Measuring the adoption of internet and communication technologies among practitioners in routine disability services: a stepped inclusive approach

Abstract

Purpose

Existing measurements of the use of information and communication technologies (ICTs) among practitioners in disability services often treat ICT adoption as a monolithic concept, overlooking its multifaceted nature within the disability field. This study introduces a stepped, inclusive approach to capturing this complexity, elucidating disparities in the utilization of various ICT dimensions, the present vs. anticipated use, and variations among different clinical-demographic groups.

Materials and methods

A cross-sectional survey was conducted in Hong Kong, gathering valid data from 324 practitioners spanning diverse disciplines and disability services.

Results

Data analysis produced a three-factor model categorizing ICT tools into (1) information and communication tools, (2) screening and monitoring tools, and (3) treatment and rehabilitation tools. The first category was identified as the predominant ICT utilized currently, with significant projected growth in the latter two categories’ usage. Variances in current ICT adoption were influenced by practitioners’ roles, clientele, positions, affiliating agencies, and educational attainments.

Conclusions

This research provides a deeper understanding of the key dimensions of ICT adoption within disability services. It underscores the importance of devising specific and customized strategies for the effective integration of ICTs, ensuring a more tailored approach to meeting the unique demands of the disability field.

IMPLICATIONS FOR REHABILITATION

• Future studies focusing on information and communication technologies (ICTs) adoption or relevant concepts, such as e-health and telerehabilitation may consider employing similar methodological approaches accustomed to one’s local context.

• Future attention and investment in ICT adoption in disability services should focus more on domains directly relevant to clinical and rehabilitation practice, tailoring strategies to the specific needs of the field.

• There exists an urgent imperative to enhance ICT training, especially for psychosocial and medical professionals, while also increasing investments in non-governmental organizations.

• Such support needs to be gender- and age-inclusive, ensuring it meets the diverse needs of practitioners at all organizational levels.

Keywords:

• Internet and communication technologies
• ICT adoption
• telerehabilitation
• practitioner
• disability service
• digital tools
• ICT-mediated intervention

Introduction

With the popularization of computers, mobile devices, and the internet, the use of digital interventions in health and social services has grown rapidly. Various forms of information and communication technologies (ICTs) are now available or have been specifically designed to support the needs of people with disabilities. These ICTs encompass electronic devices (e.g., computers, smartphones, tablets), hardware/software/mobile applications, information storage systems, communication platforms (e.g., email, WhatsApp, Zoom), social networking sites (e.g., Facebook, Twitter), and assistive devices leveraging computer or internet-related technologies (e.g., Screen Reader). A range of terms, such as e-health, m-health, telehealth, telemedicine, telerehabilitation, e-care, and telecare have been used, occasionally interchangeably, to refer to interventions employing ICTs to enhance service access, delivery, and outcomes. Growing evidence demonstrates the effectiveness and advantages of incorporating ICTs within health and social care, highlighting its significant potential in improving service accessibility [1,2], intervention outcomes [3,4], user engagement [5], record management [1], and service cost-effectiveness [6]. For policymakers and service organizations, the integration of ICTs has emerged as a strategic approach to delivering more cost-effective services, facilitating access and participation with minimal barriers [7].

In disability services, despite the recognized potential and advantages of ICTs, their expansion over the past two decades has been notably slow, particularly as observed in research conducted before the COVID-19 pandemic [7,8]. Many ICT-mediated interventions have been integrated into services as supplementary features or for singular applications, utilized in ways disparate from the existing service structure [9,10]. Within the body of literature on the barriers to the digital transition of routine services, a recurring theme is the acceptance of practitioners. Although research on their perspectives, as a recent systematic review concluded, remains nascent [7], an emerging consensus before the pandemic underscored practitioners’ reluctance or hesitancy to embrace these technologies in service delivery [11–13]. Among the various factors that may impede the adoption of ICTs among practitioners, such as digital literacy, perceived benefits, and institutional resources [14–17], a key concern is the impact of ICT integration on the client-practitioner relationship, a crucial determinant of human service outcomes [18]. Many practitioners in health and social care have questioned the compatibility of impersonal interventions with human services, especially regarding the development of genuine and strong rapport with service users [6,19]. Additionally, some expressed concerns that the boundaries in their relationships with clients may become obscured, for instance, when communicating through social media or instant messaging [20].

The COVID-19 pandemic significantly accelerated the adoption of ICTs in disability service delivery. As physical distancing measures became imperative, the demand for remote access to care and support surged. ICTs emerged as essential tools in maintaining continuity of care, offering efficient, accessible solutions for individuals who were otherwise isolated or unable to access traditional in-person services [8,10,16]. Additional funding from the government has been allocated to service organizations to leverage ICTs in supporting people with disabilities [21]. Meanwhile, research into the adoption of ICTs among practitioners is proliferating, contrasting with the minimal attention the issue received before the pandemic [7,8]. The reluctance of practitioners appears to have shifted post-pandemic, stemming from a deeper, practical understanding of the pros and cons of ICTs gained during the social distancing period. Recent studies have observed a substantial increase in the actual use of ICTs by practitioners in routine services and their willingness to integrate ICTs into future practice [16,17,22].

However, before drawing a robust conclusion regarding the current status of ICT adoption among practitioners and considering subsequent strategies for future development, a critical yet often overlooked challenge must be addressed: the complexity of measuring the adoption of ICTs within disability services. The extant studies often employed popular models and theories to gauge users’ intentions and actual use of technologies, and to identify various predictors of adoption (e.g., performance expectancy and effect expectancy). The most commonly used model in telerehabilitation research is the Unified Theory of Acceptance and Use of Technology (UTAUT) and its modified counterpart, UTAUT2, which were originally developed to investigate the adoption of the mobile application Pokémon Go [7]. For instance, the recent study conducted by Seebacher and colleagues on the adoption of telerehabilitation among rehabilitation professionals adapted the UTAUT questionnaire, substituting “Playing Pokémon Go” with “using telerehabilitation” [16]. Hennemann and colleagues similarly altered the UTAUT questionnaire to inquire about practitioners’ intentions to explore “eHealth interventions” in future practices [12]. This substitution of terms, using concepts, such as ICT, eHealth, or telerehabilitation as an overarching entity, has also been noted in studies that employed other well-known ICT adoption theories, such as the Technology Acceptance Model [23], as well as in non-theory-driven studies [24].

The challenge is that terms, such as e-health and telerehabilitation differ considerably from a single concept like “Pokémon Go”. They typically allude to a suite of ICT-mediated interventions that vary in medium (e.g., mobiles, internet, computers, and assistive devices leveraging ICTs) and intent (e.g., treatment, assessment, and monitoring). Specifically, in the disability context, treating the adoption of ICTs as a unidimensional construct can seriously compromise measurement reliability. When practitioners from varied training backgrounds or those serving clients with different types of disability are asked about general terms like telerehabilitation or eHealth without specific explanations or ICT examples, they might interpret these terms in relation to unique intervention tools they have encountered or adopted at varying degrees in practice. In this sense, the trustworthiness of other related findings (e.g., factors influencing the adoption of telerehabilitation) generated from these studies also becomes questionable.

A few studies have tackled this issue by examining the adoption of ICTs, using a self-developed list of ICT items to measure the frequency of use. For instance, Feijt et al. employed 11 items to gauge the use of digital mental health interventions, covering a range of digital tools, such as emails, videoconferencing, wearables, and VR/AR [22]. Similarly, Bezuidenhout and colleagues defined ICTs for telerehabilitation with individuals with neurological diseases or older adults by measuring the adoption of nine digital tools (e.g., SMS, video calls, and wearables) in practice [25]. While these measures might enhance the trustworthiness of assessing the actual state of ICT adoption in practice, the incorporation of an extensive list of variables can render the concept less interpretable. Variations in actual use would have to be identified based on comparisons of individual digital tools. Challenges might also arise when trying to analyze the relationship between influencing factors and the collective adoption of ICTs. Therefore, while the detailed breakdown of ICT adoption into specific tools or applications can enhance precision and reliability, it can also add complexities that hinder interpretation, analysis, comparison, and generalization in research, potentially constraining progress in the field.

Objectives of this study

The primary aim of this study is to introduce a stepped, inclusive approach to navigating the complexity of measuring ICT use for research on its adoption in the routine disability service context. First, drawing from a review of existing literature and expert advice, the research team categorized the purposes of ICT adoption in the disability service context into significant dimensions. A list of digital tools, featured in prior research, was sorted under this purpose-based framework to create a measurement tool protocol for gauging the frequency of ICT use by professionals. In the next phase, more experienced disability service practitioners from diverse training backgrounds provided feedback on the overall structure, item list, and question format, leading to a substantial revision of the measurement tool. The subsequent data analysis identified the main categories underlying the broad construct of ICT adoption through exploratory factor analysis (EFA), building on the original structure developed earlier. This approach allowed the intricate list of ICT variables in the scale to be grouped into primary components (latent factors), enhancing measurement interpretability. High-quality research that delves into the major components of ICT adoption is sparse [7]. While some studies have organized ICT adoption based on functions (e.g., communication, assessment, and treatment [25]), these frameworks were typically self-developed without empirical support. This paper is the first to report how practitioners’ ICT adoption in routine disability services can be measured across distinct dimensions, backed by literature, practical insight, and empirical evidence.

By distilling the concept of ICT adoption into pivotal dimensions, the study’s secondary objective is to investigate if disparities exist in the use of different ICT dimensions, and between the current and anticipated use of ICTs across each dimension. Last, this paper will also probe how various professional and demographic characteristics, such as training backgrounds, clientele in terms of disability type, and agency types, are associated with the adoption of ICTs within each dimension. Understanding the nuanced aspects of ICT adoption with these specific needs and characteristics will empower policymakers and organizations to engage practitioners and craft more tailored strategies for ICT development in the disability sector.

Materials and methods

Design and participants

This study employed a cross-sectional survey conducted with practitioners of disability services in Hong Kong between late 2021 and early 2022, during the latter stages of the pandemic. The target participants were professionally trained practitioners who serve people with disabilities living in the community. These practitioners are responsible for designing and/or delivering interventions and are more likely to utilize ICTs in practice than those who solely support clients’ daily care needs (e.g., personal care workers) or those working with clients in inpatient settings. In Hong Kong, the target population was estimated to be around 2,000 practitioners, based on staffing needs for each type of community rehabilitation service as suggested by the Social Welfare Department [26]. However, this estimate might differ significantly from the actual number due to the autonomy of service providers in allocating resources and the increasing number of practitioners working for multiple service units or organizations simultaneously.

Purposive sampling was employed to capture the diversity within disability services. According to the government’s list of agencies offering rehabilitation services in Hong Kong [27], the two predominant sectors in the disability field are services for individuals with mental health issue (MHI) and for those with intellectual and developmental disability (IDD). As such, the recruitment primarily focused on these two disability types but also encompassed services for other sensory/physical disability (SPD), including visual, hearing, and physical disabilities. An invitation letter was dispatched to contacts within the targeted agencies, inviting all qualifying participants to engage either online via Qualtrics or through a paper survey. In total, the survey garnered valid responses from 324 practitioners. The study received approval from the Social Behavioral Research Ethics Panel of *** University (SBRE-20-705).

Measures and procedures

The core objective of this paper, which is to explore the adoption of ICTs among practitioners in disability services, follows the stepped, inclusive approach previously discussed. Specifically, for item development, the research team searched PubMed, EBSCO, ProQuest, and PsycINFO for articles on ICT adoption in the health and social care contexts. We opted to draw from the study by Feijt and colleagues on e-mental health [28] and Nanda and Ramesh’s study on ICT use among disability professionals [29], resulting in an initial item pool of 12 ICT tools. To formulate the purpose-based framework of ICT adoption in direct practice, in addition to prior research evidence, we also referenced the Case Management Manual published by the Social Welfare Department [30], which delineates the primary purposes of direct clinical services in the local context. Three broad purposes—namely, relationship building, assessment, and treatment—were identified, and the 12 ICT items were categorized under each dimension. This entire process was undertaken by the research team, supplemented by substantial advisory input from a panel comprised of a social worker, nurse, and occupational therapist, each with over a decade’s experience in disability service. As for the question format, frequency of use was selected over the proportion of work because the panel members considered it clearer and simpler to answer.

In the subsequent stage, the draft scale underwent reviews during focus group sessions and individual meetings with a diverse group of practitioners. This included social workers, nurses, occupational therapists (OTs), physiotherapists (PTs), speech therapists (STs), and welfare workers who worked with people with MHI, IDD, hearing impairments, visual impairments, and physical disabilities. These consultation meetings enabled the research team to refine the ICT tools list, its wording, and the scale’s three-dimensional structure. The first dimension comprised six items: email, WhatsApp/text messaging, videoconferencing, online service record systems, social media, and informational websites. The second dimension included online screening tools, self-monitoring apps, and wearable/biofeedback devices. The third dimension encompassed virtual/augmented reality, rehabilitation training apps/technologies, and assistive technology for individuals with disabilities. Experts recommended presenting the items in both English and Chinese to prevent misunderstandings. Two sets of questions were finalized: one querying the current actual use of each ICT item in participants’ direct interactions with those with disabilities and another concerning their willingness to use each item in future practice. A 5-point Likert scale (e.g., never, approximately 1–2 times per half year, approximately 1–2 times per month, approximately 1–2 times per week, and daily) was employed, influenced by the study conducted by Feijt and colleagues [28], and validated as appropriate by experts. Table 1 summarizes the initial pool of ICT tools and the scale measuring their adoption.

Table 1. The initial item Pool of the scale measuring adoption of ICTs among practitioners in disability services.

Item	Never	Approximately 1–2 times per half year	Approximately 1–2 times per month	Approximately 1–2 times per week	Daily
Purpose1: relationship building
1. Email	□	□	□	□	□
2. WhatsApp/text messaging	□	□	□	□	□
3. Videoconferencing	□	□	□	□	□
4. Online service record systems	□	□	□	□	□
5. Social medial	□	□	□	□	□
6. Informational websites	□	□	□	□	□
Purpose2: assessment
7. Online screening tools	□	□	□	□	□
8. Self-monitoring apps	□	□	□	□	□
9. Wearable/biofeedback devices	□	□	□	□	□
Purpose3: treatment
11. Virtual/augmented reality	□	□	□	□	□
12. Rehabilitation training apps/technologies	□	□	□	□	□
13. Assistive technology for individuals with disabilities	□	□	□	□	□

Note. The initial labels of the three dimensions were later refined to (1) “Information and Communication Tools”, (2) “Screening and Monitoring Tools”, and (3) “Treatment and Rehabilitation Tools” for greater clarity, drawing on the results of the EFA.

The demographic questions include gender, age, and education, along with characteristics about participants’ clinical practice, as shown in Table 2. This includes training backgrounds, service years, positions, clientele in term of disability types, and types of employing organizations. To ensure a comprehensive and locally-sensitive approach to these clinical-demographic variables, they were established based on feedback from and pilots with experienced practitioners in Hong Kong. Some variables underwent recoding for analysis. These include age (1 = early adult; 2 = young adult; 3 = middle-aged adult; 4 = older adult), education level (1 = diploma or below; 2 = bachelor’s degree; 3 = master’s degree or above), training backgrounds (1 = psychosocial practitioners [i.e., social workers, counselors, and clinical psychologists], 2 = rehabilitation therapists and assistants [i.e., OT, PT, ST, and their trained assistants], 3 = medical professionals [i.e., doctors and nurses], 4 = training and welfare practitioners [i.e., teachers, instructors, and welfare workers]), position (1 = executive positions; 2 = supervisory positions [i.e., center-in-charge and advanced practitioners]; 3 = frontline professionals), years of experience (1 = junior [≤3 years]; 2 = mid-level [>3 and ≤10 years]; 3 = senior [over 10 years]), and types of employing organizations (1 = non-governmental agencies; 2 = governmental agencies). Concerning the clientele’s disability type, participants were asked which group, out of the 11 types of disabilities commonly listed in Hong Kong disability policies (e.g., mental illness, intellectual disability, autism, visual impairment, etc.), they primarily work with in practice. This variable was further recoded into three categories: (1) IDD; (2) MHI; and (3) SPD.

Table 2. Clinical-demographic characteristics of participants (n = 324).

Characteristics	n (%)	Missing
Gender		1
Female	230 (71.2%)
Male	93 (28.8%)
Age (range 20–69, mean 38.7, standard deviation [SD] 9.8)
Early adult (20–25)	18 (5.8%)	14
Young adult (26–40)	176 (56.8%)
Middle-aged adult (41–55)	94 (30.7%)
Older adult (56–69)	22 (7.1%)
Education
Diploma or below	103 (31.8%)
Bachelor’s degree	101 (31.2%)
Master’s degree or above	120 (37.0%)
Occupation
Psychosocial practitioners	156 (48.1%)
Rehabilitation therapists and assistants	65 (20.1%)
Medical professionals	40 (12.3%)
Training and welfare practitioners	63 (19.4%)
Position
Executive positions	17 (5.2%)	2
Supervisory positions	53 (16.5%)
Frontline professionals	252 (78.3%)
Years of experience (range 0.25–40, mean 9.8, SD 8.1)		5
Junior (= <3)	86 (27.0%)
Mid-level (>3 and = <10)	118 (37.0%)
Senior (>10)	115 (36.0%)
Employing organization		2
Non-governmental agencies	278 (86.3%)
Governmental agencies	44 (13.7%)
Clientele (disability type)
Persons with IDD	99 (30.6%)
Persons with MHI	136 (42.0%)
Persons with SPD	89 (27.5%)

MHI: mental health issue; IDD: intellectual and developmental disability; SPD: sensory/physical disability.

Data analysis

Before the analysis, the dataset underwent a comprehensive cleaning process to ensure the accuracy and integrity of the data. Descriptive statistics were utilized to provide an overview of the participants’ clinical-demographic characteristics and to outline the distributions of the variables of interest.

For the EFA of the current state of ICT adoption, the method of categorical weighted least squares with Oblimin oblique rotation was used. EFA has been commonly considered an appropriate method for uncovering latent factors of scales [31]. When dealing with surveys using Likert-scale ratings, it is advisable to use weighted least squares accompanied by robust standard errors and an oblique rotation [32]. Items that did not achieve the .40 threshold were viewed as not being associated with any of the factors [33]. The model fit was examined by traditional model fit indices: the model chi-square, the Comparative fit index (CFI), the Root mean square error of approximation (RMSEA), and the Standardized Root Mean Residual (SRMR). The model was also examined for the same scale for measuring the intended frequency of ICT adoption in future work.

Once the underlying structure was identified with fit statistics, a subscale score (average across items) was computed for each participant, which was then utilized in subsequent analyses. A paired-sample t-test was applied to examine the differences between the use of different dimensions, as well as the current adoption and intention of use across each dimension. MANOVA, which protects against type I errors [34], was performed to assess the differences in each factor of ICT adoption by clinical-demographic variables. Pillai’s Trace was reported due to the presence of uneven sample sizes among certain independent variables and its applicability to violations of MANOVA assumptions [35]. Partial eta-squared (η²) was used to measure the effect size for MANOVA to explain the proportion of variances in the set of dependent variables. For partial η², Cohen suggested the values of 0.0099 (approximately 1%), 0.0588 (approximately 6%), and 0.1379 (approximately 14%) to define small, medium, and large effects, respectively [36]. Drawing on the MANOVA results, follow-up univariate tests were performed for independent variables with significant effects on ICT adoption. The data analysis was performed using IBM SPSS for Windows (Version 26) and JASP software package (JASP TEAM 2023). The significance level was set at 0.05.

Results

Participants

Among the 324 participants who completed the survey, the mean age was 38.7 (SD = 9.8). A majority, 71.2%, were female, and 68.2% held a bachelor’s degree or higher. Most participants identified as psychosocial practitioners (48.1%), held frontline positions (78.3%), and were affiliated with non-governmental agencies (83.9%). The average years of experience stood at 9.8 (SD = 8.1). Around 40% of the participants worked with people with MHI, while 30% worked with people with IDD and SPD, respectively. Table 2 provides an in-depth breakdown of the clinical-demographic characteristics of the participants.

Factor structure

The Kaiser–Meyer–Olkin metric for sampling adequacy was calculated to be 0.76, comfortably exceeding the recommended threshold of 0.50 [37]. The Bartlett’s test of sphericity yielded acceptable results (p < 0.001), indicating the data’s suitability for factor analysis.

Three factors presented eigenvalues exceeding 1.0. The structure identified closely aligned with the initially conceived three-dimensional structure, derived from literature reviews and expert consultations, with only a single alteration. While the loading factors for the majority of items spanned from 0.531 to 0.838, the “online screening” item did not meet the 0.40 benchmark (registering at 0.369) and deviated from the intended factor structure. Consequently, it was removed from the scale. The refined three factors were labelled as the use of (1) information and communication tools (accounting for 20.7% of variance), (2) screening and monitoring tools (accounting for 17.6% of variance), and (3) treatment and rehabilitation tools (accounting for 15.9%). The three-factor model accounted for 53.8% of variance in total. Table 3 presents the loading factor of each item included and the overall structure of the final model. The Cronbach’s alpha of the whole sample is 0.7, indicating an acceptable level of reliability. The final model shows good fit indices (χ²₂₅ = 59.5, p < 0.001; CFI = 0.95; RMSEA = 0.065; SRMR = 0.036). Examining this model against the measures of anticipated ICT adoption in future work also generated similar good fit indices (χ²₂₅ = 85.2, p < 0.001; CFI = 0.95; RMSEA = 0.086; SRMR = 0.033).

Table 3. Factor loadings of the 12 ICT items.

	Factor 1: Information and communication tools	Factor 2: Screening and monitoring tools	Factor 3: Treatment and rehabilitation tools
1. E-mail	0.701
2. WhatsApp/text messaging	0.577
3. Videoconferencing	0.590
4. Online service record system	0.588
5. Social media	0.531
6. Informational website	0.631
8. Self-monitoring apps			0.747
9. Wearables and biofeedback			0.777
10. Virtual/augmented reality		0.663
11. Rehabilitation training apps/technologies		0.817
12. Assistive technology for persons with disabilities		0.837

Note. Applied rotation method is oblimin.

Differences between the current and predictive use of ICTs in practice

Having identified the key structure underlying ICT use, the study proceeded to conduct a comparative analysis between the use of ICTs in different dimensions, and between practitioners’ actual ICT utilization and their intentions for future application. First, the current use of Factor 1 (i.e., information and communication tools) is observed to be significantly more frequent than that of Factor 2 (i.e., screening and monitoring tools; mean difference [MD] = 2.08, p < 0.001) and Factor 3 (i.e., treatment and rehabilitation tools; MD = 2.06, p < 0.001). Moreover, the paired-sample t-test indicates a statistically significant rise in the use of both screening and monitoring tools (MD = 0.28, p < 0.001) and treatment and rehabilitation tools (MD = 0.33, p < 0.001) as predicted by participants. No significant difference was found between the current and anticipated application of information and communication tools. Table 4 presents the mean scale scores of ICTs in both present and anticipated adoption with the t-test results.

Table 4. Mean scale scores of ICT adoption across each factor.

	Current practice, mean (SD)	Future work, mean (SD)	t-Test (df)	p-Value
Factor 1: Information and communication tools	3.71 (0.70)	3.72 (0.75)	−0.447 (323)	0.586
Factor 2: Screening and monitoring tools	1.63 (1.04)	1.91 (1.13)	−6.179 (323)	<0.001
Factor 3: Treatment and rehabilitation tools	1.65 (0.81)	1.98 (1.02)	−9.444 (323)	<0.001

Note. 1: never; 2: approximately 1–2 times per half year; 3: approximately 1–2 times per month; 4: approximately 1–2 times per week; 5: every day.

Effects of clinical-demographic factors on ICT adoption

For conducting MANOVAs, the current use of ICTs across the three dimensions were taken as the three dependent variables and while the eight clinical-demographic characteristics were treated as independent variables. Correlation analysis of the three dependent variables showed significant correlation below 0.4 (r), indicating no severe multicollinearity for MANOVA. The results of Box’s M-tests yielded p-values above 0.05, showing that the assumption of the equality of covariance matrices could be met.

The results of MANOVA (Table 5) revealed group differences for five variables: education, occupation, position, employing organization, and clientele (disability type). The effect sizes of these variables were general small. Medium effect sizes were only observed for occupation (partial η² = 0.082) and clientele (partial η² = 0.079).

Table 5. Results of MANOVA for the combination of ICT subscales.

Variables	Pillai’s trace	F	df	Error df	p	Partial η^2
Gender	0.008	0.902	3	319	0.440	0.008
Age	0.026	0.904	9	918	0.521	0.009
Education	0.042	2.274	6	640	0.035	0.021
Occupation	0.245	9.49	9	960	<0.001	0.082
Position	0.053	2.883	6	636	0.009	0.026
Years of experience	0.027	1.466	6	628	0.189	0.014
Employing organization	0.043	4.748	3	318	0.003	0.043
Clientele	0.157	9.104	6	640	<0.001	0.079

Note. Age was recoded into four groups and years of experience was recoded into three groups. The bold values indicate group differences revealed by MANOVA.

Obtaining a statistically significant effect from MANOVA permits more focused comparisons using univariate tests. The results indicate that, for the variable of education level, practitioners with a master’s degree or above reported a more frequent use of information and communication tools in practice than the group with a diploma or below (MD = 0.28, p = 0.009). However, no differences among groups were observed for other aspects of ICT adoption. In terms of their occupational training backgrounds, no group differences were identified in the use of information and communication tools, or screening and monitoring tools. A significantly higher level of adoption of treatment and rehabilitation ICTs was observed among rehabilitation practitioners (e.g., PT and OT) compared to all other groups, including psychosocial practitioners (MD = 0.94, p < 0.001), medical professionals (MD = 0.95, p < 0.001), and training practitioners (MD = 0.60, p < 0.001). Furthermore, results show that training practitioners (e.g., teachers and instructors) used treatment and rehabilitation ICTs more frequently than psychosocial practitioners (MD = 0.34, p = 0.011).

In terms of employment conditions, practitioners in executive positions used screening and monitoring tools (mean difference = 0.73, p = 0.014), and treatment and rehabilitation tools (mean difference = 0.28, p = 0.042) much less frequently than frontline practitioners. Participants employed by governmental agencies demonstrated a higher adoption rate of treatment and rehabilitation tools than those working in non-governmental agencies (MD = 0.46, p = 0.004), while other aspects of ICT adoption remained consistent between the two groups.

Last, the type of clientele, in terms of disability conditions, was found to have significant effects only on the adoption of treatment and rehabilitation ICT tools. A significantly higher usage was observed among practitioners working with individuals with SPD, compared to both professionals working with IDD (MD = 0.40, p = 0.001) and with those with MHIs (MD = 0.72, p < 0.001). Participants working with individuals with IDD also exhibited a greater adoption rate of treatment and rehabilitation tools than those working with MHI (MD = 0.32, p = 0.005).

Discussion

The study reports the findings of a cross-sectional study examining the adoption of ICTs within routine disability services by practitioners. First, it showcases a stepped, inclusive strategy for addressing the intricate nature of ICTs in empirical research aimed at gauging their adoption frequency in practice. This method was informed by an initial conceptual framework that was developed through a review of existing literature and consultations with experts during the preparatory phase, which subsequently informed the construction of the scale used in the data analysis stage. Although EFA led to minor modifications in the subscale compositions, the final model remained largely consistent with the initial framework. Establishing this factor structure for ICT adoption facilitates the computation of subscale scores for further analysis—details of which will be discussed later. Measuring ICT adoption with a structured scale of ICT examples also provides respondents with a more comprehensive and practical understanding of the concept. This, in turn, aids in the completion of the subsequent questions related to readiness for ICT adoption and support factors in this study. Therefore, it is recommended that future studies focusing on ICTs or their equivalents in domains, such as e-health and telerehabilitation consider adopting similar methodological approaches.

It is vital to emphasize that the study goal was not to validate an ICT adoption scale for wider application. Given the fast-paced nature of ICT evolution, items on the current scale could quickly become obsolete. Another critical consideration is the diverse contexts; the suggested pool of ICT items might not be suitable for disability services across different geographical landscapes. Applicability largely hinges on the maturity of digital technology and the penetration of the internet and electronic products in particular local environments. Nevertheless, this research outlines an approach to devising an up-to-date, localized measurement tool for assessing ICT adoption. It emphasizes the merit of involving experienced practitioners throughout the research design, ensuring that the diversity of disciplines and clientele in disability services is represented. This approach can more comprehensively and accurately capture the potential purposes and types of ICTs to be measured in local research.

In exploring the disparities between present and anticipated ICT usage across the three dimensions, the findings indicate a relatively high level of adoption, predominantly concerning information and communication tools. This is not only in line with recent research that reported a significant surge in ICT usage during the COVID-19 pandemic [16,17,22], but also offers a nuanced understanding of which types of ICTs have driven this rapid increase. Notably, in a study carried out in Netherlands during the pandemic, Feijt documented a marked rise in the utilization of “basic” tools, such as video conferencing, email, and text, with mean scores between 3 and 4 [22], using the same 5-point Likert scale employed in this study. The result is akin to the mean value of the Factor 1 (i.e., communication and information tools) in this study. This suggests that the recent public health crisis has substantially contributed to the integration of ICTs into routine practice globally, particularly endorsing tools that bolster accessibility amidst service interruptions. Interestingly, study findings further reveal a significant increase in the anticipated use of the more “innovative” technologies for screening and monitoring, as well as treatment and rehabilitation, while the integration of more common communication-based ICTs appears to have reached a satisfactory level. This suggests that future attention and investment in ICT adoption in the disability field should focus more on domains directly relevant to innovative, clinical, and rehabilitation practice, tailoring strategies to the specific needs of the field.

The study also identified small-to-medium level differences in the current use of ICTs between groups based on certain clinical-demographic variables. Examining each dimension of ICT adoption, the study found that these group differences predominantly pertained to the use of treatment and rehabilitation tools (factor 3). Practitioners’ occupation, job position, employing agencies, and clientele were found to influence the adoption of these ICT tools. Rehabilitation and training practitioners, such as PTs, OTs, and teachers, used these ICTs more frequently than psychosocial and medical practitioners. This may suggest that the current forms of ICT are more compatible with physical and occupational rehabilitation than with psychosocial and medical treatment. However, in the delivery of disability services, psychosocial and medical practitioners are often expected to take a case management approach to ensure holistic support for service users [30,38]. Hence, it would be beneficial for them to have a comprehensive understanding of interdisciplinary, innovative treatment options to facilitate client service planning. Therefore, this study advocates for further research to explore the knowledge, motivation, expectations, and confidence of psychosocial and medical practitioners in employing ICTs during the treatment phase in service services.

The second variable leading to differences in the use of treatment and rehabilitation tools was one’s job position. Practitioners in executive positions used these tools, as well as screening and monitoring tools, much less frequently than frontline practitioners. This observation aligns with the common understanding that executives usually have fewer direct contacts with service users, focusing more on management and administration. However, as higher-level practitioners often oversee resource allocation, it is imperative for them to stay updated on the rapidly evolving field of digital health and to remain informed about the needs of frontline services. Recent research has also highlighted the crucial role that managers play in the innovation and digital transformation of healthcare organizations [39]. Therefore, this finding is a great reminder that as they may not utilize these ICTs frequently, such knowledge is crucial for integrating innovative digital interventions without misdirecting social resources on tools and devices that do not align with frontline practice.

The use of treatment and rehabilitation tools also varied among practitioners employed by different types of agencies. Governmental agencies, such as hospitals and publicly funded special schools, exhibited more frequent use in routine services. This could be attributed to these agencies often having a more stable source of funding than non-governmental agencies, especially during economic downturns, enabling them to invest in pricier equipment, such as VR and assistive devices. This highlights the significance of resource support for non-governmental agencies looking to integrate ICTs in this area.

Another pronounced difference was based on clientele, with the most frequent adoption observed among practitioners primarily working with people with SPD and the least among those mainly serving individuals with MHI. This once again underscores the compatibility of these rehabilitation-and-treatment-related ICTs with physical/occupational therapy as delivered by practitioners, such as OTs and PTs, which often play a pivotal role in the rehabilitation plans for individuals with visual impairments, hearing impairments, and physical disabilities. Although a variety of treatment-related ICTs have been developed for mental health treatment [40] and a substantial body of research exists on their effectiveness [41,42], the findings indicate that their utility was relatively low, even during the pandemic. This calls for continuous promotion activities (e.g., exhibition and on-site workshop) of ICTs for mental health professionals, to showcase the use of ICTs in routine services and convince them of the advantages that innovative tools can offer to treatment outcomes.

In addition to treatment and rehabilitation ICTs (factor 3), the study observed that the relatively poor adoption status of screening and monitoring tools (factor 2) showed no significant variances based on almost all clinical-demographic factors, with the exception of the variable of job position. This result may not be surprising, as these types of ICT tools (e.g., self-monitoring apps, wearables, and biofeedback) often require end-users to possess a certain level of operational ability and the persistence to continue their use. Moreover, although digital assessment and monitoring can offer advantages for disability services, the lack of standardized regulations or guidelines for wearables and self-monitoring tools in practice raises concerns, such as ambiguity in data protection [43,44] and potential conflict of interests associated with stakeholders [45]. This undermines practitioners’ evaluations of their usefulness. Further research on the adoption of these tools is needed to guide their safe and effective use in frontline practice and to develop appropriate functions and interfaces, thereby increasing accessibility for persons with disabilities.

Last, findings suggest that practitioners’ demographic factors have only a minor relationship with their use of ICTs in routine practice. Only a more frequent use of information and communication tools was noted among practitioners with a master’s degree or above compared to those with a diploma or below. This aligns with the extant literature indicating that a higher education level often predicts greater digital competencies and a stronger willingness to use technology [16,46]. However, regarding gender and age, the study found no influence on the use of ICTs across all three dimensions. Concerning these two factors, the existing literature offers inconsistent results. Some studies pinpointed a low adoption of ICTs among older, experienced practitioners [16,46,47], and females [16,48], while others found a higher ICT affinity among senior practitioners [49] and females [49,50]. There are also some studies that identified no differences in the intention to use ICTs between age [51] and gender groups [52]. This research also suggests that gender, age, and service years are not significant factors influencing the level of ICT use across all dimensions. This implies that promotional and in-service training activities for ICTs in disability services may not need to target a specific gender and age group among practitioners.

Limitations

The study has some limitations. First, a non-probability sampling method was employed due to constraints in resources and sampling frames in the local field. Although potential sampling bias cannot be eliminated, the study attempted to capture the diversity of practitioners in disability service through a purposive sampling process. For a target group with such a high level of diversity, the sample size (i.e., 324) of this study might appear insufficient. However, theoretically, it provides adequate power for the statistical analyses conducted in this study. Future research with more resources and networks should adopt a larger, more representative sample to enhance the external validity of study findings. Second, this study uses a cross-sectional design to inquire about participants’ current frequency of ICT use and their anticipated future use. While many studies in this area have adopted a cross-sectional design, it is important to recognize that such research can only predict, not capture, the real changes in ICT adoption over time. Last, the scale developed to measure the use of ICTs, and the three-factor model generated by the study, may have limitations when applied in contexts outside of Hong Kong or the disability field. It is worth reiterating that the primary objective of this study isn’t to validate any scales but to offer a reference for future research aiming to manage potentially inconsistent interpretations of ICT and its related concepts in measurements.

Conclusion

In examining the adoption of ICTs by practitioners within routine disability services, this study introduces a methodical, inclusive methodology for quantifying the multifaceted use of ICTs and its associated realms, such as e-health and telerehabilitation. Data analysis generated a three-factor model, condensing variables that gauged the utilization of 12 ICT tools into three components: (1) information and communication tools, (2) screening and monitoring tools, and (3) treatment and rehabilitation tools. Subsequent analysis pinpointed a predominant use of ICTs pertaining solely to information and communication tools during the advanced phases of the pandemic. However, there was an anticipation among practitioners for an augmented use of the remaining ICT categories in future practice. It is imperative that future endeavors and investments targeting ICT development in the disability sector concentrate on domains intrinsically linked to innovative treatment and rehabilitative practices, thereby calibrating strategies to align with the sector’s unique demands. Furthermore, the research provides fresh perspectives on the modalities of training and facilitating ICT integration in disability services, emphasizing the impact of variables, such as practitioners’ occupational roles, clientele job positions, employing agencies, and educational levels on their prevailing ICT usage. For the pragmatic application of these insights, there is a pressing need to bolster ICT training, particularly for psychosocial and medical professionals, and to amplify investments in non-governmental entities. This reinforced support should adopt a gender- and age-inclusive approach, specifically crafted to address the requirements of practitioners across varying hierarchical positions.

Acknowledgements

I thank all colleagues who facilitated the implementation of the study and participants who offered their views in the survey.

Disclosure statement

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

Additional information

Funding

This study was financially supported by the Madam Tan Jen Chiu Fund.