Disability-Inclusion Training in a Financial Institution: A Descriptive Implementation Case Study from Bangladesh

1. Introduction

Access to appropriate financial services supports autonomy, economic participation, and resilience to financial shocks [1]. People with disabilities, however, remain disproportionately excluded from formal banking and other financial services, particularly in low- and middle-income countries [2,3]. Reported barriers include discriminatory assumptions about decision-making and creditworthiness, inaccessible branches and automated teller machines, communication barriers, rigid identification and authentication procedures, inaccessible information, and digital platforms that do not work reliably with assistive technologies [4,5,6].

These barriers rarely occur in isolation; a screen reader, accessible mobile application, Braille information, or video-based sign-language service can only support participation when staff understand how to offer it, organizational procedures permit its use, and customers can access the wider service pathway. The World Health Organization and United Nations Children's Fund describe assistive technology as encompassing assistive products and their related systems and services. Consistent with this definition, the WHO 5P framework emphasizes people, policy, products, provision, and personnel as interdependent elements of access [7]. A recent Assistive Technology editorial similarly encourages adoption of a common WHO-based definition [8]. An assistive-systems lens is therefore relevant to accessible banking without implying that banking itself is assistive technology: it draws attention to the human and organizational conditions that determine whether accessible and assistive technologies can be used in practice.

Workforce training is commonly proposed as a response to attitudinal and communication barriers. Evidence from disability training in other service sectors suggests that training can improve knowledge and confidence, particularly when people with disabilities contribute directly to its design and delivery [9]; however, transfer into sustained workplace behavior is less certain and depends on opportunities to practice, leadership support, reinforcement, and aligned organizational systems [10]. Recent financial-inclusion research similarly identifies a need for staff capability alongside institutional policies, accessible products, and disability-disaggregated monitoring [2,3]. Evidence of how financial institutions implement and evaluate disability-inclusion training remain limited.

Bangladesh provides an important setting for examining this issue, a nation of more than 177 million people. National estimates vary according to the definition and measurement approach used: the 2021 National Survey on Persons with Disabilities reported that 7.14% of the population experienced at least one functional difficulty, while 2.80% met the survey's government-defined disability measure [11]. People with disabilities in Bangladesh have described attitudinal, procedural, and accessibility barriers to microfinance and digital financial services despite national disability-rights legislation and international obligations under the Convention on the Rights of Persons with Disabilities [5,6,12]. To our knowledge, this is among the first published case studies describing the collaborative development, delivery, and early evaluation of disability-inclusion training within a commercial financial institution.

This study aimed to: (1) describe the collaborative development and delivery of a disability-inclusion training intervention within a commercial bank in Bangladesh; (2) report immediate and follow-up participant outcomes; and (3) document organizational responses reported during the follow-up period. The training was the intervention. Organizational activities occurring after training were treated as exploratory downstream outcomes rather than intervention components or evidence of causal impact.

2. Materials and Methods

2.1. Study Design

This study used a descriptive, single-site implementation case-study design incorporating a single-group pre-training and post-training evaluation and follow-up, with questionnaire responses collected approximately 9-10 weeks after training. The case was bounded by the development and delivery of a three-day training program and the first approximately three months after delivery. Reporting was informed by the Template for Intervention Description and Replication and the Standards for Reporting Implementation Studies [13,14]. The study was designed to describe implementation and preliminary outcomes, not to test intervention effectiveness.

2.2. Setting and Implementation Partnership

The intervention was implemented with Prime Bank PLC, a commercial bank in Bangladesh. At the time of planning, the bank reported operating 147 branches and 153 automated teller machines across 140 locations. The bank provided the organizational setting, participants, operational knowledge, and senior-leadership engagement. The case study focused on staff from central divisions and branch-facing functions and on organizational responses reported during the follow-up period after training.

The implementation partnership brought together Prime Bank PLC, Team Inclusion Bangladesh, and Monash University. Team Inclusion Bangladesh contributed local disability-rights expertise and lived-experience leadership; Monash University contributed expertise in financial inclusion, organizational behavior, instructional design, and evaluation. The partners jointly developed and delivered the intervention.

2.3. Formative Development and Theoretical Frameworks

Formative development occurred before the training and included separate group discussions with banking professionals and people with disabilities, a facilitated joint dialogue to identify shared barriers and possible solutions, a baseline survey administered using KoboToolbox, and iterative planning meetings. These activities informed the curriculum, examples, and delivery methods.

Three complementary frameworks were used for distinct purposes. The Analysis, Design, Development, Implementation, and Evaluation (ADDIE) model structured the instructional-design process [15]. The Capability, Opportunity, Motivation-Behavior model informed the selection of activities intended to strengthen knowledge and skills, create opportunities to practice inclusive service, and increase motivation to change behavior [16]. Kirkpatrick's four-level model organized the intended evaluation domains of reaction, learning, workplace behavior, and organizational results [17]. The follow-up questionnaire explored participant-reported workplace application, but no independent observation or objective measure of behavior was undertaken. These frameworks guided development and reporting; they were not tested as explanatory models.

2.4. Participants

Twenty-five employees participated in the training. Eligibility criteria were developed collaboratively by the research and training team and provided to Prime Bank's Human Resources department. Recruitment used a combined purposive and voluntary approach. Prime Bank nominated staff members whose roles aligned with the eligibility criteria, while an internal noticeboard announcement invited expressions of interest from employees wishing to participate. Human Resources reviewed all nominated staff and volunteers against the predefined eligibility criteria before invitations were issued. Consequently, the final cohort comprised both bank-nominated participants and employees who volunteered following the internal announcement. Regardless of recruitment pathway, participation depended upon individual interest, willingness to participate, and fulfilment of the eligibility requirements.

2.5. Training Intervention

The intervention was delivered face-to-face over three consecutive days, 23-25 May 2025, at an accessible venue in Dhaka. The training was delivered from Friday to Sunday. Friday and Saturday were weekly holidays in Bangladesh, while Sunday was a normal working day. Consequently, two training days occurred outside routine working hours and one during normal work time. Participants did not receive financial incentives from the research team for participation.

People with disabilities participated in the program’s design, acted as facilitators, trainers, and role-players. The training combined rights-based content with disability-led experiential and applied learning. Table 1 summarizes the principal components. Materials were provided in digital and large-print formats, and sign-language interpretation supported sessions involving Deaf facilitators. The venue was selected to accommodate wheelchair users. On the first day, participants who did not routinely use wheelchairs used them while moving through the training environment and reflected on physical and procedural barriers. On the third day, people with disabilities took simulated banking roles while participants experienced service scenarios in which accommodations were deliberately absent. The program concluded with individual and group action planning.

2.6. Data Collection

Immediate participant outcomes were assessed using the complete study-specific electronic questionnaires administered before and after the training through KoboToolbox (Supplementary Files S1 and S2). The questionnaires were internally reviewed and pilot tested, but were not psychometrically validated. They primarily assessed perceived familiarity, confidence, and frequency of considering inclusion rather than objective knowledge, demonstrated skill, or a validated psychological construct.

Four parallel five-point items were used for direct pre/post comparison: (1) self-rated familiarity with disability types and rules or policies for inclusive banking (Q2; 1 = not familiar at all to 5 = very familiar); (2) confidence identifying and addressing physical, digital, or communication barriers (Q3; 1 = not confident at all to 5 = very confident); (3) confidence supporting a customer with a disability to use ATMs or mobile banking or complete forms (Q5; 1 = not sure at all to 5 = very sure); and (4) frequency of considering disability inclusion during daily customer service (Q8; 1 = never to 5 = always). The pre-training Q5 form displayed 'Quite sure' for both response categories 4 and 5, although the scale instructions defined category 5 as 'Very sure'; categories 4 and 5 were therefore combined for descriptive reporting. The multi-select barriers item (Q7) was not treated as a scored outcome because it assessed respondents' perceptions and did not have a defensible correct-answer threshold. Questions 10 and 11 assessed views about hiring and working with people with disabilities but were not sufficiently equivalent across the pre- and post-training forms for direct comparison. No validated empathy measure was administered. Post-training questions also asked which training components participants found useful and what personal action they intended to take.

During the planned three-month follow-up period, the same cohort was invited to complete a separately designed questionnaire (Supplementary File S3). The questionnaire included retrospective open-ended prompts; a five-category item whose stem referred to comfort and whose response labels referred to confidence in initiating conversations and interactions with people with disabilities; a five-category compound item on using inclusive language and avoiding assumptions about capabilities; and open-ended items on commitments, intended timing, and perceived impact. It did not repeat the principal immediate pre/post items and was not used to assess maintenance of those outcomes.

The contemporaneous evaluation report recorded 21 responses, but the archived KoboToolbox extraction available for reanalysis contained 20 participant-level records. Attempts to locate or reconcile the additional record were unsuccessful. Analyses were therefore restricted to the 20 retrievable records, and no values were imputed.

Organizational responses were extracted from a three-month evaluation report, follow-up interviews conducted by the implementation partner, internal policy material supplied by the bank, and a documented next-step plan. Each response was classified as reported operational, in progress, proposed, or planned. No systematic pre-training organizational baseline was available, and implementation dates, coverage, fidelity, uptake, and customer outcomes were not independently verified. These data were supplied by implementation partners and were not independently audited.

2.7. Analysis

Quantitative data were summarized using counts, percentages, and reported means. For Q2, Q3, and Q5, ratings 4 or 5 were grouped as an upper-category response. This grouping was used for descriptive presentation, was not a validated competency threshold, and the available contemporaneous records did not document that it had been specified before analysis. Q8 was summarized using the reported group mean on its 1-5 scale. Available-case denominators are reported, and no missing values were imputed.

The source data available for manuscript analysis contained aggregate pre/post counts and did not permit reliable participant-level linkage; therefore, paired analyses, effect sizes, and confidence intervals could not be calculated. Given the small single-site cohort, ordinal study-specific measures, and absence of linkable paired records, analyses were restricted to group-level descriptive statistics. Follow-up categorical items were summarized from the 20 retrievable records. Open-ended responses and follow-up interviews were used by the implementation team to contextualize the descriptive findings, but no independent qualitative analysis was undertaken for this manuscript. Accordingly, quotations are not presented as thematic evidence.

2.8. Ethics

The Monash University Human Research Ethics Committee approved the study (Project No. 2025-42969-119271). All participants provided informed consent before their evaluation data were used for research. Individual responses were reported in aggregate, and no role-labelled quotations are presented because the small cohort created a risk of deductive identification.

3. Results

3.1. Participant Characteristics

Twenty-five employees completed the training. Participants represented senior and mid-level roles across multiple divisions (Table 2). The cohort included 11 participants at vice-president level or above and 14 executive or officer-level participants.

3.2. Intervention Delivery and Participant Reaction

All planned training days were successfully delivered with full attendance. Role-play, communication practice, sign-language instruction, lived-experience contributions, and group case work were among the components most frequently selected as useful in the post-training evaluation. One multi-select chart in the source report recorded more selections than participants; because participants could select multiple components, those frequencies were not interpreted as unique participant counts.

3.3. Immediate Participant Outcomes

The contemporaneous evaluation provided group-level pre/post summaries for three parallel five-point self-report items and the frequency-of-consideration item (Table 3). The proportion selecting response categories 4 or 5 for self-rated familiarity with disability types and inclusive-banking rules or policies was 4/25 (16%) before and 21/25 (84%) immediately after training. The corresponding proportions were 12/25 (48%) and 23/25 (92%) for confidence identifying and addressing physical, digital, or communication barriers, and 8/25 (32%) and 20/25 (80%) for confidence supporting customers with disabilities. The group mean for frequency of considering disability inclusion during daily customer service was 2.5 before and 4.4 immediately after training on a five-point scale. These measures were study-specific self-reports and should not be interpreted as objective knowledge, demonstrated skill, validated empathy, or competency. Because participant-level records could not be reliably linked, these are group-level descriptive comparisons rather than within-person estimates of change.

3.4. Follow-Up Participant-Reported Observations

Participant-level follow-up data were available for 20 of the 25 training participants (80% of the cohort). Five respondents (25%) selected 'moderately confident', 13 (65%) selected 'very confident', and two (10%) selected 'extremely confident' in response to the item concerning comfort/confidence initiating conversations and interactions with people with disabilities. Overall, 15/20 respondents (75%) selected very or extremely confident.

Self-reported frequency of using inclusive language and avoiding assumptions about the capabilities of people with disabilities varied. Nine respondents (45%) selected 'sometimes', five (25%) selected 'often', three (15%) selected 'always', two (10%) selected 'rarely', and one (5%) selected 'never'. Because no comparable item was administered before or immediately after training, these responses describe reported practices at follow-up and cannot establish improvement, maintenance, decline, or sustained transfer into routine behavior. Open-ended responses included retrospective perceptions, commitments, reported actions, and intended actions; they were not independently verified and are not interpreted as observed behavioral outcomes.

3.5. Reported Organizational Responses Following the Training

The follow-up records described several organizational responses after training, although their implementation status and temporal relationship to the intervention varied (Table 4). An HR representative reported that disability-inclusive wording had been added to a proposed Talent Acquisition Policy. Video-call sign-language support and Braille materials were reported as available at follow-up; their initiation dates, coverage, quality, and uptake were not independently verified. Digital accessibility and onboarding improvements were reported as commencing.

Wider training rollout, a train-the-trainer pathway, dedicated Disability Inclusion Officers, annual accessibility budgets, inclusion dashboards, and disability-led accessibility audits appeared in recommendations or next-step plans. They were therefore classified as proposed or planned, not implemented outcomes.

4. Discussion

This case study documents the implementation of a three-day disability-inclusion training intervention and the participant and organizational observations reported during the follow-up period. Immediately after training, the group-level proportions in the upper response categories were higher for self-rated familiarity with disability types and inclusive-banking rules or policies, confidence identifying and addressing barriers, and confidence supporting customers; the group mean for frequency of considering disability inclusion was also higher. At follow-up, most respondents reported high comfort/confidence initiating interactions, while the compound measure of inclusive-language use and avoidance of assumptions varied. Because the follow-up instrument did not repeat the immediate measures, maintenance of the immediate findings could not be assessed. Organizational records described a proposed recruitment-policy change, communication and accessible-information services reported as available, and digital accessibility work underway; several additional actions remained plans rather than completed changes.

The intervention's strongest feature was the central role of people with disabilities in formative consultation, collaborative design, facilitation, practical instruction, and role-play. Disability-led delivery moved the program beyond generic awareness content and connected banking scenarios to lived experience and local communication practices. Evidence syntheses in other service sectors similarly suggest that direct involvement of people with disabilities can improve the relevance and credibility of workforce training, although evidence for sustained behavior and service outcomes remains limited [9]. This case also illustrates the practical value of assigning distinct functions to ADDIE, COM-B, and Kirkpatrick rather than treating them as interchangeable frameworks: ADDIE organized development, COM-B shaped the intervention logic, and Kirkpatrick organized evaluation.

The higher immediate post-training group-level self-ratings and the variable follow-up reports concern different constructs measured with different instruments and should not be interpreted as a longitudinal pattern of initial gain followed by incomplete transfer. The broader training-transfer literature nevertheless indicates that learning acquired in a workshop does not automatically become routine behavior; transfer is influenced by opportunities to practice, supervisory and peer support, task relevance, and reinforcement [10]. In the present case, limited everyday contact with customers with disabilities was reported as one reason that new terminology was not used consistently. Future implementation should therefore plan transfer supports from the outset, such as brief refresher modules, accessible job aids at service points, manager prompts, peer coaching, observation with feedback, and structured disability-led service walkthroughs. These activities would address opportunity and reinforcement rather than relying on individual motivation alone.

The wheelchair-based exercise warrants particular caution. Disability simulations can produce distress, reinforce assumptions about incapacity, or encourage participants to equate a brief constrained experience with the complexity of living with disability [18]. A meta-analysis found equivocal effects of mobility simulations on attitudes and environmental-accessibility outcomes [19]. In this program, the activity occurred alongside disability-led facilitation, environmental reflection, and later reverse role-play; these contextual elements may have directed attention toward barriers rather than impairment. Nevertheless, future iterations should explicitly frame such activities as environmental accessibility audits, state that they cannot reproduce lived experience, and prioritize disabled-led walkthroughs and co-analysis over attempts to simulate disability.

The findings also clarify the relevance of the study to assistive technology. The intervention was not an assistive-technology intervention in the narrow sense, and banking should not be relabeled as assistive technology. Rather, the training addressed the personnel and process conditions that influence whether accessible and assistive technologies can function within a service system. Screen-reader-compatible digital banking, accessible authentication, Braille or structured electronic information, and video-based sign-language communication can fail when staff are unaware of them, procedures prohibit flexible use, or customers must rely on others to complete transactions. This interpretation aligns with WHO's systems framing of assistive technology and with evidence that digital financial applications often contain barriers in labeling, navigation, authentication, and compatibility with assistive technologies [5,7,8].

The organizational findings should be interpreted as early implementation signals rather than results-level proof. The most concrete policy response was wording reported in a proposed recruitment policy; it should not be described as an adopted policy until formally approved. Similarly, sign-language video support and Braille materials were reported as available, but the study did not verify coverage, quality, uptake, or whether they pre-dated the training. Digital accessibility work was reported as underway without user testing or conformance assessment. Future evaluation should distinguish implementation outcomes such as adoption, fidelity, penetration, and sustainability [20]; retain documentary evidence of policy approval; and test digital services with people who use screen readers, magnification, alternative input, and sign language. WCAG 2.2 provides a current technical reference point for web content, but conformance testing should complement, not replace, user testing [21].

For financial institutions, the principal implementation lesson is that training can be a useful initiating strategy but is unlikely to be sufficient by itself. Organizational readiness, leadership commitment, resources, accountability, and operational opportunities influence whether staff can act on new knowledge [22]. A stronger next phase would combine refresher training with approved policy, accessibility audits, accessible-service protocols, disability-inclusive performance indicators, complaint and feedback mechanisms, and governance involving people with disabilities. Customer outcomes should be prospectively specified and measured, including independent service use, privacy, completion of transactions, time and assistance required, complaints, satisfaction, and equitable access to products.

5. Conclusions

A collaboratively designed, disability-led, three-day training intervention was feasible to implement within a commercial bank in Bangladesh and was followed by higher immediate group-level self-ratings, high follow-up comfort/confidence among respondents, and several reported organizational responses. The immediate measures were study-specific and unpaired, and the separately designed follow-up did not assess sustained effects. Self-reported use of inclusive language and avoidance of assumptions varied, and some prominent organizational actions were still proposed rather than implemented. The study supports the use of training as one initiating strategy within a wider accessibility program, while underscoring that training alone cannot establish accessible financial services. Longer-term, independently verified evaluation involving customers with disabilities is required to determine whether participant learning translates into sustained behavior, accessible technology use, equitable service delivery, and improved financial participation.