Healthcare professionals’ acceptance of BelRAI, a web-based system enabling person-centred recording and data sharing across care settings with interRAI instruments: a UTAUT analysis

The 21st century is the century of the ‘ageing population’. In Belgium (population 2011: 11 million), there are now 1.88 million people aged 65 and older which equals 17% of the population [1]. In 2020, they will be 2.22 million, which is plus 20%. Over the next 10 years the number of people over 80 will grow to nearly 578,500, compared with 479,000 in 2011 [2].

As people live longer, healthcare environments are increasingly confronted with older persons, characterised by chronic conditions and/or comorbidities necessitating long-term care [3‐6] and rising costs. There is also a substantial shift away from institutional caregiving. This evolution increases the complexity of caregiving since multiple service providers are delivering care to individuals [3, 7, 8]. The need to receive support from multiple service providers has important implications for persons with complex care needs [9].

As these people transfer between settings, accurate communication and sharing of crucial information is fundamental in providing high quality care [10]. A lack of information (transfer) can result in errors [11], increased assessment burden, and frustration among care recipients and their supporting network [12]. Deficient information transfer or communication can result in uncoordinated care, adverse events and iatrogenic complications influencing morbidity, mortality and hospital outcomes [10, 11]. It appears that clinical information systems that typically have been designed to support single service providers in one setting no longer meet the necessary requirements [8].

As people migrate through this maze of healthcare providers, the need for integrated and coordinated assessment systems providing comparable information increases [13], as does the need for service provisions that overcome barriers [3, 8, 14‐16]. The importance of good communication in healthcare environments [17] to enable person-centred care and increase safety cannot be emphasised enough. Collected data must be available and useful to those who have to decide at the personal, clinical, managerial, and public policy levels [3, 18]. In complex healthcare and social care systems in which people with the same needs might receive different services in different sectors and in which changes in one subsystem might affect other subsystems, thinking at the system level is crucial. To this end, it is necessary to have an integrated, standardised and computerised collection of information in a language that is clear and understandable to all users [3, 8, 19, 20].

The interRAI suite of instruments [21] addresses the changing strengths, preferences and needs of vulnerable people rather than the organisation which provides services at a single point in time [6, 8]. This ‘third generation’ multi-domain suite of compatible instruments has the potential to transfer and share high quality person-centred information and to compare people, services and outcomes across settings [12, 22‐26].

Although these instruments provide opportunities to improve continuity of care, as well as efficiency and quality of care [27‐37], they cannot reach their full potential with paper-based recording only. Paper-based recording has severe limitations: it is applicable only to slow, manual processing, it is difficult to share, it may lack information concerning the history of treatment, it is centred upon institutional boundaries, it is vulnerable to errors, it lacks interoperability with other information systems and its information may not be presented in ways relevant to specific care contexts [38]. Moreover, the complex algorithms of the interRAI suite of instruments that generate decision support outcomes and guidelines are highly dependent on automation.

These problems can only be addressed by the use of computer-based information technologies [8, 19]. Information technology (IT) is already being used to prevent human error in healthcare and to improve patient care outcomes worldwide. In 2000, the Institute of Medicine in Washington DC recommended increasing efforts to incorporate information technology into the delivery of patient care [39].

In Belgium, web-based software has been developed to support the use of the interRAI instruments in different settings and in different languages^a. This software has been tested and evaluated. As, to our knowledge, there is no evidence of a fully computerised set of interRAI suite instruments throughout the whole world, this research can be considered innovative.

However, developing and implementing information technology to support a shared dataset is a difficult, slow and gradual process [40]. There is also a gap between the two different professional approaches. On the one hand, information and communication technologies (ICT) scientists may have the technical skills to develop an application, but they are unlikely to fully understand how the application must operate in order to support the organisation. On the other hand, healthcare organisations may be able to identify and express the ‘what’, ‘when’, ‘for whom’ and ‘where’ issues concerning an application to support their goals, but they may not have the technical skills to make it happen [41]. Consequently, in order to achieve a successful implementation, substantial challenges have to be tackled [42]. There is a need for good-quality software, ongoing IT support, access to computers, and a degree of computer literacy in care organisations [8, 19, 43]. Moreover, there are numerous possible reasons for why change may be opposed: misunderstandings about the need for change, fear of failure, fear of loss of professional autonomy, fear of additional workload, fear of transparency, reluctance to share instrumentation [8], concerns about confidentiality, concerns about the impact on human resources and the professional demands that come with such a change. In addition, the ICT professionals need to have a grasp of what is actually happening and needed in care practice.

With all this in mind, caregivers will not necessarily accept a (new) health information system and use it in daily practice [16, 44]. Information technology that is not used, or is only partly used, cannot reasonably be expected to contribute to improving the quality of care. On the contrary, it can be an extra burden or cause other problems [45]. It is clear that the impact of such a change could be a contributing factor leading to an organisational crisis, causing conflict among professional groups. It is therefore crucial to understand how aspects of real-world activities relate to one another [41] and to clarify what the influencing factors are of the acceptance of and the use or failure of the information system.

In this paper we explore the applicability of the unified theory of acceptance and use of technology (UTAUT) theoretical model in the BelRAI healthcare project in analysing the acceptance of the BelRAI web application by healthcare professionals in three Belgian care settings (home care, nursing home care and acute hospital care for older people with disabilities). This will help to broaden our understanding of the development and implementation of healthcare information systems.

This paper describes a study concerning the utilisation of the UTAUT model to provide an explanation on professional caregivers’ acceptance of the BelRAI web application. Given the fact that UTAUT was originally developed outside the healthcare context, we built upon this framework to complement prior related research on technology acceptance and use [45, 48, 58, 59] by exploring the applicability of the UTAUT model in the specific BelRAI healthcare project - home care, residential care and acute hospital care for older and disabled people.

This study shows that performance expectancy, effort expectancy and social influence do not have a significant influence on behavioural intention as shown in the original UTAUT study [46] and other related research on technology acceptance and use [45, 48, 58, 59]. However, behavioural intention is influenced by facilitating conditions associated with the presence of the resources and the knowledge required to use the system, the degree of compatibility of the new system with other systems in use and the availability of assistance in the case of system difficulties. Similarly, self-efficacy associated with the possibility to complete a task using the new information system with only the built-in help facility, having the possibility to call someone for help and having enough time to practice, has a significant influence on behavioural intention to use the BelRAI web application.

These results seem to be in line with the findings by Declercq et al. [68]. Inadequate computer equipment (hardware, software, internet connection) causes frustration among users, is counterproductive, and constitutes significant barriers to accepting the new technology. At the same time it is important to achieve the right balance between confidentiality and accessibility. The BelRAI web application meets the requirements set by the Belgian Commission for the Protection of Privacy, and the privacy and security standards have been positively evaluated by the users. However, rigid security regulations also result in access procedures that are complex and time-consuming, as a result of which they negatively influence the application’s user-friendliness. Occasionally, when it is possible to get around the security measures (by exchanging the eIDs, for instance), a result will be reached that is opposite to the one intended.

In any case, there is need for high quality education, both internally and externally. The expectations for external education and support can strongly depend on the context or setting. In one setting there might be a need for practical education, whereas in another a rather theoretical training might be appropriate. In an initial phase the training can be organised per setting, so that each setting gets specific information based on its needs. During a subsequent phase, the training groups can be composed rather heterogeneously in order to put an emphasis on the interaction and collaboration between different settings. This is particularly appreciated by home care caregivers since they hardly get in touch with other caregivers active in both home care and other care settings. Also, the periods in between different training phases should be long enough in order to enable caregivers to acquire experience and give feedback. The group sizes and the number of teachers must be in tune with one another so that teachers can adopt a problem-solving approach. The internal training sessions, organised by the organisations themselves, should be very practical. For an organisation, it is appropriate to appoint a reference person who can hold the position of internal trainer and act as a contact person. Assigning some staff members to work on the project can facilitate a full-time follow-up and implementation. The learning period does not end after the final training session. In order to be able to fluently use new technological tools, each caregiver should pass through a post-training learning process. Moreover, each caregiver should be given the opportunity to gain sufficient experience in a variety of situations. In addition to the technological aspect, the content (assessment) as such constitutes an indispensable part of the learning process. Frequent and intensive exercises are required in order to have good command of all applications. However, frequent staff changes can seriously disturb the learning process.

Time is an important resource. In practice, a high workload, a lack of personnel, high client turnover, and the presence of acute problems (acute hospitals) turn out to be significant barriers standing in the way of implementing new information technology. In order to improve efficiency and avoid an increased workload, the degree of integration of existing information technology (e.g., electronic patient records) with new technologies is decisive in the success of implementation. It is in everyone’s interest that everything can happen in a user-friendly and safe way, without technical failures and with a minimum of intermediate steps. Integrated technology leads to fewer difficulties and will be accepted and used more rapidly.

During the implementation of new technologies, it is also important for active caregivers not to feel alone, but rather to feel supported in times of doubt by the care team, the (middle) management and other caregivers involved. Other factors that can contribute to a successful implementation are a help desk, a wiki website, and the possibility of being able to work in a quiet environment. Moreover, the collaboration process can be influenced by the physical proximity of various disciplines: people working within the same organisation collaborate more easily than people employed at different sites.

Although the use of the BelRAI application was mandatory, it is plausible that the behavioural intention of professionals is influenced by the short-term nature of this project. Simulating mandatory use of new information technology in a public healthcare setting is not straightforward. Therefore, since the future professional value of BelRAI is not guaranteed, performance expectancy may not have influenced behavioural intention. Similarly, effort expectancy may not have a significant relationship with behavioural intention because people tend not to exert great effort during a pilot project. Social influence may not have an impact on behavioural intention because of a difference in attitude between caregivers and typical technology users. Future study designs should disentangle these effects inside the UTAUT framework to understand user acceptance in the context of a permanent implementation. The measurement of behavioural intention to use new information technology could be expanded to take into account the emotional disposition of professional caregivers. Ultimately, when new information technology is implemented on a large scale, professionals should be looking forward to using it.

The practical implications of this research are that good-quality software, interoperability and compatibility with other information systems, easy access to computers, a degree of computer literacy, training facilities, built-in and online help and ongoing IT support are critical influencing factors in stimulating the behavioural intention to use new technology. These findings can be used by policy makers to maximise the acceptance and the success of new technology initiatives.

For researchers, this study indicates that a bottom-up approach should be preferred when building upon the UTAUT model. The conclusions of the original UTAUT study [46] with regards to the item and scale construction should not be copied blindly across different information systems. In the specific context of BelRAI, self-efficacy is an important influencing factor of behavioural intention while this factor is not retained in the original study.

Our elaboration of the UTAUT model is limited since no information is available about the actual use of the BelRAI web application. Other research should encompass variables that relate to actual use (e.g. frequency and duration) in order to investigate the influence of behavioural intention. In this research it is not possible to gain insight into non-response bias and to know who is cooperative and who is not. It is likely that respondents who did not complete the survey have a more negative attitude towards new technology. Future research should provide insight into the characteristics of the entire study population. Future research should also take into account experience with the system in a longitudinal evaluation and examine the effects of age, gender, healthcare context, profession, computer literacy and specific training sessions not only related to behavioural intention but also to actual usage of the system. To better understand healthcare professionals’ acceptance of BelRAI, more research is needed to generalise the findings of this feasibility project to a future, broad implementation.

^aSince 2006, different government-funded feasibility studies concerning the implementation of interRAI assessment instruments at a national level were carried out by an interuniversity team of the KU Leuven (Elderly Care Research Unit at LUCAS& Center for Health Services and Nursing Research) and the Université de Liège (Département des Sciences de la Santé Publique) [68, 87‐91].

^b85 questionnaire items related to the UTAUT constructs were measured: Performance expectancy (PE), 24 items; Effort expectancy (EE), 10 items; Attitude towards using technology (ATUT), 14 items; Social influence (SI), 9 items; Facilitating conditions (FC), 11 items; Self-efficacy (SE), 10 items; Anxiety (ANX), 4 items; Behavioural intention (BI), 3 items.

^cSelection based on item loadings is consistent with recommendations in the psychometric literature [84]. “This approach favours building a homogenous instrument with high internal consistency but could sacrifice content validity by narrowing domain coverage” – items from some of the models not being represented in some of the core constructs [46].

^dThe p-value of this t-value should be significant at the 0.05 alpha protection level [80].

^eBy the latest version of PLS-Graph 03.00 build 1126 of 2003 AVEs are generated automatically using bootstrapping as a resampling procedure. The AVE measures the variance captured by a latent construct being the shared (or explained) variance. The square root of AVE of each construct should be much larger than its variance shared with any of the other constructs in the model and should be at least 0.50. Conceptually, it is equivalent to saying that the correlation between the construct and its measurement items should be larger than its correlation with the other constructs [45, 76, 80].

^fThe corresponding t-values suggest significance of the coefficients (t-values > 1.96, significance level p < 0.05) [80]. PLS generates the square of the correlation coefficient (R²) for each dependent construct in the model. R² is interpreted as the percentage of shared (or explained) variance and thus represents the proportion of variance in the dependent constructs which can be explained by the independent ones [45, 77].