Empirical Study of Nova Scotia Nurses’ Adoption of Healthcare Information Systems: Implications for Management and Policy-Making

Background: This paper used the Theory of Planned Behavior (TPB), which was extended, to investigate nurses’ adoption of healthcare information systems (HIS) in Nova Scotia, Canada. Methods: Data was collected from 197 nurses in a survey and data analysis was carried out using the partial least squares (PLS) technique. Results: In contrast to findings in prior studies that used TPB to investigate clinicians’ adoption of technologies in Canada and elsewhere, this study found no statistical significance for the relationships between attitude and subjective norm in relation to nurses’ intention to use HIS. Rather, facilitating organizational conditions was the only TPB variable that explained sampled nurses’ intention to use HIS at work. In particular, effects of computer habit and computer anxiety among older nurses were signified. Conclusion: To encourage nurses’ adoption of HIS, healthcare administrators need to pay attention to facilitating organization conditions at work. Enhancing computer knowledge or competence is important for acceptance. Information presented in the study can be used by administrators of healthcare facilities in the research location and comparable parts of the world to further improve HIS adoption among nurses. The management of nursing professionals, especially in certain contexts (eg, prevalence of older nursing professionals), can make use of this study’s insights.


Implications for policy makers •
Healthcare information systems (HIS) are expensive. Evidence shows that administrators around the world spend millions of dollars to acquire such tools. • The benefits of HIS will probably not be fully realized if healthcare professionals, such as nurses who are expected to use such tools to serve patients, underutilize them. This is a major challenge to healthcare policy-makers and administrators. • Encouraging clinicians, including nurses, to adopt HIS and similar technologies is deserving of attention. • Empirical studies of nurses' adoption of HIS has the potential to influence policy-making and the management of nurses who use such technologies at work. • Facilitating organizational conditions where nurses work positively affect their intention to use implemented HIS, so do their computer habits and anxiety (lack thereof). • Computer knowledge or skills is an important antecedent to acceptance of HIS.

Introduction
Healthcare information systems (HIS) encompass computer hardware and software used for capturing, processing, storing, retrieving, sharing, and presenting data and information for decision-making in healthcare facilities such as clinics, health centers, and hospitals. 1 Examples of HIS include electronic medical records (EMR), electronic health records (EHR), mobile nursing information systems (MNIS), clinical decision support systems (CDSS), electronic patient records (EPR), and patient care systems (PCS). [2][3][4][5][6][7] Healthcare facilities across the world expend large amounts of money to acquire and maintain various types of HIS. 4,6,[8][9][10][11][12] HIS increase the timeliness and accuracy of patient care and enhance organizational efficiency by reducing costs and improving standards in healthcare delivery. 5, 6 Tung et al 6 while citing previous work commented that "if hospitals do not adapt [HIS], they will be inefficient and lose the trust of their patients. " Favorable use of HIS ultimately impacts patient satisfaction and outcomes. 13,14 In spite of the benefits of HIS to healthcare, the literature indicates that in some instances nurses, have been reluctant to accept HIS and have at other times underutilized such technologies. [15][16][17][18][19] Studies have been conducted in Canada [20][21][22][23][24] and across the world 2,3, [15][16][17][18][19][25][26][27] to assess nurses' adoption of HIS and related technologies. Studies in this study's location that examined nurses' perceptions are rare but more are beginning to emerge. 7,28,29 For example, Ifinedo et al 28 showed that perceived usefulness of HIS and favorable computer habits have positive effects on Nova Scotia nurses' acceptance of HIS. It is argued that knowledge accumulation and theory development with respect to nurses' adoption of HIS stand to benefit from efforts that present insights and observations from diverse contexts. For instance, Poon et al 8 noted the adoption of HIS varies markedly between clinicians in America and Holland. The foregoing insight suggests that it may be erroneous to accept that perceptions uncovered in prior studies would apply to all contexts, including Nova Scotia, a province of Canada, and this study's research setting.
In fact, there are particular socio-economic conditions existing in Nova Scotia that need to be highlighted; chief among these are shrinking economy, low levels of technology acceptance, declining populations, and the presence of the oldest nursing professionals in Canada who have a less than favorable attitude toward new healthcare technologies. [28][29][30][31][32] In fact, prior healthcare studies have implied that socio-economic factors may permit a better understanding of nurses' acceptance or rejection of new nursing practices and technologies. 26 Thus, investigating the realities of this study's research location is a worthwhile exercise. Nova Scotia implemented a HIS, ie, NSHIS, at an approximate cost of over $55.7 million; the system is comprised of EMR, EHR, CDSS, EPR, PCS, and related technologies. 9 No prior empirical information exists regarding adoption of the system. Given that Nova Scotia nurses and other healthcare professionals are mandated to use the system at work, it would be reasonable to gain an understanding of salient factors likely to influence adoption of the acquired system. This present study is designed to enlighten in this aspect. Information from this study would help hospital administrators and policymakers in Nova Scotia and comparable regions to develop strategies and policies that enhance clinicians' adoption of HIS and similar technologies. Results reported in studies conducted in more socioeconomically endowed provinces of Canada, ie, Quebec and Ontario, that used one of the commonly used psychosocial behavioral theoretical frameworks for examining healthcare professionals' adoption of technologies, ie, the Theory of Planned Behavior (TPB), 33 34 that included TPB's three constructs. They found that facilitating conditions was the only significant predictor of acceptance of a new telemonitoring system. Research efforts from across the world that included the variables of attitude, subjective norm, and facilitating conditions have also provided mixed results in relation to the pertinence of these constructs in explicating nurses' adoption of HIS and other technologies. 35,36 For instance, subjective norm was found to be the most influential factor in terms of safety behaviors displayed by nurses in a study conducted in Iran. 26 Studies in Kenya 37 and Turkey 38 showed that nurses had favorable attitudes toward computerization.
Simpson and Kenrick 16 also reported that nurses' attitudes toward computerization in clinical practice in a British general hospital were generally positive; however, they found significant differences in relation to nurses' age and length of service. Chung and colleagues' study of nurses' intention to adopt HIS in Taiwan showed that attitude and subjective norm were key determinants. 39 In Spain, Asua et al 40 6,[42][43][44] because this particular social cognitive theory has proven to be helpful in explaining nurses' use of HIS and related technologies. 45 Importantly, research has shown that TPB explained significant variance in healthcare professionals' intention to use technologies. 35,44 In their systematic review, Godin et al 35 revealed that TPB explained 59% of the variance of healthcare professionals' intention to adopt various medical practices and systems to underscore its suitability for this study. Several healthcare researchers, including some cited in this paper, have used the theory in comparable studies. It is worth adding that the adoption of HIS, even in mandatory settings, is still susceptible to the behavioral intentions of healthcare professionals. 35,44 That noted, TPB posits that individual behavior is influenced by attitude, subjective norms, and perceived behavioral control. Attitude refers to an individual's positive or negative feelings toward engaging in a specified behavior. Subjective norms describe an individual's perception of what significant others think about a given behavior. Perceived behavioral control refers to an individual's beliefs regarding the resources needed to facilitate a behavior. In this study, this variable is described in terms of 'facilitating organizational conditions'; others have used a similar term. 21,22,24 Intentions refer to the willingness to engage in a behavior. 33 As actual usage of HIS was not directly measured in the study, nurses' self-reported usage of HIS is employed in lieu. Healthcare researchers 36 have signified the need for focused attention on issues of concern to specific populations of healthcare workers being studied. Heeding this advice, an informal study was conducted to gauge nursing professionals' views on factors likely to encourage or discourage adoption of HIS among nurses at work in the research location. The informal study solicited the participation of 6 registered nurses (RNs) in the province; one participant being the former head of the provincial RNs' association. Among other concerns, participants stressed the inability to use everchanging healthcare technologies effectively due to the age of nurses, less than desirable computer habits, poor computing competencies, and phobia of technologies. It is difficult to posit with certainty whether the issues noted in the informal study are specific only to this study's location given that such issues have been highlighted in the literature. 46 Griebel et al noted that computer competencies and anxiety are among the 10 additional factors recommended for consideration alongside established constructs in technology adoption frameworks for healthcare research. 46 Accordingly, TPB was extended in this study by three factors, ie, computer anxiety, computer habit, and computer knowledge, that were found to be important for nurses' adoption of new technologies in the research setting. Computer anxiety refers to a state of heightened tension or a feeling of apprehensive expectation toward computers. 29,47 Habit, as proposed in TIB, refers to behavior that has become automatic often requiring only minimal mental effort to accomplish. Thus, computer habit refers to a computer behavior that has become automatic. 40,48 Computer knowledge refers to the skill level an individual has with basic computer hardware and software such as internet browsers, word processors, spreadsheets and presentation software.

Research Model and Hypotheses
To understand the phenomenon of nurses' adoption of HIS in Nova Scotia, it is imperative to base the study on a theory (this has been accomplished). It is also a common practice in quantitative research to draw insights from prior empirical studies and observations in formulating predictions or hypotheses. This study followed such guidelines. The formulated hypotheses and relevant paths or relationships are depicted in Figure 1. It has been demonstrated that nurses having favorable attitudes toward HIS tend to have willingness to use such systems at work. 3,21,36,41,49 Nurses tend to use adopted computing systems and technologies at work when their peers use such tools. 20  Nurses facilitating organizational conditions matter for the acceptance of HIS and related technologies. [20][21][22]27 Namely, higher levels of organizational support and assistance boded well for nurses' acceptance of technologies. 21,41 Past studies showed a negative relationship exists between clinicians' computer anxiety and their attitudes toward technology use. 2,50 Evidence exists to show that nurses' computer knowledge and competence enhance positive attitudes toward technology use in work environments. 2,3,7,18 By the same token, nurses with adequate computer knowledge tend to be less anxious about using technologies for work. 2,50,51 Thus, the following set of hypotheses are proposed: H1: Nurses' attitude toward HIS will have a positive effect on behavioral intention to use HIS H2: Nurses' subjective norms will have a positive effect on behavioral intention to use HIS H3: Nurses' facilitating organizational conditions will have a positive effect on behavioral intention to use HIS H4: Nurses' computer anxiety will have a negative effect on attitude toward HIS H5: Nurses' computer knowledge will have a positive effect on attitude toward HIS H6: Nurses' computer knowledge will have an inverse relationship with computer anxiety Studies have shown that peer influence plays a significant role in the behavioral intention of healthcare workers, including nurses, to accept and integrate computing technologies in their work or practice. 21 Sample and Instrument More than 70% of participants had university education and 96% were females, which is an indication of the characteristics of RNs in Canada and elsewhere. 18,55 The sample included different age categories; however, approximately 75% of the respondents were aged 41 years and above. This information is consistent with data from a report, which indicated that Nova Scotia RNs are the oldest in Canada. 31 The majority of participants work in public healthcare facilities where HIS are deployed. Most Canadian RNs are government employees. 55 On average, participants have worked 14.5 years with their current employers (standard deviation [SD] = 10.8). Table 1 shows the rest of the participants' demographic information.
Consistent with the survey method, this study used measuring scales validated in prior studies. Items used to operationalize organizational facilitating conditions (FAC), behavioral intention to use HIS (BEH), subjective norm (SUB), and self-reported HIS use (IUS) were adapted from Venkatesh at al. 42 Measures for computer anxiety (ANX) were adapted from Compeau et al. 56 Items used to represent computer habit (HAB) were adapted from Limayem and Cheung. 48 Measures used to capture attitude toward HIS (ATT) were adapted from Davis. 43 Measurement items were anchored on a 7-point Likert scale ranging from "strongly disagree" (1) to "strongly agree" (7) in which participants were asked to indicate an appropriate response. For the construct of computer knowledge, participants were asked to indicate their level of skill in terms of software applications such as internet browsers, word processors, spreadsheets and presentation software. The scale used to assess their skill levels was anchored a scale that ranged from "No skill at all" (0) to "Very competent in using it" (6). A full list of measuring items for the constructs and their descriptive statistics is provided in Table 2.

Data Analysis
This study used the partial least squares (PLS) technique for data analysis, which other healthcare researchers have used in similar studies. 4,7,27 PLS is suitable for this study because it places minimal demands on sample size and residual distributions; it allows the use of observed items to represent a construct. 57,58 This study used WarpPLS 5.0, which was chosen for its ability to handle both formative and reflective constructs in a model. 59 PLS tests the reliability and validity of measures. Reliability of reflective constructs with values of 0.6 or more are usually considered acceptable. 58 Reliability indicators -composite reliability and Cronbach alpha -shown in Table 3 are consistently above 0.6. Regarding validity of the constructs, each reflective measurement item is expected to load highly on its latent construct 58 and standardized item loadings exceeding 0.707 are considered adequate. 58,59 Item loadings shown in Table 2 are all above this threshold. Additionally, a study's constructs are expected to be distinct. [57][58][59][60] To show distinctness (ie, discriminant validity), a minimum value of 0.5 average variance extracted (AVE) is recommended and the square root of AVE should be larger than correlations between that construct and all other constructs in the model. 60  Table 3 shows that this study's constructs meet these criteria. For the formative construct (in bold line in Figure 2), it is recommended that item weights and the presence of multicollinearity be checked. 61 In assessing multicollinearity among items used to represent computer knowledge, the variance inflation factors (VIF) of items were found to range from 2.112 to 3.217, which are below the threshold cutoff value of 3.33. 61 In addition, items are expected to have statistical significance. 61 The items used to represent computer knowledge are significant at the level of P < .001 to show that the construct items have adequate psychometric properties. With reliability, convergent validity, and discriminant validity of the constructs established, the structural model for the hypothesized paths was then tested. The structural model indicates the significance of hypothesized relationships using the path significance (p), beta (β) coefficients, and the coefficient of determination (R 2 ), which is the amount of variance explained by the indicators. Table 4 and Figure 2 show the results of PLS analysis and supported hypotheses. Computer knowledge alone explained 46% of the variance in attitude toward HIS. Computer knowledge, subjective norm, computer anxiety, and organizational facilitating conditions explained 42% of the variance in attitude toward HIS. Computer anxiety, attitude toward HIS, computer habit, subjective norm, and organizational facilitating conditions explained 51% of the variance in behavioral intention to use HIS. Behavioral intention to use HIS alone accounted for 32% of the variation in nurses' self-reported HIS use. Falk and Miller 62 recommended a minimum value of 0.10 (10%) for R 2 ; results in this study are above this threshold. Overall, the results indicate the proposed research model is relevant for theory development. 57,58,62 Discussion This paper describes a study that used TPB, complemented by computer anxiety and computer habit, to investigate nurses' adoption of HIS in Nova Scotia. Analysis of the collected data showed that 9 out of 11 formulated hypotheses were supported; two were unsupported. The unsupported predictions will be discussed first. Sampled nurses' attitudes toward HIS did not positively influence their behavioral intention to use HIS at work. Plausible explanations for this result might be related to  extraneous factors in the research setting. The study's sample was comprised of nurses who were mainly aged 40 years and above, it is likely their attitudes toward implemented technologies may not be as favorable compared to the views of younger colleagues who may be more receptive to technological innovations. 16,38 In fact, past studies 4,18 found statistically significant relationships between nurses' age, attitude toward computers, and intention to use HIS. Participants in the study, especially older nurses, did not seem to have positive attitude toward computerization and innovative medical technologies such as HIS; this might have negatively impacted the result in this aspect. Some feedback representative of older respondents' comments is provided: " The variable's lack of significance in the study also may be related to the fact that most nurses in the sample are about the same age, ie, older, and to a large extent, share the same beliefs about implemented HIS and related technologies in their work environments (see the comments above). These participants may not have the belief that they can influence peers to accept implemented systems at work. The literature suggests that older nursing staff can act as role models for others with regard to technology adoption in hospital settings. 63 Unfortunately, this may not be possible in the research setting because the sample has a disproportionate number of older nursing professionals. In brief, the result demonstrated that peer influence or social pressure played little or no role in encouraging nurses to accept HIS in Nova Scotia. The study showed that organizational facilitating conditions encompassing infrastructural, technical, and management support are critically important in stimulating nurses' willingness to use or accept implemented HIS in work environments. This result mirrors the result reported by others in Canada and elsewhere to affirm that organizational facilitating conditions is the most important factor among TPB's constructs that influence nurses' intention to use HIS. 21,24,27,35 This result indicated that nurses in Nova Scotia who believed their healthcare organizations had high levels of facilitating conditions had corresponding high intention and favorable usage outcomes. The result supported past findings indicating that nurses with less computer anxiety tend to develop a more favorable attitude toward computerization and implemented HIS and have more willingness to accept and use such technologies in work environments. 27,35,50,52 Nurses' computer knowledge was found to have a significant positive effect on attitude toward HIS in the research location. This result is consistent with the viewpoint noting that computer knowledge continues to be viewed as an important factor that can enhance nurses' acceptance of healthcare technologies. 7   have sufficient skills to search best evidence and to use the Internet and online databases for information seeking and retrieval. " Computer anxiety was also shown to be negatively linked to computer knowledge. Notably, nurses who had computer knowledge or skills were less likely to be anxious about using HIS at work. The result supports observations reported in similar studies. [50][51][52] The relationship between subjective norm and attitude was confirmed by the data. It is worth noting that "attitude" and "intention to use" are dissimilar constructs. Support and influence from peers or colleagues do impact nurses' attitude toward implemented HIS and related systems. Others have offered similar insight. 20,23,41 Nurses in the study showed that where their organizations are able to minimize impediments to HIS utilization or provided necessary assistance to using such systems, attitudes toward the system tend to improve. Previous nursing studies have affirmed the foregoing fact. 2,3,36 The result showed that nurses' computer anxiety had a significant negative relationship with intention to use HIS. This result is inconsistent with studies of healthcare professionals, including nurses, that indicated that computer anxiety had no effect on whether a nurse would or would not accept a particular technology. 65 Nurses' computer habits are positively associated with HIS usage behaviors. That is, nurses whose computer behaviors have become automatic, over time, tend to be more willing to use implemented HIS at work. This result lends credence to a prior study conducted in Canada 53 ; however, it is inconsistent with observations in a Spanish study. 40 The data confirmed that nurses' behavioral intention to use HIS strongly influenced self-reported use of the technology. In other words, where the willingness to use implemented HIS was high, favorable levels of acceptance of such systems ensued. Previous studies 7,65 provided support for the existence of a strong, positive relationship between nurses' behavioral intention to use information technologies and use outcomes.

Contribution and Implications for Research
This present study contributes to research literature by demonstrating that TPB is a good theoretical model for predicting and explaining nurses' adoption of HIS. The extended TPB has more predictive power compared to the original TPB, which explained 39% and 27% of the variances in intention and behavior. 54 The extended TPB used in this study explained 51% and 32% of variations in intention and usage behavior of HIS, respectively. This study adds to viewpoints indicating that extending or modifying established theoretical frameworks for studying healthcare professionals' adoption of technologies is a useful exercise. 6,36,46 While this study's results, in particular those relating to attitude and subjective norm in TPB agree with previous research findings on clinicians, ' including nurses, ' adoption of HIS and similar technologies in Canada and elsewhere, 20,23,36 a somewhat interesting result was also noticed. Namely, only one TPB construct was found to be positively related to intention; another study in Canada presented a similar result. 24 This study argues that extraneous factors (eg, older population of nurses) in the research setting might have influenced the relationships between attitude and subjective norm, and intention in the model. Admittedly, further research is needed to quash or corroborate the foregoing proposition. What is more important is the fact that healthcare researchers need not downplay influences arising from contextual influences in their research locations. This study's focus on Nova Scotia with its unique characteristics produced results that are not in total agreement with the tenets of TPB. In light of the growing body of work that used TPB to investigate nurses' adoption of healthcare technologies and the sorts of results reported in such studies, it may be worthwhile for future inquires to examine the extent to which socio-economic, and even cultural factors impact findings across contexts.
To provide an answer to the question: what factors influence Nova Scotia nurses' adoption of HIS, this study revealed that organizational facilitating conditions is the most influential variable that predict Nova Scotia nurses' intention to use HIS. Unfortunately, attitude and subjective norm did not offer meaningful results. With respect to the question dealing with the nature of the relationships among the study's constructs, it was shown that computer anxiety (lack thereof), computer knowledge, subjective norm, and organizational facilitating conditions are strongly associated with nurses' attitudes toward HIS. Computer habit and lack of anxiety about computers enhance intention to use HIS. Overall, the study's results provide further empirical support for challenges espoused by nurses in the research location. Recall, less than desirable computer habits, poor computing competencies, and phobia of technologies are among the inhibitors that were pin-pointed. Future research in the region and comparable locations with similar demographics can build upon the information provided in this study.

Implications for Practice and Policy-Making
This study has implications for healthcare practitioners. Given that nurses are considered important actors in providing healthcare services to the population, managers of healthcare professionals in Nova Scotia could strive to provide necessary infrastructural, technical, and managerial support to encourage nurses' adoption of HIS and similar technologies needed to improve healthcare. Comments received from this study's participants indicated the availability of necessary support (eg, helpful IT desk, training facilities' , system usage manuals) and encouragement from management as worthy of note. Externally, nurse educators in Nova Scotia and comparable regions of Canada and the world can also play a role in helping curb computer phobia both for older and younger nurses. For example, older nurses can be given short-term courses in nursing informatics and new nurses can benefit from a welldesigned curriculum that emphasizes appropriate computing skills and competence. 5,19 Healthcare organizations ultimately benefit from such programs. A study noted that nursing professionals who received medical informatics classes during their formal education are more likely to have positive attitudes toward technologies used for healthcare delivery. 38 Older nurses who receive regular computing trainings are also more accepting of computer use at work. 67,68 Internally, the provision of in-house educational sessions on the importance of implemented HIS for nursing tasks, initial practical training, and support related to newly acquired HIS can help diminish fears associated with using such technologies. Such training, especially for older nurses, could emphasize that mistakes are a normal part of learning and using a new system. 68 Assurance of this sort can help diminish any phobias, which in turn improve attitudes toward acquired HIS. Management could provide continuous technology training to all nurses. For example, Kuo et al 66 commented that "continuous educational programs can be provided for nurses to enhance their information technology literacy, minimizing their…discomfort about information technology. " Because the opinions of peers may influence attitude, it may be worthwhile to encourage nurses' involvement with implementation teams when new HIS are being procured. 49 Other nurses would be able to learn from their colleagues and gain first-hand knowledge of how the new system functions; such measures may help improve attitude toward the system. Administrators could also explore other motivators that will enhance nurses' general attitudes toward accepting new technologies. For example, incentives (ie, monetary and otherwise) can be given to nurses who utilize implemented HIS effectively in their tasks. Computer habit matters for nurses' HIS adoption. It may also be rewarding for healthcare establishments to consider developing game applications (apps) that mimic the features and functions of to-be implemented HIS. Such apps (web-based or handheld) can then be made available to clinicians, including nurses, prior to system acquisition or implementation. The supposition is that habits acquired from such endeavors may become automatized to favorably benefit perceptions and attitudes of users toward the actual system when it is eventually deployed.

Limitations
The results reported in this study should be interpreted against its limitations. First, there is a noticeable gender bias in this study-96% of the participants are females. Nonetheless, the sample was not significantly different from the population of nurses in Nova Scotia and Canada. Second, the data sample was collected from just one region of Canada: Nova Scotia. It is difficult to claim that the results can be generalized to all socio-economically disadvantaged parts of Canada. Third, the study's participants provided answers to self-reported usage of HIS; the possibility of social desirability bias exists with such an approach. Logs of actual usage of HIS may offer better insight. Fourth, this study did not focus on any specific HIS. Factors affecting the use of disparate HIS, such as MNIS, CDSS, and EPR, are likely to be different; this reality might be limiting to this study. Fifth, data for the study was obtained in a crosssectional survey, which only presents a snapshot of insight; a longitudinal study would offer deeper information. Sixth, as it was not explicitly noted that the same measuring items for TPB constructs were used for this study and others, direct comparisons of results should be done with caution.

Conclusion
In the 21st century, HIS have become a critical resource to improve healthcare delivery across the world, including Nova Scotia. Moreover, acquiring such healthcare infrastructure is very expensive. As such, it is important for researchers to assess the adoption of such technologies by clinicians. This study was conducted in Nova Scotia where HIS has been implemented, but very scant empirical information exists on clinicians, ' including nurses, ' adoption of the technology. This study contributes to the literature by presenting perspective from this setting. Knowledge accumulation and theory development in the area benefit from the endeavor. Using an established theoretical underpinning, ie, TPB which was extended by relevant factors, the study revealed that computer habit, computer knowledge, lack of computer anxiety, and organizational facilitating conditions are among key factors predicting and explaining nurses' adoption of HIS in Nova Scotia. Information presented in the study can be used by administrators of healthcare facilities in the research location and comparable parts in Canada and elsewhere to design strategies and policies that further encourage HIS adoption among their nurses and other healthcare professionals. Future studies can build upon this work by adding other relevant constructs such as compatibility and self-efficacy.