The United States Bureau of Labor and Statistics, American Time Use Survey reported that employed persons working from home doubled from 2019 to 2020, increasing from 22 to 42% (1). De Meulder et al. (2) conducted three online “living” surveys from April to July 2020 about sign language interpreting in times of coronavirus disease 2019 (COVID-19), with each iteration of the surveys incorporating respondent feedback. In the first survey, 27% of respondents reported working remotely only occasionally prior to the pandemic, while 60% reported never working remotely. The second survey demonstrated a dramatic shift from onsite to remote interpreting with 0% of participants reporting remote work in the last 6 months of 2019 to 100% of participants in April 2020. The respondents predicted the continuation of remote work in the future, with the majority indicating that 25% of their workload would remain remote. Across all three surveys, 63% of the respondents reported never receiving any training on remote interpreting (2). Other recent studies revealed that 64% of interpreters had never worked remotely prior to the pandemic and that 46% did not feel comfortable or prepared for the transition to remote work. Regarding available resources for remote interpreting, only 18% indicated that workshops and/or training were provided by their employer (3). The shift in the overall workforce from onsite to remote due to the pandemic and lack of experience and training of sign language interpreters working remotely beckons the need for examination of interpreters' perceptions of the determinants of implementation behavior upon shifting from onsite to remote work. Because the demand for remote interpreting is expected to continue, such research will be beneficial in facilitating the transition of future interpreters.

In an effort to more effectively implement innovations and evoke change, several studies have recognized the importance of understanding how to define and measure behavior change (4–7). The Theoretical Domains Framework (TDF) was created by an interdisciplinary group of psychological theory, health services research, and health psychology experts (4). The original TDF conceptualized behavior into 12 domains with 100 constructs (4). Cane et al. (5) updated the domain structure by refining and expanding the original 12 behavior domains to 14 with 84 constructs (5). Huijg et al. (6) developed a TDF questionnaire to measure behavior change based on the original 12-domain version (4) and assessed its psychometric properties (8). The Determinants of Implementation Behavior Questionnaire (DIBQ) was used to measure healthcare professionals' perceptions of factors influencing implementation to effectively integrate innovations into routine healthcare practice (specifically physical therapists in Sweden and Denmark delivering various physical activity interventions to patients with arthritis, chronic obstructive pulmonary disease, diabetes, or obesity) (8). Some items were modified from previously published questionnaires (9, 10), while other newly developed items were informed by the literature (4) and results of past studies (6, 11). The DIBQ was expanded from 12 to 18 domains with 26 constructs and reduced from a total of 100 to 93 items (8).

The DIBQ (8) has been previously adapted to measure expectations of the implementation process for a low back pain program in primary care not yet delivered in practice (12). With the aim of including a realistic number of items and evaluating the implementation process at organizational, individual, social, and contextual levels, content experts in clinical or methodological research with backgrounds in the musculoskeletal system or implementation research were asked to rate the relevancy of the 93 DIBQ items. After undergoing confirmatory factor analysis, a 26-item, 10-domain tailored version unique to implementing a low back pain program (DIBQ-t) was developed. The example by Ris et al. (12) demonstrates that the DIBQ can be successfully adapted to a different context and used to guide the examination of interpreters' perceptions upon transitioning from onsite to remote interpreting in this study.

The goal of this study was to examine sign language interpreters' perceptions and experiences of the determinants of remote interpreting implementation from home during the COVID-19 pandemic. To accomplish this, we adapted the DIBQ into the Determinants of Implementation Behavior Questionnaire for Sign Language Interpreters (DIBQ-SLI) to assess the perceptions of contextual determinants of implementation behavior of sign language interpreters throughout the changing work conditions of the pandemic as defined in the TDF (4, 5, 8). We hypothesized that sign language interpreters across the primary interpreting settings of staff, educational, community/freelance, video remote, and video relay (Table 1) would present with differing experiences of the implementation process.

The goal of this study was to examine sign language interpreters' perceptions of the determinants of remote interpreting implementation from home during the COVID-19 pandemic. A descriptive analysis of the determinants of implementation behavior was performed, and the findings were compared across interpreting settings. The video relay interpreters had the most favorable ratings for skills, self-efficacy, perceived behavior control, and innovation strategies. The staff interpreters had the most favorable ratings for innovation characteristics and organizational resources and support, while the knowledge and social support constructs were rated most favorably by the community/freelance interpreters. Outside of the fairly equivalent distribution of ratings across settings for social support, the results from this study were in support of our hypothesis that interpreters across settings would present with differing experiences of the implementation process. Interpreters in the video relay setting consistently had more positive perceptions than interpreters in the educational setting across all significant aggregate determinants of implementation behavior constructs. The staff interpreters also had higher innovation characteristics, organizational resources, and support, and innovation strategies construct ratings than the educational interpreters. The video remote interpreters also had higher skills and innovation strategies when compared with the educational setting. Higher ratings may signify that conditions in the interpreting setting were more conducive for the remote interpreting implementation behavior, but not necessarily. Our ability to detect causation is limited by the cross-sectional design; however, this research can now inform future studies with sign language interpreters exploring similar perceptions of behavior change.

To provide a broader understanding of the interrelatedness across the DIBQ-SLI, we conducted PCA for data reduction and affirmation of the most prominent constructs and items. Because multiple items from different DIBQ-SLI constructs loaded on one component, differences across interpreting settings in the component scores reveal differences in the underlying dimension shared by these items. Component loadings in this work were higher than those reported in the DIBQ-t (12), possibly indicating greater interrelatedness of the constructs and items within the dimensions for ensuring positive perceptions of remote interpreting implementation compared to implementation of a low back pain program. We found that items most pertinent to change in work conditions for sign language interpreters during the pandemic clustered on components consistent with what was selected by Ris et al. (12) to measure expectations of the implementation process (12). We expected the components to represent the organizational, individual, and social dimensions. If these dimensions did not appear in the PCA, it would call into question the construct validity of the DIBQ-SLI items for the interpreter population.

The first principal component resembled an organizational dimension, explained the greatest amount of variance in the DIBQ-SLI, and demonstrated a substantial impact on interpreters' perceptions and experiences of the determinants of implementation. The organizational dimension revealed significant differences across interpreting settings; specifically, the staff and video relay settings had more positive perceptions than both the educational and community/freelance settings. Educational interpreters who responded to the survey were not receiving sufficient managerial support, training, materials and resources, experience with remote interpreting before having to commit, financial reimbursement, and insight when compared with staff and video relay interpreters, to believe that it is possible to tailor interpreting services to be delivered remotely. Strategies for improved perceptions of remote interpreting implementation in the educational setting may include ensuring an administrative infrastructure devoted to addressing the concerns of the educational interpreting team. Educational interpreting administrators might consider inquiring with staff interpreting departments and video relay service providers and, as appropriate, emulate a similar administrative infrastructure. Due to the self-employed nature of the community/freelance interpreters, they are unable to reap a similar support infrastructure when compared to the staff and video relay settings. Community/freelance interpreters may be able to dialogue openly with one another and interpreter referral agencies with whom they maintain independent contracts to ensure that organizational-level standards are beneficial for all.

The second principal component resembled an individual dimension within the DIBQ-SLI, which was compared across settings. Interpreters' beliefs in their abilities to transition from onsite to remote regardless of what other interpreters were doing, as well as interpreters' beliefs in their interpreting skills and in their abilities to practice delivering sign language interpreting remotely, were seemingly critical. Interpreters in the video relay setting likely had more favorable conditions across the self-efficacy and skills constructs, probably because of their experience with the virtual work environment in call centers prior to working remotely from home. One option for educational interpreters lacking self-efficacy and with a desire to practice their remote interpreting skills might be to seek mentorship. Past evidence has explored best teaching practices of video remote interpreting in interpreter education programs (18), however, further integration of remote interpreting practices and training into interpreter education, in general, could also help to promote interpreters' self-efficacy and skills when working remotely.

Lastly, the third principal component resembled a social dimension. Although the community/freelance interpreters demonstrated the most favorable experiences of the implementation process for the third principal component, the social dimension was fairly equally distributed, as there were no significant differences found across interpreting settings. To foster perceptions, interpreting administrators and interpreting leaders across settings could intentionally create virtual environments for remote interpreting colleagues to engage with one another and offer peer support. Even though the social support construct was fairly high performing as it ranked fourth out of the eight constructs studied, virtual team-building activities will still likely be needed to assist interpreters in maintaining the social dimension. Another suggestion to help mitigate elements of professional isolation for interpreters who are predominantly remote workers is to intersperse face-to-face or onsite interpreting (29).

The evidence from this study for tool validation of the DIBQ-SLI appears promising, but more studies are needed. Three of the four organizational resources and support items, five of the 11 innovation strategies items, two of the four self-efficacy items, and two of the three skills items had the highest component loadings in the first and second principal components, indicating that the other items from these constructs might not be needed in future iterations of the DIBQ-SLI. Social support included three items, and all three had the highest component loadings in the third principal component. Initial data reduction and affirmation did not include the three items from perceived behavioral control, five items from innovation characteristics, and one item from knowledge, indicating that these constructs may not be needed in future iterations of the DIBQ-SLI. Interpreters already seemed to have developed a sense of control from onsite interpreting (26) that they were able to transfer over to the remote context. Interpreters' abilities to tailor sign language interpreting to varying needs, thoughts on whether remote interpreting was simple to deliver or compatible with daily practice, and knowledge of working remotely all seemed to be less critical for ensuring positive perceptions of the contextual determinants of implementation. Fear and concerns regarding the risk of contracting COVID-19 did not offer interpreters the choice of whether or not remote interpreting was possible or if they had the knowledge to be successful; rather, interpreters likely sought out remote interpreting because of the heightened demand and need for security amid all the uncertainty from the pandemic.

A broad perspective of individual-level determinants of remote interpreting implementation by way of the DIBQ-SLI was presented in this study; however, there were a few limitations. First, readers are cautioned not to construe individual-level ratings as being reflective of organizational-level perceptions and vice versa (30). Second, the positive perceptions of determinants of implementation and component scores gathered in this study do not necessarily reflect that the actual implementation of remote interpreting was better, as actual implementation measures were not gathered. Third, the authors recognize that strategies employed in one context or interpreting setting to address problematic determinants of implementation may not be transferrable to another context or interpreting setting. Fourth, the DIBQ-t (12) had 26 items and ten domains compared to our 30-item, 7-domain version with eight constructs. The items on each measure differed slightly, so comparisons between this study and previous ones (12) should be made cautiously. Fifth, this study combined K-12 and postsecondary interpreters to represent the educational interpreting setting. Postsecondary or university and college settings likely have access to more resources and technical support to assist interpreters in their transition from onsite to remote work. With the least favorable ratings across all determinants of behavior constructs, future research garnering greater insights on the differences in determinants of implementation between K-12 and postsecondary interpreting is needed to inform ongoing and future remote educational interpreting needs. Finally, the sample sizes represented across the primary interpreting settings were small. While readers are cautioned about deriving generalizations using these limited data as the study may be underpowered, the internal consistency within each domain and overall were good to excellent and comparable to the DIBQ (Cronbach's alpha 0.68 to 0.93) (8), and the DIBQ-t (Cronbach's alpha 0.717 to 1 and overall 0.896) (12). Our findings of significance across specific interpreting settings can now be used to power and inform future studies.

Roughly 60% of sign language interpreters have never worked remotely prior to the COVID-19 pandemic and have never received any training on remote interpreting. This study adapted the DIBQ to assess perceptions and experiences of the determinants of implementation across interpreting settings throughout the changing work conditions of the pandemic. Interpreters scored most to least favorable across the constructs of knowledge, skill, self-efficacy, social support, perceived behavioral control, organizational resources and support, innovation characteristics, and innovation strategies. The video relay interpreters consistently demonstrated more favorable perceptions of the determinants of implementation than the educational interpreters. Organizational resources and support, innovation strategies, self-efficacy, skills, and social support had the highest component loadings across the organizational, individual, and social dimensions, and thus were deemed the most prominent constructs relevant to the context of the study. The staff and video relay interpreters had greater organizational dimensions than the educational and community/freelance settings, and the video relay setting had a greater individual dimensions than the educational interpreters. This study reinforced that the DIBQ can be modified across different contexts to evaluate perceptions of the implementation process. Since the remote interpreting demand is expected to continue, these results can facilitate the transition of future interpreters from onsite to remote work.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Research Subjects Review Board at the University of Rochester. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

GR, VS, SB, and RY-N: conceptualization and methodology. GR: data curation, project administration, and writing—original draft. GR and VS: formal analysis. GR, VS, MM, DO, SB, and RY-N: visualization. GR, VS, DO, MM, SB, and RY-N: writing—review and editing. All authors contributed to the article and approved the submitted version.

The project described in this publication was supported by the University of Rochester CTSA award number TL1 TR002000 from the National Center for Advancing Translational Sciences of the National Institutes of Health (GR) and the University of Rochester Center for Community Health and Prevention Mini-Grant (GR). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.