The Arabic psychosocial impact of assistive devices scale: Development, translation, and evaluation

ABSTRACT

This paper describes the development, translation, and early evaluation of the Arabic Psychosocial Impact of Assistive Devices Scale (AR-PIADS), an outcome measure instrument for the subjective impact of Assistive Technology on a person with a disabilities’ quality of life. Developing the AR = PIADS instrument involved forward and backward translation by two independent teams of bilingual, Arabic-English speakers (n = 5) and a quality and usability review by a panel of people with disabilities (n = 18). The emergent version was evaluated with a group of experienced Arabic-speaking Assistive Technology users (n = 67) for its psychometric properties. Initial results demonstrate a favorable comparison for 16 of the 26 questionnaire items with scores recorded for the original, English language version. Internal consistency, measured using Cronbach’s alpha, yielded a range of 0.97–0.99 for AR-PIADS while the new instrument’s reliability was assessed using an intraclass correlation coefficient resulting in scores within the range of 0.86–0.97 for the overall instrument. Despite these positive results however, the translation process did highlight a number of challenges with language and cultural interpretation of the translated instrument. This suggests that further work is warranted to explore its utility in service provision.

KEYWORDS:

• Arabic
• outcomes
• psychometric properties
• quality of life
• translation

Introduction

The importance of measuring the impact of Assistive Technology (AT) services has been well articulated (Fuhrer et al., 2003; Scherer, 1996). However, finding appropriate measures, implementing them correctly, collecting sufficient data, and satisfying the multiple requirements of stakeholders remain as barriers for successful AT service delivery (Jutai et al., 2005). The importance of outcome measurement for AT services for well-established service delivery models has been highlighted. A successful mechanism for measuring outcomes in a service can benefit organizations seeking internal or external validation from constituents within the service process (Parette & Dikter, 2007).

In June 2010, the State of Qatar established Mada, the first AT center in the Arabic speaking world. The global population of native Arabic speakers is approximately 300 million people of which 290 million use it as their first language (Lewis, 2009). It is spoken over a wide geographical area from North Africa (Maghreb) to the Arabian Peninsula, and much of the Middle East. Arabic belongs to the family of Afro-asiatic languages and currently there are three variants of Arabic in use: Classical Arabic; Modern Standard Arabic (); and colloquial or spoken Arabic. In addition to these, there are also numerous country and regional dialects and variations (Versteegh, 1997).

Prior to the establishment of the Mada Qatar Assistive Technology Center which was given a pan-disability remit in the Arab speaking world, access to such services in Qatar was sporadic and generally managed by non-governmental organizations representing specific disabilities or specialist education providers such as the Qatar Social and Cultural Centre for the Blind, the Al-Noor Center and the Shafallah Center. There was no legislative framework or statutory mechanisms to ensure the provision of technology to people with disabilities, as such the aim of the new center was to partially satisfy the State’s obligations to the commitments set out in the Convention on the Rights of People with Disabilities that had been ratified in 2009 (Avilés et al., 2019). The establishment of a new service quickly necessitated the development of a robust, well-validated mechanism for measuring outcomes. To this end, an Arabic language version of an established instrument: the Psychosocial Impact of Assistive Devices Scale (PIADS) was developed and implemented to support the provision of AT to people with disabilities in Qatar. This paper presents the Arabic Psychosocial Impact of Assistive Devices Scale (AR-PIADS), outlining its development and translation and presents the results of a preliminary psychometric evaluation conducted with experienced Arabic speaking AT users.

Background

AT refers to tools, devices and equipment used by people with disabilities to increase their independence, improve their ability to perform tasks and, enhance their participation within their home, community, and social life (R. O. Smith et al., 2018). It encapsulates a range of technology from low-tech solutions, for instance vision magnifiers or mobility aids, to high-tech, such as eye-gaze controlled computers or home automation systems (Federici & Scherer, 2017; Jutai et al., 2005). Developments in technology has seen people with disabilities using AT as an enabler to facilitate their participation in education, employment, and community life (Borg et al., 2011; E. M. Smith et al., 2022).

Over the past twenty years, there has been an increased interest in examining the outcomes of AT services and products within health and social care service provision (Edyburn & Smith, 2004; Kinney et al., 2016), reflecting the need to demonstrate success and value for money in an increasingly competitive context. However, measuring the success of AT use and service intervention is a complex and multifaceted process. It can include measures such as satisfaction (Borgnis et al., 2023), changes in functional status (Ryan et al., 2017), participation (R. O. Smith, 1996) and, also Quality of Life (QoL) (Hersh, 2010). Health-related professions recognize QoL assessment as the gold standard to measure treatment impact (Kalfoss, 2016). QoL refers to the subjective experience and views of the impact of the health condition on a person’s life (Paap et al., 2018). The areas often included in QoL measures are life satisfaction, subjective well-being, and a positive general affect. Measuring whether technology enhances an individual’s subjective perception of their own QoL is challenging and is reported to have limitations (Raphael et al., 1996). As such, QoL is closely associated with satisfaction in work, social relationships, and ability to go places regardless of the physical capability to do so.

PIADS: psychosocial impact of assistive devices scale

The Psychosocial Impact of Assistive Devices Scale (PIADS) was developed in Canada to provide a measure of impact of AT on the QoL of people with disabilities. It comprises three subscales which assess fundamental dimensions of QoL; 1) Competence, provides a measure of the perceived impact of the technology on people’s functional independence and task performance; 2) Adaptability, relates to the effect technology has on people’s perceived capability to seek and try new experiences or opportunities; and 3) Self Esteem, measures people’s subjective emotional evaluation of their own worth and the effect technology use has on their self-confidence. PIADS can be self or interviewer-administered and involves respondents considering a list of 26 words/phrases that describe how using technology may affect them as individuals. Respondents rate each item on a scale ranging from −3 (maximum negative impact) to + 3 (maximum positive impact).

Studies have reported on the use of PIADS with a range of ATs including eyeglasses (Day et al., 2001); speech recognition software (DeRosier & Farber, 2005); closed-circuit television (Huber et al., 2008); and wheelchairs (García et al., 2015). PIADS has also been used in a wide range of health and social care contexts with diverse AT users including people with physical disabilities (Devitt et al., 2004; Harada et al., 2014), deaf or hard of hearing (Saladin & Hansmann, 2008) and visually impaired users (Strong et al., 2003). More recently, PIADS has been utilized as a tool to examine the impact of web-based eHealth services (Wiklund Axelsson et al., 2013), mobility technology for those with neurological conditions (Jiménez Arberas et al., 2021), to monitor the progress of children using open license communication technologies (Draffan & Banes, 2022), and the use of a technology loan-bank for adults with neurological disabilities (Pousada et al., 2021).

The development of a French version of PIADS aimed to serve the needs of the French speaking population in its original Canadian context (Demers et al., 2002). It has since been translated into many languages including Italian (Tofani et al., 2020), Spanish (Díez et al., 2021) and Japanese (Inoue et al., 2011) and implemented in over 14 countries (Chae & Jo, 2014; Hsieh & Lenker, 2002; Orellano & Jutai, 2013). The instrument’s psychometric properties have been studied and shown good performance (Day et al., 2002). PIADS has also been developed as a context-specific outcome measure for particular healthcare technologies intervention (Long et al., 2014; Nordströ et al., 2014), and as an instrument to measure the impact of AT on families’ QoL (Ryan et al., 2007). Considering the extensive body of research on the use of PIADS and previous successful translations, it was chosen by Mada as an instrument that could be translated and used as part of its AT provision service. It is worth noting that at the time this study was conducted 2014/2015, there were no comparable outcome measurement tools available in the Arabic language, however, reports of efforts were emerging on the translation into Arabic of the QUEST 2.0 (Bakhsh et al., 2014). Following ethical approval for the study from the Research Department at the Mada Qatar Assistive Technology Center and from the Department of Computer Science and Statistics at Trinity College Dublin, the process of translating and conducting preliminary psychometric testing of the Arabic Psychosocial Impact of Assistive Devices Scale (AR-PIADS) commenced. The forthcoming section outlines the three-phase procedure followed to translate, pilot, and test the AR-PIADS.

Method

Phase one: PIADS translation

The objective of a translation process is to ensure continuity of meaning and understanding, from the original source language to the target language, without loss of the original meaning (Farghal & Shunnaq, 1999). Translation is a process of establishing equivalence between source language and target language. However, given the semantic, grammatical, and lexical differences among languages, achieving perfect linguistic equivalence is almost impossible. Translation is not only a linguistic exercise but also a social-cultural one and requires consideration of the socio-cultural milieu of the target language (Orellano & Jutai, 2013). In this regard, when translating tools that require people to reflect on the impact of AT on their QoL, the interpretation of cultural influences on the text translated from one language to another requires particular attention.

Differences among variants in languages such as Arabic can pose challenges when translating and establishing cultural equivalences (Alaa & Al Sawi, 2023; Mohammad, 2018). In the absence of guidelines for translation from English into Arabic (Khalaila, 2013), a method of translation and cross-cultural adaptation (Guillemin et al., 1993) that has been applied to other translations of PIADS (Demers et al., 2002; Hsieh & Lenker, 2002) was selected for this process. Central to this process is “back translation,” previously used in work on health and social care instruments (Khalaila, 2013). It involves a double translation process by which the translation from the source language into the target language is blind translated back again into the source language (Brislin, 1970). This process however required some modification to accommodate.

The procedure to translate PIADS into Arabic involved three steps: 1) backward translation; 2) review of backward translation; and 3) quality and usability evaluation of AR-PIADS. The backward translation procedure (step 1) involved two teams of bilingual Arabic-English speakers who independently translated the questionnaire items, accompanying instructions, glossary, and introductory text from English into Arabic. The first team comprised two third-level engineering students with no prior AT experience. The second team comprised two bilingual AT professionals and one communication professional conversant with the domain. Once the teams completed the work, their Arabic translations of PIADS were swapped from one team to the other and the teams independently translated them back into English. Although it could be argued that asking each group to backward translate each other’s Arabic version of the original PIADS may not be considered a “blind” translation process per se, it did however make use of the bilingual resources that were available and allowed the team to build a familiarity with the questionnaire items as has been done in other translation studies (Lee et al., 2019; Sperber, 2004; Tsang et al., 2017). To identify and address problems in translation, of words or concepts, it is good practice to have an independent group of bilingual speakers evaluate the translations (Drennan et al., 1991). The translations and back translations produced by the two teams were reviewed by a five- member committee. This committee also produced a final draft AR-PIADS. Finally, to provide quality assurance and assess the usability of the final draft AR-PIADS, a focus group with nine bilingual AT practitioners with no prior experience with PIADS was conducted. They were provided with the source English version of PIADS, the final draft AR-PIADS and the accompanying glossary in both languages. During the focus group, items identified as ambiguous were discussed and alternatives were agreed and a final version of AR-PIADS was delivered.

Phase two: AR-PIADS pilot trial

To gain further clarity and understanding on individual items of AR-PIADS a pilot trial was conducted with a convenience sample of 18 (14 females and 4 males) visually impaired, full-time AT service users (with more than one year experience). The participants had never been administered neither the English nor the Arabic versions of PIADS prior to the trial. The pilot procedure involved: 1) Face-to-face interview with the participants (in Arabic) introducing the study and AR-PIADS; 2) Administration of AR-PIADS; 3) Follow-up focus group to explore the participants’ understanding of the items as well as the relevance and cultural appropriateness of the same. During the last phase, participants were asked to contribute alternatives or modifications to items they considered inappropriate or problematic. A comprehensive description of the genesis of this project and the pilot has been reported elsewhere (Boyle & Zahid, 2013; Zahid & Boyle, 2013).

Phase three: AR-PIADS field trial

The AR-PIADS field trial was conducted with a convenience sample of 67 AT users. These were drawn from the client database of an AT Centre in Qatar. The inclusion criteria were: 1. Age: 18 years of age or over; 2. Does not have an intellectual disability; and 3. AT users in full-time possession of their own AT. Of the sample, 63% (n = 42) reported to be Arabic speakers only and 37% (n = 25) as bilingual Arabic-English, with Arabic as their mother tongue. Based on information recorded during the participants’ original assessment at the Mada Center, the disabilities presented in the sample were: blind/visual impairment (61%); physical disability (20%); deaf/hearing impairment (6%); and learning difficulty¹ (13%). Although the participants’ experience using AT varied, the majority (38%) reported having between three to five years’ experience using a range of devices.

In terms of recruitment procedure, the participants were contacted by telephone to first, explain the study and the nature of their participation in it; second, seek their consent for inclusion in the study; and third, schedule a time for a follow-up telephone interview. Previous work using this method produced comparable results to studies using face-to-face interviews (Day & Campbell, 2003). So, the administration of the PIADS and AR-PIADS was via telephone interview to ensure higher levels of participation.

Psychometric/statistical analysis

To examine the psychometric validity of AR-PIADS, its concurrent validity, test/retest reliability and internal consistency were tested. To this end, comparing the scores for the AR-PIADS with an accepted benchmark, such as the English PIADS, provides a measure of how closely related both tests are and the level of concurrent validity of AR-PIADS. In this study, the experimental AR-PIADS was administered to a group of Arabic only and bilingual Arabic-English speakers (n = 67). A single-sample two-tailed t-test was calculated to allow comparison of the scores achieved for each of the 26 questionnaire items in the source English PIADS and the experimental AR-PIADS. Typically for studies of translated instruments such as this, concurrent validity is determined when the results gathered by the new, translated version can be seen to correlate as similar construct such as with the original source version (Taherdoost & Group, 2017). Other translation studies of PIADS have compared individual item scores to the data originally collected by the tool developers (Day & Jutai, 1996) as part of the original PIADS validation study. Similarly, in the current study, the data collected from the administration of AR-PIADS was compared with the same data. In previous studies, a p of 0.05 was used (Chae & Jo, 2014; Day & Jutai, 1996; Hseing & Lenker, 2002) although in the French translation there is further discussion as to what the value of p should be for translation studies with relatively small sample sizes such as this (Demers et al., 2002). To further test the concurrent validity, the PIADS (source English version) and the AR-PIADS (Arabic experimental version) were both administered with bilingual participants (n = 25), which represented all of the bilingual speakers in the original sample, with an interval of one month separating the administration of the source and experimental versions. This sub-sample was broadly of a similar diagnostic category, 1 and all had in excess of one year of experience using a similar range of technologies. This process was similar to one employed by Demers et al. (2002) to demonstrate the validity of the French version of PIADS.

For evaluation of the test/retest reliability of AR-PIADS, a sub-sample (n = 20) Arabic only speakers were randomly selected to complete the AR-PIADS twice. This sub-sample was administered the AR-PIADS twice, with an interval of a month between interventions, by the same researcher and always by telephone. Thus, given that the same tool was administered twice, to the same sample and under the same conditions, any variations detected in the score should be attributed to measurement error. The interrelatedness of individual scores over a fixed time period was measured using the intra-class correlation coefficient (ICC) as was the case with the French and Korean language translations of PIADS (Demers et al., 2002; Hsieh & Lenker, 2002). To test the internal consistency of AR-PIADS: its reliability regarding the consistency of the results provided by different items measuring the same construct; the Cronbach’s α was used. This method provided a measure of the overall consistency of the AR-PIADS questionnaire and its three subscales.

Results

Translation issues

Feedback from participants during the initial pilot study (described as “phase one” above) identified the following translated items; 4. Adequacy and 11. Usefulness, from the source English PIADS, as the most challenging translations. Discussions recorded during the subsequent focus group (described above as “phase two”) highlighted further disagreement for the translations of the following items: 1. Competence; 4. Adequacy; 10. Frustration; 11. Usefulness; and 23. Willingness to Take Chances.

The Arabic translations for the remaining items were confirmed as being exact translations or suitably close to the source English items. Nonetheless, following discussion, strong disagreement remained regarding the translation of item 4. Adequacy. In consultation with a professional translator, the translation of this item was changed from to The final version for psychometric testing with services users is presented in Table 1.

Table 1. Comparison of psychometric properties: AR-PIADS and original PIADS.

PIADS items	AR-PIADS	Original PIADS Test Values	AR-PIADS Mean (SD)	df	T-test and p-value
COMPETENCE Subscale		1.44	1.4901 (0.4195)	66	−0.115	0.911
1. Competence		1.71	1.881 (1.3543)	66	1.031	0.306
3. Independence		1.33	1.269 (0.5663)	66	−0.887	0.378
4. Adequacy		1.27	1.09 (1.2027)	66	−1.228	0.224
5. Confusion		1.09	1.134 (2.1454)	66	0.169	0.866
6. Efficiency		1.83	1.806 (1.6259)	66	−0.121	0.904
8. Productivity		1.70	1.761 (1.0602)	66	0.472	0.638
11. Usefulness		1.59	2.09 (1.0259)	66	3.986	0
13. Experience		1.10	0.97 (1.8577)	66	−0.565	0.574
14. Skilfulness		1.35	0.8806 (1.3983)	66	−2.748	0.08
16. Capability		1.69	1.955 (0.9118)	66	2.381	0.2
17. Quality of Life		1.32	1.687 (1.3507)	66	2.221	0.3
18. Performance		1.79	1.358 (1.5926)	66	−2.219	0.03
ADAPTABILITY Subscale		0.89			1.766	0.3084
15. Well-being		1.05	0.866 (1.8821)	66	− 0.802	0.426
22. Willingness to take chances		.50	0.9701 (1.68748)	66	2.281	0.026
23. Ability to participate		1.06	0.94 (1.8081)	66	−0.542	0.59
24. Eagerness to try new things		0.63	1.254 (1.4496)	66	3.522	0.001
25. Ability to adapt to the activities of daily living		1.26	1.134 (1.8414)	66	−0.559	0.578
26. Ability to take advantage of opportunities		1.02	0.9851 (1.49233)	66	−2.002	0.049
SELF-ESTEEEM Subscale		0.77	1.289375 (0.555460154)	66	2.645	0.033
2. Happiness		0.81	2.194 (1.0904)	66	10.389	0
7. Self-Esteem		0.61	1.09 (1.6398)	66	2.394	0.02
9. Security		1.25	1.672 (1.7354)	66	1.989	0.4216
10. Frustration		0.85	1.343 (1.6195)	66	2.493	0.015
12. Self Confidence		0.90	1.284 (1.3905)	66	2.258	0.027
19. Sense of Power		0.62	1.045 (1.5612)	66	2.227	0.029
20. Sense of Control		0.99	1.433 (1.6351)	66	2.217	0.03
21. Embarrassment		0.38	0.254 (2.5664)	66	−0.403	0.688
OVERALL SCORE		1.03		66

Psychometric properties

To evaluate the concurrent validity, the p values calculated for AR-PIADS these were compared with the scores obtained in the original PIADS validation study (Day et al., 2002).

Table 1 presents the one-sample, two tailed t-test for independent samples scores for each of the 26 individual items, the three subscales and the overall instrument. The non-significant t = test scores outlined in Table 1 indicate the Arabic translations for each item do not produce different results to those obtained with the original English version of PIADS.

Previous studies have used a significance threshold of 0.05 (Demers et al., 2002) which means the overall scale and three subscales were not significantly different. This indicates each of the scales and the overall experimental instrument should yield scores comparable to the original English language PIADS. The analysis of individual p scores in this study shows 16 of the 26 instrument items are above the significance threshold (p > 0.05). These results certainly cast some doubt about the translation and suggests that items 2, 7, 10, 11, 12, 18, 19, 22, 24, and 26 merit further investigation. The p-scores for the Self-Esteem subscale and the overall instrument did not reach the significance threshold of p > 0.05 suggesting that equivalence between AR-PIADS and the source instrument requires further, more extensive psychometric testing.

Test–retest reliability was evaluated using data from the administration of the AR-PIADS twice over the course of one month with a subset of participants (n = 20). The intra-class correlation coefficient (ICC) values calculated are presented in Table 2, alongside the corresponding results for the measure of test–retest stability of the English language version of PIADS (Day et al., 2002). The range of scores compares favorably with the original source and indicates that AR-PIADS demonstrates good stability over time.

Table 2. Reliability and consistency of AR-PIADS subscales and overall scores.

		Internal Consistency (n = 67) Cronbach’s α
	Original	AR-PIADS	Test-Restest Reliability (n = 20) ICC
Competence	0.92	0.976	0.917 (0.74–0.97)
Adaptability	0.88	0.988	0.758 (0.007–0.962)
Self-Esteem	0.87	0.974	0.95 (0.773–0.990)
Overall	0.95	0.992	0.937(0.864–0.971)

Internal consistency was determined by calculating Cronbach’s ALPHA for AR-PIADS and each of its three subscales. The range of scores from 0.97 to 0.99 suggests a good internal consistency and is comparable to scores presented in other translation studies. Furthermore, the results obtained for the adaptability and self-esteem subscales are greater than those found in similar studies where scores reported are in the order of less than 0.90 (Day & Jutai, 1996; Saladin & Hansmann, 2008).

Discussion

The objective of this study was to establish an instrument that could assess the impact of AT on the QoL of Arabic speakers with disabilities and to collect data on the impact of technology that could be internationally benchmarked. The motivation was to provide an outcome measurement tool to be used in an Arabic AT service delivery. PIADS was considered a benchmark against which results from an Arabic context could be compared due to its extensive history of translation into other languages and contexts.

The results of our psychometric study suggest that AR-PIADS performs comparably with the original PIADS. However, issues remain with the translations a large proportion (n = 10) of the 26 questionnaire items that comprise the AR-PIADS. Furthermore, results suggest that further evaluation of the application of the overall instrument with particular attention focussed on the self-esteem subscale is required before claims can be made regarding the equivalence of AR-PIADS with the English PIADS. Considering some of the inconsistencies of translation in this study, it is worth noting that Arabic is spoken across a large geographical and cultural area and is characterized by significant regional variations, difference, and dialects; not only from country to country but also within cities and regions (Grigore & Bițună, 2015). This study was carried out in a Qatari context, as such, translations in this context may not be applicable or appropriate across the wider region because of significant variances in dialect from one country to another. Another consideration particular to Arabic is that as a language it exists in both vernacular and formal varieties, or al-lugha(t) al-fuṣḥā. In common parlance, Arabic speakers often use a mixture of both colloquial and formal versions of the language interchangeably depending on the communication context. Although this study was conducted within the context of a service provision agency in the State of Qatar, the nationality of Arabic speaking participants in the sample was mixed with representation from across the Middle East region. Thus, further work will be necessary to examine country and regional variations for translated items in AR-PIADS.

It is worth considering some of the limitations of the study reported here, in particular the fact that its scale is the relatively small drawing upon a limited sample size and that a convenience sampling method was employed. The sample size was impacted by the low number of service users in the context of this study. Two main factors contributed to the previous: 1) the AT service delivery organization in this study had been operational for less than four years at the time of this study; 2) the relatively small population within the State of Qatar. A further limitation that bears consideration is the focus of the study on a limited application of the AR-PIADS. Further work extending the evaluation of AR-PIADS psychometric properties to evaluate for instance its construct, convergent and predictive validity would be desirable. Such work may also involve comparison of the application of AR-PIADS with an extended range of comparable outcome tools. Notwithstanding the previous, findings from this study indicate AR-PIADS can be considered a reliable instrument for measuring and predicting the impact of the provision and use of AT on the QoL of people with disabilities.

Acknowledgements

The authors wish to acknowledge the support of the Mada Qatar Assistive Technology Center, in particular Mr Anirban Lahiri. The authors also wish to extend their gratitude to Dr Jeff Jutai for his guidance and support in the early stages of this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was supported by funding from the charity RESPECT and the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme (FP7/2007-2013) under Research Executive Agency (REA) grant agreement no. PCOFUND-GA-2013-608728.

Notes

¹ Participants with Learning Difficulties included dyslexia, dysgraphia and other specific learning difficulties. None of which had a diagnosis of an intellectual disability.