 Chapter 4 Mobility2nd Level d450 Walking3rd Level d4503 Walking around obstaclesQualifier (.Performance, d4503.3 Severe difficulty without anCapacity) assistive technology to help.

that through their absence or presence respec-tively improve functioning and reduce disabilityor limit functioning and create disability.

(b) Personal Factors: contextual factors that relateto the individual such as age, gender, social sta-tus, but there is only the single top-level headingand no further classification is given.

However, the focus on ‘capacity limitations’ and‘performance problems’ due to ‘change in body struc-tures’ or ‘change in body function’ indicates a contin-uing medical model focus on the problems resultingfrom the individual’s impairments rather than a socialmodel focus on social, infrastructural and other barri-ers. In addition, the sole mention of assistive technol-ogy is in the Products and Technology classification ofthe Environmental factors component in which a dis-tinction is made between general technology productsand assistive technologyper se(see Fig. 4.) Further,the assistive technology areas are covered by just sixsub-classes.

A complex alphanumeric coding system is used torepresent the classification. This is illustrated by thedecomposition of a particular code shown in Table 1,with ‘d’ representing activities and participation. Fur-ther complexity results from the potential for using theActivities and Participation’ list in four different wayswith associated coding options [27].

It has been suggested [27] that the ICF can be used asa statistical and research tool for collecting and archiv-

ing health related data and quality of life data, as a clin-ical tool for assessing needs and evaluating outcomes,as well as a social policy tool to support the benefits andcompensation systems. However the potential useful-ness of the suggested educational social awareness andadvocacy applications would be significantly increasedif the model had an explicit way of including the per-spectives of disabled people. It should also be notedthat such perspectives are frequently better describedin qualitative rather than quantitative terms and that theICF approach is better suited to capturing and encodingquantitative rather than qualitative data.

3.2.2. Human Activities Assistive Technology(HAAT)model

The human activities assistive technology model isdescribed by Cook and Hussey in their book AssistiveTechnologies: Principles and Practice [6]. The modelis given pictorial representation as shown in Fig. 5.

The HAAT model is a re-statement of a well-knownhuman performance framework used by human factorsengineers and psychologists to study the operationalbehaviour and performance of humans doing techno-logical tasks. The original human performance modelcomprised the three components of “human, activity,and context” [2]. Cook and Hussey added the assistivetechnology component as shown in Fig. 5, giving thefollowing framework:

– Context, which defines the social framework andthe physical environment in which the person andthe assistive technology have to operate.

– Human personrepresents the person at the centreof the HAAT model, who is considered to havethe attributes of sensory inputs, central processingand effectors (motor outputs).


In Part 2:

Environmental

Factors

Products and technology

Natual and human-made environments

Support and relationships

Attitudes

Services, systems and policies

Fig. 4. Tree diagram for the ICF component, Environmental Factors.

Context

HUMAN PERSON

Activity Assistive Technology System

Fig. 5. HAAT model of an assistive technology system (after Cook and Hussey [6]).

– Activity defines the procedure, operation or taskthat the person would like to achieve. The activitycomponent is one of the more flexible terms withinthe model and depends on the use of the model.

– Assistive technologydefines the external enabler(device) used to overcome any contextual barrieror obstacle.

This approach is one of the very few attempts topresent a general systems structure for the technologyof the assistive system. For this, Cook and Hussey useblock diagrams to capture the general input-processor-output nature of the assistive technology system.

Cook and Hussey [6] have used the HAAT modelas a general introductory conceptual framework for theperformance of a person-plus-assistive technology sys-tem. This allowed them to discuss the general contentof the four components in the model and provided anumber of technical labels for technology componentsin the assistive technology system. In the USA, themodel has been influential in providing a pedagogicalframework for assistive technology studies. However,there has been very little work developing new exten-sions of the HAAT model approach, one of these be-ing where the HAAT model has been embedded in a

human-centred design procedure for the developmentof an application using a robot [14].

3.2.3. Assistive Technology outcomes modellingThis approach covers the whole process from deter-

mining what assistive technologies and services end-users could benefit from through to the assessment ofend-user requirements and characteristics, followed byassistive technology device procurement, and conclud-ing with (evaluation of) the longer-term outcomes forthe end-user. The methodologies in this area providea framework for modelling and measuring the perfor-mance of the assessment service, the supply service, theassistive technologyper seand ultimately, the end-usersatisfaction with the assistive technology provision.

Probably the best known and most commonly usedmodelling framework of this type is the Matching Per-sons and Technology (MPT) Model [13,23], which isdivided into the three main components of the personusing the technology, the technology and the milieu orenvironment. However, the model is very closely tiedto the associated six-step assessment procedure for usein determining outcomes and the appropriate assistivetechnology for a particular person in a given environ-


Introductory

Use

Shorter-Term

Outcomes:Effectiveness

Efficiency

Device

satisfaction

Psychological

functioning

Subjective

well-being

Longer

Term

Use

Procurement

of a

Device

Longer-Term

Outcomes:Effectiveness

Efficiency

Device

satisfaction

Psychological

functioning

Subjective

well-being

Continued

use of

device

Discontinued use of device

Fig. 6. Conceptual framework for AT outcomes assessment, based on MPT (after Fuhrer et al., [13].

Recommend

and

Report

Initial

Evaluation:Needs identified

Skills evaluated

sensory

physical

cognitive

language

Device

characteristics

Implement:

Order and

setup

Delivery and

fitting

Training

Referral

and

Intake

Follow-up:

Maintenance

Repair as

needed

Follow along:

Re-evaluate

Maintenance

Repair as

needed

Fig. 7. Assistive technology service delivery model (after Cook and Hussey [6].

ment [23], rather than having a detailed classificationstructure. There are two versions of the assessmentprocedure for adults, developed in Ireland and the USA,with the US version translated into French and Italianand one for children under 15, known as the Match-ing Assistive Technology and Child (MATCH) version.The assessment procedure requires the service providerand end-user to complete slightly different versions ofa number of forms, followed by discussion of outcomesand action. The approach is influenced by the medi-cal model of disability and aims to determine ‘limita-tions’ on functioning and identify goals and technolo-gies that could be used to improve functioning, as wellas characteristics of the person, environment or tech-nology that could lead to inappropriate use or abandon-ment of these technologies. In addition, the model in-cludes some personal characteristics, experiences andattitudes to technologies and degree of satisfaction withdifferent aspects of life.

Figure 6 shows the schematic of a conceptual frame-work to be used in assistive technology outcomes as-sessment [13]. It is based on an adaptation of the ap-proach frequently used in the health sciences and mostof the factors considered to influence the outcome ofdevice and service provision are obtained from the MPTmodel. There are three key stages: device procure-ment, introductory use, and longer-term use and twodecision points in the model, indicated by a diamondrepresenting the flow chart ‘OR’, where the end-usermay continue to use the device or reject it. Each of

these three stages is a complex process that will supportdetailed models incorporating activities such as assess-ment, training, and outcomes analysis. Thus the modelrecognises that there are time-dependent elements inassistive technology use, including feedback from pre-vious use, changes in circumstances, increasing skill indevice use and emotional factors.

A similar framework for assistive technology servicedelivery has been obtained from the HAAT model [6]and is shown in Fig. 7. It has the same three main stagesas the conceptual framework. However, the samplequestionnaire [6] to support these process stages hasa very strong rehabilitation emphasis and there is lim-ited investigation of the user’s requirements for moregeneral assistive technology support. There are only avery few examples of the use of the HAAT model inassessment procedures, one of which is due to Fay [11].

Other instruments that have been developed for mea-suring assistive technology device outcomes includethe Psychosocial Impact of Assistive Devices Scale (PI-ADS) [7,21] and the Quebec Users Evaluation of Sat-isfaction with Assistive Technology (QUEST 2.0) [8].

3.2.4. Discussion and comparative analysis of theICF, HAAT and MPT models

A model is generally designed to be used for a par-ticular (group of) application(s) and is influenced by aparticular philosophy, which originates with the mod-eller and/or the context in which the model has beenderived. Although it is possible to design a model for


multiple or even multi-dimensional applications, thereare generally tradeoffs between the range of applica-tions and the degree of complexity of the model. Whilemodels that cover all possible applications are general-ly so complex as to be of little practical value, carefulmodel design to cover a wider range of applicationsthan those considered each by the ICF, MPT and HAATmodels on its own is possible.

In terms of their intended applications and underly-ing philosophies, the ICF, MPT and HAAT models areall good models. The drawbacks of these models withregards to modelling assistive technology systems re-late to the influence of the medical model of disability,whether implicitly or explicitly, and the much widerrange of modelling applications required than are cov-ered by these models, either singly or collectively.

Human centred design approaches are particularlysuited to assistive technology, due to the importance offocussing on what the end-user wants. This leads to thefour components of the human factors framework ap-proach [6], namely person, context, activity and assis-tive technology, as an appropriate basis for modellingassistive technology systems. These components areall present to some extent in the four models, thoughnot always explicitly. However, the presentation of thefour components in the three modelling frameworksand their degree of detail is determined by the intendedapplications and therefore in most cases unsuitable forthe modelling of assistive technology systems. Boththe HAAT and MPT models are better suited to regionsor countries with a modern infrastructure, due to thelack of categories covering infrastructure, societal atti-tudes and legislation. These categories are present inthe ICF model (often under different names), but thenature of the detail is more appropriate to health datathan assistive technology applications.

All four components are present explicitly in theHAAT model [6] and this modelling approach is im-portant as the first attempt to apply this framework toassistive technology applications. However, the focuson providinga conceptual framework for assistive tech-nology applications means that the treatment of thesecomponents is not necessarily suitable for other aspectsof assistive technology modelling. In particular thediscussion of the activities component [6] refers to thethree performance areas of ‘daily living, work and pro-ductive activities and play and leisure’, in the UniformTerminology for Occupational Therapy [10] activitiesmodel, but does not specify it in detail. End-user as-pects of the assistive technology component, such asusability, documentation and training are also lacking.

The detail of the ICF model categories is much moreappropriate to its intended application of the collectionof health related data than assistive technology applica-tions. However, it should also be noted that even for thehealth data context the proliferation of detail and theneed to encode the data reduces the user friendliness ofthe model. The four components of the human factorsapproach are spread across the two part structure ofFunctioning and disability and Contextual factors.

The ICF classification is based on health and health-related states rather than people, though it does includea completely unclassified personal factors component.The very large number of medically inspired categoriesof body functions and structures, the lack of classifi-cation of personal factors, and the division of personrelated classifications between different categories ofthe ICF does not facilitate its use in assistive technol-ogy modelling and means it does not support designfor all approaches or particularly encourage holistic ap-proaches based on consideration of social, infrastruc-tural and environmental barriers. Assistive technologyclassifications are only considered in one chapter of theEnvironmental Factors component of Contextual Fac-tors. The granularity of the sub-classes in this Productsand Technology chapter is very poor [1,22] and thereis some overlap between the different classes. The sys-tem of classifiers seems to be designed purely for datacoding purposes and therefore does not facilitate ob-taining an overview of the field or insight into holis-tic results and outcomes. The activity classification isbased on the Component Activities and Participation,but the WHO documentation [27] recognises that thereis some ambiguity and overlap between the two terms.

The MPT model is intended to be used in outcomesmodelling and the matching of assistive devices to aparticular end-user. Therefore, the person componentof the model, including personal characteristics, prefer-ences and social interactions and support, is better de-veloped in the MPT that the other two models. Howev-er, though all four components are included, the modellacks a formal structure and is based on data obtainedthrough the associated MPT assessment forms. Theseinclude questions about performance of, and/or satis-faction with a number of activities, but there is no sys-tematic framework or classification of activities and thechoice seems to be based on assumptions of what typesof fundamental activities disabled people might expe-rience problems with. The MPT model does considerthe human and assistive technology system interactionas part of the consideration of outcomes. This includesissues, such as device effectiveness, device satisfaction


and end-user well-being, which should be an importantpart of assistive technology. However, technology de-tails and factors such as ergonomic design, technicalreliability and cosmetic appearance are lacking.

In summary, the following comments can be made:

– All three modelling approaches have been influ-enced by the the medical model of disability todiffering extents.

– The three models are reasonably well suited totheir intended applications, the collection andanalysis of health related data for ICF, provision ofa conceptual framework for the general discussionof assistive technology applications for HAAT andmatching assistive technology to the end-user andassessing outcomes for MPT. However, they arenot suited to the much wider range of assistivetechnology applications being considered and de-fined in this paper.

– Although all three models consider the four com-ponents of the human factors framework to someextent, even if not explicitly, the treatment has anumber of limitations and disadvantages. Theseinclude the focus on health and health related statesand the use of coding rather than a more userfriendly classification in ICF; the lack of detailof many of the model components and tendencytowards an engineering rather than end-user fo-cus in HAAT; and an implicit classification basedon factors identified in the associated assessmentprocedures, rather than an explicit classification inMPT.

– An inappropriate level of classification detail,missing classifications and /or implicit rather thanexplicit classification characterise all three mod-els. They all lack a number of important classifica-tion areas, particularly in the person and assistivetechnology domains. The ICF classifications aretoo detailed to be user friendly, whereas the MPTclassifications are only defined implicitly throughthe associated assessment framework, rather thanexplicitly.

With regards to the three modelling approaches onwhich the ICF, HAAT and MPT are based, classifica-tion approaches have the tendency, illustrated by theIFC, to include excessive detail, but ignore the interac-tion between categories, whereas outcomes modelling,as illustrated by MPT, has too limited a range of ap-plications, namely all aspects of assistive technologyprovision and outcomes assessment. Despite the limi-tations of the HAAT approach, systems modelling has

considerable potential and therefore the new Compre-hensive Assistive Technology model will be developedwithin a systems modelling framework.

3.3. Proposals for a Comprehensive AssistiveTechnology(CAT) model

3.3.1. Need and applications for a CAT modelThe model review in the previous section has iden-

tified that the established approaches to modelling as-sistive technology systems only provide a basic outlinestructure for such models and are far from meeting allthe requirements of a well-developed framework. Forinstance, the existing modelling structures cannot beapplied to investigate the adequacy or quality of ex-isting assistive technology provision and identify areaswhere it is lacking. In addition, they are unable to pro-vide a technological framework for the design speci-fication (synthesis) and development of new assistiveapplications and devices.

The first stage in any modelling process is often spec-ifying or clarifying what is being modelled. In thiscase this has been done through the derivation of a newdefinition of assistive technology in Section 2.2. Thisdefinition will form the basis of the subsequent mod-elling process. For ease of reference this definition isrepeated here:

‘Assistive technology is a generic or umbrella termthat covers technologies, equipment, devices, appa-ratus, services, systems, processes and environmentalmodifications used by disabled and/or elderly peopleto overcome the social, infrastructural and other bar-riers to independence, full participation in society andcarrying out activities safely and easily.’

Drawing on the modelling review in the previous sec-tion and the new definition of assistive technology, thefollowing requirement for a new modelling frameworkfor assistive technology systems can be stated:

(i) Applicable to any assistive technologysystem,with assistive technology characterised in thebroader sense of the above definition.

(ii) Able to provide a classification and archivalframework for assistive technology systems.

(iii) Capable of identifying where new assistivetechnology applications are required. Thiswill require accessibility barriers to be identi-fied.

(iv) Provide a framework for modelling the genericstructure of assistive technology systems andlend itself to analysis and synthesis (designspecification) procedures.


(v) Facilitate the development process for new as-sistive technology systems.

(vi) Provide a framework for the evaluation of theperformance of an assistive technology sys-tem.

(vii) Able to support the process of providing assis-tive technology for a particular user, with theaim of avoiding device rejection and abandon-ment.

(viii) Help engineers and other professionals obtaina clearer understanding of how assistive tech-nology systems function in a social context.

(ix) Applicable to a wide range of different con-texts, not just where there is a modern andwell-developed infrastructure.

Modelling human activities is an important feature inall the three modelling approaches described in Section3.2. For example, a checklist of activities can be usedto identify where accessibility barriers exist for an indi-vidual or group of individuals. Thus, prior to present-ing the detail of the model in Section 4, the modellingof ‘activities’ is discussed next.

3.3.2. Modelling human activitiesOne aspect of assistive technology modelling that

has received some attention is that of modelling human‘activities’. A model is needed to determine whetherand what type of assistive technology support may berequired for the individual or within the societal in-frastructure generally. There are several different waysof creating a human activities model depending on thedesired end-use of the model and/or the perspective ofthe modeller. The discussion here follows the overviewmodelling review of Section 3.1.

Part 1 of the ICF approach contains the category ofActivities and Participation. For the purpose of thisdiscussion the components of Activities and Participa-tion will be represented in a tree diagram form, simi-lar to that to be used in one of the representations ofthe Comprehensive Assistive Technology model, andshown in Fig. 8.

The ICF definesactivityas the execution of a task oraction by an individual andparticipationas the person’sinvolvement in a life situation. The figure shows thepartition of the classification into these two groupings.Although, as recognised by WHO [27], this distinctionis rather imprecise, there is an important distinction be-tween what can be considered fundamental or basic ac-tivities and activities which are specific to a particulartype of situation or context. Examples of basic activ-

ities include face-to-face communciation and movingshort or long distances, whereas examples of a situationor context specific task include teaching and shopping.It should be noted that both teaching and shopping fre-quently involve both face-to-face communication andmoving short or long distances.

Thus the CAT model includes a partition of activi-ties into fundamentalandcontextual activities, whichis related to, but not identical to the ICF distinctionbetweenactivity andparticipation. The ICF compris-es nine activity classifications with extensive subdivi-sions. However, the number of categories is so largeas to be unwieldy and making it difficult to obtain anoverview of the model. In addition, there is an exces-sive degree of detail. Therefore, appropriate choice ofa smaller number of less detailed activity categories isrequired to give a model structure which can be easilycomprehended and overviewed. The selection in theCAT model is discussed in Section 4. It should be not-ed, that, despite the similarities at the higher levels, theCAT model structure is not identical to that of the ICF.

Another route to modelling human activity is to usethe classification of assistive technology products anddevices. Although only indirectly a model of humanactivities rather than assistive technology, this is thebasis of the ISO 9999 standard. It is also the approachused in the EU HEART project after the first level tech-nical area categories. The HEART project [15] definesthe four technical areas of communication, mobility,manipulation, and orientation, with each area furtherdivided into activities and categories of assistive tech-nology devices, as follows:

(i) Communications: interpersonal communica-tions, reading/writing, computer access/user in-terfacing, telecommunications

(ii) Mobility: manual mobility, powered mobil-ity, accessibility, private transportation, pub-lic transportation, orthoses, prosthetics, seatingand positioning.

(iii) Manipulation: environmental control, robotics,orthotics and prosthetics, recreation and sports.

(iv) Orientation: orientation and navigation sys-tems, cognition.

The HAAT model uses the Occupational Therapycategories to describe the activities component of themodel. The occupational therapy categories derivefrom the Uniform Terminology for Occupational Ther-apy [10] and these are based on the three basic perfor-mance areas of daily living, work and productive activ-ities, and play and leisure. Although these categories


In Part 1:

Activities

and

Participation

Learning & applying knowledge

General tasks and demands

Communication

Mobility

Self-care

Interpersonal interactions/relationships

Domestic life

Major life areas

Community, social and civic life

Activities

Participation

Fig. 8. Tree diagram for the ICF component, Activities and Participation.

are useful for end-user assistive technology system as-sessment, there is a degree of ambiguity and overlapbetween them. It is interesting to note that the end-userquestionnaire devised by Cook and Hussey only hassections on the fundamental activities and on the singlecontextual activity, Daily Living. Categories in the ICFare not used.

DiscussionAll the approaches described above have their lim-

itations. In particular, they may restrict the activitiescovered to those for which there are assistive technol-ogy categories or which are considered appropriate insome sense for disabled people. As a result, some ac-tivities, including those where barriers exist for somedisabled individuals or groups of disabled people, maybe missed. In addition, the ICF and the occupationaltherapy categories are characterised by overlap, as wellas some categories that are too finely detailed and oth-ers that are too broad. It is not easy to define the ap-propriate breadth and specificity of activity categoriesand a trade-off is required between sufficient precisionto avoid overlap and maintaining a reasonable numberof categories at each level.

Over-precise task decomposition is also generallyinadvisable for the following reasons:

(i) Many higher-level activities take place withina particular context and often have more thanone possible decomposition as a sequence of ba-sic tasks. It is generally the ability to achievethe higher-level activity that is important ratherthan the ability to carry it out in a particularway. Since the contextual information is gener-ally lost when higher-level activities are decom-posed into basic tasks, this may result in a mis-

understanding of what is required. For this rea-son, it is valuable to have fundamental(basic)ac-tivities and contextual activities (participation)categories.

(ii) A very detailed decomposition into basic taskscan lead to overlap between some of the basictasks resulting from high-level activities, whichshould then be removed. For example, a numberof high-level activities, including listening to aradio, choosing the programme on some designsof washing machine or choosing the temperatureof an electric oven, involve turning dials. Conse-quently, a model basedsolelyon decompositioninto simple tasks is inappropriate.

The above discussion shows that a new approach toclassifying human activities is required to support assis-tive technology development, analysis, and selection.The approach used in the new Comprehensive AssistiveTechnology model presented in Section 4 is thereforerequired to meet the following specifications:

(i) To cover all (major) human activity areas whilstminimising overlap between categories

(ii) To provide a systematic structure with an un-derlying logical justification for the categorieschosen

(iii) To define activity areas of an appropriatebreadth and with a precise degree of specificity

(iv) The activity areas should be able to identify thereal barriers encountered by different groups ofdisabled people in carrying out a wide range ofsocial and domestic activities.

(v) The activity areas should be based on a hierar-chical structure with two main levels and notmore than eight categories on each branch to


give a model of manageable size and complex-ity.

4. The Comprehensive Assistive Technology (CAT)model

4.1. Justification of the choice of model representation

Model development generally involves tradeoffs be-tween comprehensiveness and simplicity, since com-plexity increases with the number of factors and aspectscovered. An assistive technology framework which canbe used for the classification, synthesis, and analysisof particular devices and for supporting the develop-ment of new devices and the appropriate matching oftechnology to the user is likely to be very complicat-ed. This complexity can be managed by the choice ofan appropriate modelling framework. The CAT modelpresented here was developed out of the HAAT mod-elling framework introduced by Cook and Hussey [6]and also has some relationship to the MPT model andICF. This approach has been chosen for the followingreasons:

(i) The model can be represented by a set of treediagrams, as shown in figures 9–14, with a lim-ited number of variables on each branch. Thismakes the model much easier to understand.However, the diagrammatic form of the treestructure is purely for the convenience of sight-ed readers and the model is not inherently visu-al.

(ii) The top level of the model contains the fourcomponents that define all assistive technologysystems, namely the user, who should be at thecentre of assistive technology design; the con-text(s) in which they will be using the assistivetechnology; the activities for which they wantto use it; and the technology.

(iii) The CAT approach gives a generic frameworkfor the categorisation, development, assessmentand person matching of assistive technologysystems. This framework includes all the mainfactors, including the social and engineering di-mensions of assistive technology.

(iv) The tree-structure approach is very flexible. Aspresented, the model is comprehensive, withoutbeing complicated and it can easily be simpli-fied by omitting variables that are either not rel-evant or important in a particular context. Thisstructure also facilitates software implementa-

tion. Furthermore, the tree diagrams lead toother representations, in particular, labelled at-tributes and tabular forms that can be used indifferent applications studies.

This is a considerable extension of the approach inthe HAAT model. The model of the person has beenextended to avoid reinforcing the tendency of engineersand designers to ignore the wider aspects of human-centred design, such as aesthetics and values. Thesewider human aspects often determine whether a deviceis actually used or rejected and whether it really meetsusers’ needs. Using the model review and the con-sideration of previous human activities models, a newdecomposition of the activity category has been intro-duced. The model is globally applicable to assistivetechnology both within and without the industrialisednations.

4.2. Description of the Comprehensive AssistiveTechnology(CAT) model

The Comprehensive Assistive Technology model hasa tree structure that does not use an excessive numberof branches at any level. This has the advantages ofbeing easy to understand, navigate, and modify, as wellas facilitating the development of interactive software.The model is based on decomposed layers of attributesthat cover the relevant aspects of a person, their en-vironment, and the assistive technology system beingused to support their activities. These attributes are or-dered to give a structure that makes sense of the differ-ent components, and this structure can be representeddiagrammatically, as attribute sets or in words, amongstother ways. The first or top level has the followingfour branches, illustrated in Fig. 9, Person, Context,Activities, and Assistive technology system.

The model is then developed by carefully definingfurther branching levels for each of these top-level cate-gories. The definitions of these branches and levels aremotivated by the requirement to use the model to anal-yse existing technologies and develop new ones. De-tails of the sub-branches for the first level branches ofthe person, context, activities, and assistive technologysystem will now be presented.

4.3. Person attribute in the CAT model

The person, or group of people, who are going touse a particular assistive technology is central to thesuccess of the system. The second level components


Comprehensive

Assistive

Technology

model

Person

Context

Activities

Assistive

Technology

Fig. 9. Comprehensive Assistive Technology (CAT) model.

Person

Social aspects

Attitudes

Community support

Education & employment

Attitude towards assistive technology

General attitudes

Characteristics

Personal information

Skills

Impairments

Preferences

Fig. 10. Model for person attribute.

are Characteristics, Social aspects, and Attitudes (seeFig. 10).

At the third level, the Characteristics variable con-sists of Personal information, Impairments, Skills andPreferences. Personal information includes age, gen-der and ethnic origin. Impairments involves sensory,physical, cognitive, mental health and other impair-ments. These impairments affect the activities wherethe person may experience barriers and therefore mightrequire assistive technology and the suitable designof this assistive technology so they do not experienceproblems in using it. Skills may be innate or the resultof education and training. Most people have a widerange of very different types of skills, not all of whichare relevant to the use of assistive technology. Pref-erences include preferences for the type of interface,device appearance, receiving basic or detailed informa-tion and the way information is presented, for instanceas speech, text, pictures or as a combination of severalformats.

The social aspects variable comprises (at the thirdlevel) Community support, and Education and employ-ment. Community support may have emotional, prac-tical, and/or financial aspects. It involves the availabil-

ity of support and friendship from the local communityand/or family and friends. This support and encour-agement can be an important factor in a user having asuccessful experience with an assistive device. Educa-tion and employment include current employment sta-tus, employment history, education and training historyand qualifications. Education and employment historycan contribute to confidence or lack of it and educationand training will have an effect on determining levelsof knowledge and skills.

At the third level, the Attitudes variable compris-es Attitudes to assistive technology and General atti-tudes. User attitudes to assistive technology includeexperiences with assistive and other technologies, andhow willing they are to try new technology. It alsoincludes preferences for assistive technology, personalassistance or a combination of the two. General atti-tudes include self-esteem, self-identity, attitudes to dis-ability, (self-)motivation and their degree of persever-ance.

4.4. Context in the CAT model

The second level describes the main types of con-text, whereas the third level specifies these contexts in


Context

Cultural &

Social context

National

context

Local

settings

Wider social and cultural context

User's social and cultural context

Infrastructure

Legislation

Assistive technology context

Physical variables

Location and environment

Fig. 11. Model for context attribute.

more detail, as shown in Fig. 11. It is important thatassistive technology design is based on the user’s exist-ing context and does not require them to change it, forinstance by requiring a modern infrastructure or the useof English. However, the provision of assistive tech-nology can enable the user to work towards improvingtheir context, either on their own or together with otherpeople.

There are three main types of context at the secondlevel (see Fig. 11):

– Cultural and social context– National context– Local settings

At the third level, the cultural and social context isdivided into Wider social and Cultural context and Us-er’s social and cultural context (See Fig. 11). These twocontexts are often very similar, particularly for peoplewho form part of the dominant cultural and social con-text. There are frequently differences for members ofminority groups, including disabled people. Variablesof interest in both these contexts include language, oth-er cultural factors, attitudes to disabled people and atti-tudes to assistive technology. Language and other cul-tural factors are particularly important since many fea-tures of existing assistive technology devices are onlyprovided in English and, sometimes, a small numberof European languages. Both speech output, if any,and documentation and manuals needs to be in the lo-cal language. In addition, the choice of symbols orother labels for controls should be unambiguous in theparticular cultural context.

The national component is able to distinguish be-tween the contexts in countries with very different his-tories, socio-economic situations, and states of infras-tructure development. Together with the wider socialcontext, this is one of the features of this model thatmakes it applicable beyond the industrialised countries.The third level of the national component is divided in-

to Infrastructure, Legislation, and Assistive technologycontexts. The infrastructure context includes factorssuch as the availability and reliability of an electricitysupply and the state of development of information andtelecommunications technology, as well as the propor-tion of the population that have access to them. Oth-er infrastructure factors are the development of a roadand rail infrastructure and the extent to which publictransport and public and private buildings are acces-sible, as well as the state of repair of the infrastruc-ture. It would clearly not be very useful to supply anassistive device that requires a mains power supply ina country where the electricity supply is intermittentand unreliable and/or the majority of the population arenot connected to the supply. The legislative contextincludes legislation about accessibility and social in-clusion for disabled people and any relevant buildingor other regulations and national standards that affectassistive technology. The assistive technology contextincludes the local, regional or national system(s) fordistributing, paying and providingsupport, training andmaintenance for assistive technology. This includesthe availability of direct payment schemes for disabledpeople to employ personal assistants or purchase assis-tive technologyof their choice, the extent of availabilityof sign-language interpreters and guide-communicatorsfor deafblind people and how these services are funded.

The local settings context describes the various set-tings in which the user would wish to use assistive de-vices. Its third level components are Location and envi-ronment and Physical variables. The location and envi-ronmental context includes indoors, outdoors or both,by the user on their own or accompanied by other peo-ple. Indoor settings could be further classified accord-ing to the type of building or its main use,for instance asan apartment in a multi-storey block, one storey house,multi-storey house or as a home, workplace or educa-tional establishment. Outdoor settings can be furtherclassified into urban, rural or natural/’wild’ e.g. moun-


tain or seaside. It should be noted that some devicesare intended for use in one particular setting or type ofsetting, whereas others may be used in several differ-ent (types of) settings. For instance, a lift with Braillemarkings will be used inside in a multi-storey build-ing, whereas a wheelchair may be used both inside andoutside and in a variety of different types of locationand terrain. The local settings context includes anyconstraints or limiting factors arising from the setting,such as door or room size or the need not to disturbother people. Physical variables include temperature,noise levels and types, humidity, the level and type ofillumination and the types of surface and the state ofrepair of the local infrastructure.

4.5. Activities attribute in the CAT model

The activities attribute categorises the various activi-ties a person might want to carry out, for some of whichthey may require support from assistive technology.The six activities categories based on the following twomain groupings:

– The majorfundamentalactivity categories of mo-bility, communications, and access to information,and cognitive activities.

– The majorcontextualactivities of daily living, ed-ucation, and employment, and recreational activi-ties.

It should be noted that the definition of fundamen-tal activities and contextual activities mirrors the cate-gories of activities and participation in the ICF schema.The three fundamental activity categories are similar tothose in the ICF scheme, and cover most of the activ-ities in the four HEART project technical areas. Thethree contextual activities are similar to participationactivities in the ICF but are more closely related to thethree occupational therapy performance areas. Thus,the activities model could be considered to combine theadvantages of previous approaches whilst simplifyingand reducing the excessive detail of the ICF approach.This then gives the following six main components inthe second level from the activity component of theCAT model (see Fig. 12):

– Communication and information– Mobility– Cognitive activities– Daily living– Education and employment– Recreational activities

Activities

Mobility

Cognitive

activities

Daily living

Education and

Employment

Recreational

activities

Communications

& Access to

information

Fig. 12. Model for activities attribute.

It should be noted that each of the contextual activitycategories involves activities from all of the fundamen-tal activity categories, namely, some communicationand/or use of information, cognitive activities, and mo-bility activities. This is one of the reasons that the activ-ities in these categories are considered ‘fundamental’.For instance, computers and the internet can be usedin daily living and the workplace or educational estab-lishment, as well as to obtain information about leisureactivities. Many daily living, education and employ-ment and recreational activities require some planningand organising, decision-making and analysis of infor-mation, as well as sitting and standing and/or liftingand reaching. Therefore, to avoid ambiguity and du-plication, cognitive, mobility and communication andinformation activities are not repeated in the contextualcategories.

Communication and information covers all the ac-tivities related to communication, whether carried outdirectly or via technology, as well as those related toaccessing information in all its forms. These activitiescomprise Interpersonal or face-to-face communication,Access to print media, Telecommunications, Comput-ers/internet and Communications and accessing infor-mation using other technologies(see Fig. 13a).

Interpersonal or face-to-face communication in-volves one-to-one, small group and large group com-munication. Access to print media covers both the pro-duction of print media and reading print media pro-duced by other people. All the communication cate-gories include artistic and imaginative as well as fac-tual communications. Communications and accessinginformation using other technologies includes the useof information kiosks and smart cards.

Mobility involves all the activities associated withmovement and safe travel (see Fig. 13b). This includes


Communication

& Accessing

Information

Interpersonal communication

Access to print media

Telecommunications

Computers and Internet accessibility

Communications using other technology

Fig. 13. (a) Communications and accessing information.

Mobility

Reaching and lifting

Sitting and standing

Short distance locomotion inside & outside

Long & medium distance locomotion

Movement on ramps, slopes, stairs & hills

Navigation and orientation

Obstacle avoidance

Access to environment

Fig. 13. (b) Mobility activities.

Cognitive

activities

Analysing information

Logical, creative & imaginative thinking

Planning and organising

Decision making

Categorising

Experiencing and expressing emotions

Calculating

Fig. 13. (c) Cognitive activities.

avoiding obstacles, as well as the navigation and ori-entation activities required to reach the right destina-tion by a desired route. It also includes moving ob-jects, sitting, and standing. This then gives the follow-ing categories at the third level: Reaching and lifting,Sitting and standing, Short distance locomotion insideand outside, Long and medium distance locomotion,Movement on ramps, slopes, stairs and hills, Obstacleavoidance, Navigation and orientation, and Access tothe environment.

Cognitive activities involve all the mental activitiesor activities related to thought processes i.e. Analysinginformation, Logical, creative and imaginative think-ing, Planning and organising, Decision making, Cate-

gorising, Calculating, and Experiencingand expressingemotions and feelings (see Fig. 13c).

Daily living involves all the different activities usedin everyday life. Most people will carry out some of theactivities in each category (except possibly Sexual andreproductive activities) on a regular basis and possiblyevery day or even several times a day. These activi-ties can be categorised as: Personal care, Timekeeping,alarms and alerting, Food preparation and consump-tion, Environmental control and household appliances,Money, finance and shopping, and Sexual and repro-ductive activities (see Fig. 13d).

Personal care activities include personal grooming,washing, toileting and sleeping. They also include per-


Daily living

Personal care

Timekeeping, alarms and alerting

Food preparation and consumption

Environmental control & household appliances

Money, finance and shopping

Sexual and reproductive activities

Fig. 13. (d) Daily living activities.

Education

and

Employment

Learning and teaching

Professional and person-centred

Scientific and technical

Administrative and secretarial

Skilled and non-skilled trades

Outdoor working

Fig. 13. (e) Education and employment activities.

Recreational

activities

Access to visual, audio & performing arts

Games, puzzles, toys and collecting

Holidays & visits: museums, galleries, etc

Sports and outdoor activities

DIY, art and craft activities

Friendships and relationships

Fig. 13. (f) Recreational activities.

sonal care activities carried out for a baby or smallchild. Timekeeping, alarms and alerting includes fireor smoke alarms, alerts to a range of activities, suchas the telephone ringing, the door bell (including anindication of who is at the door), a baby crying anda wake-up signal. Food preparation and consumptionincludes the use of cookers, microwaves and blendersto prepare and cook food. Environmental control andhousehold appliances includes the use of washing ma-chines and vacuum cleaners, as well as turning lightson and off, opening and closing curtains, blinds anddoors and control of the heating system. Sexual and re-productive activities include assisted conception, suchas donor insemination and in-vitro fertilisation, and theuse of contraception.

Education and employment comprises the widerange of differentactivities involved in education, train-

ing and employment, both paid and voluntary. Edu-cation and employment can further be categorised asfollows (see Fig. 13e): Learning and teaching, Pro-fessional and person-centred, Scientific and technical,Administrative and secretarial, Skilled and non-skilledtrades, and Outdoor work.

Recreational activities can be divided into the cate-gories of Accessing the visual, audio and performingarts (TV, cinema, theatre, radio, music, dance), Games,puzzles, toys and collecting, Holidays and visits: mu-seums, galleries, heritage sites, Sport and outdoor ac-tivities, DIY, art and craft activities and Friendshipsand relationships (see Fig. 13f). There is some overlapbetween the three contextual activity categories. Forinstance, painting and decorating can be either an em-ployment activity (skilled trade) or a recreation activity(DIY and craft activities) and cooking a meal can be


either a daily living activity (food preparation) or anemployment activity (skilled trade).

4.6. Assistive technology attribute in the CAT model

Since the community environment is generally notdesigned for disabled people, many disabled people re-quire assistive technology to carry out their desired ac-tivities. In some areas, appropriate assistive technolo-gy systems are not (yet) available so that some groupsof disabled people will have considerable difficultiesor even find it impossible to carry out some desiredactivities. At the second level, as shown in Fig. 14,the assistive technology system attribute can be dividedinto the following components:

– Activity specification– Design issues– System technology issues– End user issues.

At the third level, the activity specification compo-nent includes: Task specifications and User require-ments. In task specification, a task or activity is de-composed into a list of the subtasks that the assistivetechnology has to accomplish. User requirements cov-er the physical, sensory and cognitive demands madeon the user, such as the ability to lift a certain weightor to follow a sequence of instructions. In general thefewer and less demanding the user requirements, thewider the group of people who can use a particulartechnology. At the third level, design issues can becategorised as Design approach and Technology selec-tion. The design approach is the overall design philos-ophy or strategy used. This includes whether it is basedon design for all or design for specific groups of (dis-abled) people and/or design for environment or designfor maintainability. Design for all aims to design foras wide a population group as possible, independentlyof factors such as age, gender, ethnic origin, size ordisability. Technology selection involves determiningthe main technologies on which the design could be oris based. For example, in the case of an obstacle detec-tion device the technology choice may lie between in-frared and ultrasonic technology and the selection pro-cess would consider factors such as performance, theuser interface, reliability, technical specifications andcost.

At the third level, the assistive technology sys-tem comprises System interface, and Technical perfor-mance. Specification of the system interface includeswhether it is single or multi-modal and which modali-

ties are available, for instance speech, text, icons and/orBraille. Technical performance includes a range of fac-tors, such as reliability, robustness, safety features andease of maintenance. At the third level, end-user issuescan be categorised as Ease and attractiveness of use,Mode of use, Training requirements, and Documen-tation. Ease and attractiveness of use covers a rangeof factors, including whether the assistive technologyis ‘user friendly’ and intuitive to use, as well as size,weight, portability and appearance. Mode of use in-cludes whether the device is portable or remains in agiven location and whether it is stand-alone or formspart of a larger system. Training requirements speci-fy the types of training (different groups of) users arelikely to require in order to use the assistive technologysuccessfully. This may involve training before usingthe technology for the first time and ongoing training.Factors associated with documentation include the for-mats it is available in, such as on-line and hard copy,how it is structured and whether it is targeted at noviceor expert users or there are different sections for boththese groups of users.

4.7. Applications

The CAT modelling framework has a number of im-portant applications, including the following [17]:

– Identification of gaps in assistive technology pro-vision. This could lead to the development of as-sistive technology systems for areas where thereare currently no systems available, or an extensionof the capabilities of existing systems to offer moreoptions to end-users.

– Analysis of existing assistive technology systems.This could lead to design modifications to im-prove performance, user satisfaction and/or in-crease functionality, as well as the specification ofnew devices.

– Design synthesis. This could lead to the develop-ment of design specifications for new devices, forinstance to meet identified gaps in provision.

– The provision of support for design for all i.e. thedesign and construction of devices and environ-ments to be accessible and useable by as wide arange of the population as possible, including dis-abled people. The model could be used to structurethe design approach. Consideration of the person-al characteristics (person attribute) of a particulargroup or groups of end-user is used to identify thebarriers they will encounter in carrying out a de-


Assistive

Technology

Activity

specification

Design issues

System

technology

issues

Task specification

User requirements

Design approach

Technology selection

System interfaces

Ease & attractiveness of use

Technical performance

Mode of useEnd user issues

Training requirements

Documentation

Fig. 14. Model for assistive technology attribute.

fined set of activities using a particular technology(technology attribute) or in a particular environ-ment (context attribute). The (assistive) technolo-gy and context attributes can then be used togetherwith the person attribute to determine the requireddesign features to make the technology or attributeaccessible to the end-user groups. This approachcan either be carried out separately for each end-user group being considered and the design speci-fications combined or all potential groups of end-users can be considered together from the start.

A number of applications in the area of analysisand design synthesis for blind and deafblind people arediscussed by the authors in the book [17].

5. Conclusions

Assistive technology is still a relatively young sub-ject in which the underlying theory is still being de-veloped. One of the areas of assistive technology inwhich the theory is still relatively undeveloped is that ofmodelling assistive technology. Currently, only a fewmodelling approaches are available, with only a limit-ed range of applications. In addition, researchers havegenerally not taken the step of defining assistive tech-nology precisely before developing a modelling frame-work and existing models are rarely based on a specificdefinition.

Since disabled people are one of the main users ofassistive technology, definitions of assistive technolo-gy require an understanding of what is meant by theterm ‘disability’. Therefore the paper commenced with

a discussion of the two main models of disability, themedical and social model. The social model, with itsstress on overcoming barriers and design for everyone,was considered more appropriate as a basis of assistivetechnology. This led to the examination of existing def-initions of assistive technology and the development ofa new definition based on the social model of disability.

This definition provided the basis on which the mod-elling framework was developed. However, first ex-isting assistive technology modelling approaches werereviewed. It was found that they could classified intothree main categories:

(1) Classification methodologies(2) System modelling methods(3) Assistive technology outcomes modelling

Although these models were found to be useful, noneof these models covers the full range of applications.Analysis of the limitations of the existing models andfull range of desired applications was used to draw upspecifications for a new assistive technology modellingframework.

This framework was then used to develop the maincontribution of the paper, the Comprehensive Assis-tive Technology model, a new modelling frameworkderived by the authors. Although the CAT modelhas some relationship to existing modelling approach-es, particularly the HAAT model, it has significantlygreater flexibility and a much wider range of applica-tions, as well as considerable differences of detail andlower level structure.

The CAT model has a number of applications, in-cluding the following:


(i) Identifying gaps in assistive technology provi-sion.

(ii) Analysing existing assistive technology sys-tems.

(iii) Developing specifications for new assistivetechnology systems.

(iv) Supporting the provision of assistive technolo-gy for particular end-users.

Model validation will be carried out through inves-tigation of applications. This will be done in Part 2of the paper, where examples of a number of differenttypes of applications will be presented that exploit therepresentational flexibility that exists in the CAT modelframework.

References

[1] S. Arthanat and J.A. Lenker,Evaluating ICF as a Frameworkfor Clinical Assessment of Persons for Assistive TechnologyDevice Recommendation, Proc. 10th North American Collab-orating Centre (NACC) Conference on ICF, Halifax, NovaScotia, Canada, June 1–4, 2004.

[2] R. W. Bailey, Human Performance Engineering: Using Hu-man Factors/Ergonomics to Achieve Computer System Usabil-ity, (second ed.), Prentice Hall, Englewood Cliffs, N.J., USA,1989.

[3] C. Barnes,Disabled People in Britain and Discrimination:A Case for Anti-discrimination Legislation, London, Hurst &Co., 1994.

[4] A. Bowling, Measuring Disease: A Review of Disease SpecificQuality of Life Measurement Scales, Buckingham, UK, OpenUniversity Press, 1995, 1–16.

[5] A. Bowling and J. Windsor, Towards the good life: it a pop-ulation survey of dimensions of quality of life,J HappinessStudies2 (March 2001), 55–81.

[6] A.M. Cook and S.M. Hussey,Assistive Technology: Principlesand Practice, 2nd edn., St. Louis, USA, Mosby Inc., 2002.

[7] H. Day and J. Jutai, Measuring the Psychosocial Impact ofAssistive Devices: the PIADS,Canadian J of Psychology, 9(1996), 159–168.

[8] L. Demers, R. Weiss-Lambrou and B. Ska, The Quebec UsersEvaluation of Satisfaction with Assistive Technology (QUEST2.0): An overview and recent progress,Technology and Dis-ability 14 (2002), 101–105.

[9] C. Digiovine, D.A. Hobson and R.A. Cooper, Clinical prac-tice of rehabilitation engineering, in:An Introduction to Re-habilitation Engineering, R.A. Cooper, H. Ohnabe and D.A.Hobson, eds, CRC Press, Boca Raton, USA, 2007, Chapter 2,pp. 19–45,

[10] W.M. Dunn, J. Foto, B. Hinojosa, L.K. Schell, S. Thomson andS. Hertfelder, Uniform terminology for occupational-therapy-3rd edn,Am J of Occup Therapy48(11) (1994), 1047–1054.

[11] B.T. Fay,Evaluation of Individuals with Visual Impairmentsfor Educational and Vocational Applications of Assistive Tech-nology, Procs. Technology and Persons with Disabilities Con-ference, 1998, California State University, Northridge, USA.

[12] M.J. Fuhrer, Subjectifying quality of life as a medical rehabil-itation outcome,Disability and Rehabilitation22 (July 2000),481–489

[13] M.J Fuhrer, J.W Jutai, M.J. Scherer and F. Deruyter, A frame-work for the conceptual modelling of assistive technologydevice outcomes,Disability and Rehabilitation25(no. 22)(2003), 1243–1251.

[14] E. Guglielmelli, M.J. Johnson, G.A. Di Lauro, C. Suppo, C.Laschi, M.C., Carrozza, P. Dario, A. Pisetta, Y. Perrella and G.Giachetti,A human Centred Design Method for Developinga Robot Appliance, Proc. AAATE Conference, 2003, Dublin,Ireland.

[15] HEART, Line C Rehabilitation Technology Service DeliverySystems in Europe, Brussels, European Commission, 1995.

[16] M.A. Hersh and M.A. Johnson,Assistive Technology for Deaf-blind People: Current State-of-the-art, Design, and end-userIssues: A Tutorial, Proceedings of Conference on AssistiveTechnology for Vision and Hearing Impaired People, Granada,Spain, 28 August, 2007.

[17] M.A. Hersh and M.A. Johnson,Assistive Technology for Vi-sually Impaired and Blind People, London, Guildford, UK,Springer Verlag, 2008.

[18] M.A. Hersh and K. Worrell,Design Issues in the Develop-ment of Communication Devices for Deafblind People, IEESymposium on Assistive Technology, Glasgow, UK, 2003.

[19] M.A. Hersh, K. Worrell and M.A. Johnson,A survey of AT forDeafblind People: State-of-the-art, and new Developments,Proc. 7th European Conference for the Advancement of Assis-tive Technology in Europe (AAATE), Dublin, Ireland, 2003.

[20] C. Hughes, B. Heang, J. Kim, L.T. Eisenman and D.J. Killian,Quality of life in applied research: a review and analysis ofempirical measures,Am J on Mental Retardation99(6) (1995),623–641.

[21] J. Jutai and H. Day, Psychosocial Impact of Assistive DevicesScale (PIADS),Technology and Disability14 (2002), 107–111.

[22] J.A. Lenker and J. Jutai,Assistive Technology Outcomes Re-search and Clinical Practice: What Role for ICF?Procs. 8thNorth American Collaborating Centre (NACC) Conference onICF, Toronto, Ontario, Canada, 2002.

[23] M.J. Scherer and G. Craddock, Matching person and technol-ogy (MPT) assessment process,Technology and Disability14(2002), 125–131.

[24] UPIAS, Fundamental Principles of Disability, London: UP-IAS (Union of the Physically Impaired Against Segregation),1976.

[25] WHO, 1980,International Classification of Impairments, Dis-abilities and Handicaps, Geneva, Switzerland, World HealthOrganisation, 1980

[26] WHO, 1994, International Statistical Classification of Dis-eases and Related Health Problems, Tenth Revision, WorldHealth Organisation, Geneva, Switzerland, Vol. 1–3, 1992–1994.

[27] WHO, 2001,International Classification of Functioning, Dis-ability and Health (ICF), World Health Organisation Publi-cation, ISBN-13 9789241545426, ISBN-10 924 154 5429,Geneva, Switzerland.

[28] J. Swain, S. French, and C. Cameron,Controversial Issuesin a Disabling Society, Open University Press, Buckingham,UK, 2003.

[29] D. Johnstone,An Introduction to Disability Studies, 2nd ed.,David Fulton Publishers Ltd, 2001.

[30] Fast–Definition of the term ‘Assistive Technology’, http://www.fastuk.org/about/definitionofat.php,accessed 31.1.08.

[31] Audit Commission,Older People and Well-being: The Chal-lenge for Public Services, Audit Commission, London, 2004.


[32] ACT Advanced Care Technologies Programme Definition ofTerms Assistive technology, http://www.actprogramme.org.uk/content/51/definition-of-terms, accessed 31.1.2008.

[33] Access to Assistive Technology in the European Union, Studyprepared by Deloitte and Touche,European Commission Di-rectorate General for Employment and Social Affairs, 2003.

[34] M.A. Pasha and S. Pasha,Let’s Redefine Assistive TechnologyDiscussion Paper, http://222.cartuk.org/?q=node/33, 2006.

[35] Technology and Disability,Aims and Scope, http://www. ios-press.nl/loadtop/load.php?isbn=10554181, accessed 13.2.08

[36] AAATE position paper,A 2003 view on Technology and Dis-ability, http://www.aaate.net.

on modelling assistive technology systems – part i

Documents