Acceptance of Clinical Decision Support System to Prevent Venous Thromboembolism Among Nurses: an Extension of the UTAUT Model

doi:10.21203/rs.3.rs-1414024/v1

Background Venous thromboembolism (VTE) has been a major public health problem and caused a heavy disease burden. VTE Clinical decision support system (CDSS) was proved to have a positive influence on the prevention and management of VTE. As the direct users, nurses' acceptance of VTE CDSS is of great importance to support the successful implementation of the system. However, there are few relevant studies to investigate nurses' acceptance and the associated factors are still unclear.

Objective To investigate the determinant factors of nurses' acceptance of VTE CDSS with the modified Unified Theory of Acceptance and Use of Technology (UTAUT).

Methods We designed a questionnaire based on the modified UTAUT and then a cross-sectional survey was conducted among nurses in a tertiary hospital in Nanjing, China. Statistically, a Structural Equation Modeling (SEM)-Partial Least Squares (PLS) path modeling approach was applied to examine the research model.

Results A total of 1100 valid questionnaires were recycled. The modified model explained 74.7%, 83.0% and 86% of the variance in user satisfaction, behavioral intention and user behavior, respectively. The results showed that performance expectancy (r=0.254, p=0.000), social influence (r=0.136, p=0.047), facilitating conditions (r=0.245, p=0.000), self-efficacy (r=0.121, p=0.048) and user satisfaction (r=0.193, p=0.001) all had significant effects on nurses' intention. Although effort expectancy (r=0.010, p=0.785) did not have a direct effect on nurses' intention, it could indirectly influence nurses' intention with user satisfaction as the mediator (r=0.296, p=0.000). User behavior was significantly predicted by facilitating conditions (r=0.298, p=0.000) and user intention (r=0.654, p=0.001).

Conclusion The research enhances our understanding of the determinants of nurses' acceptance of VTE CDSS. Among these factors, performance expectancy was considered as the top priority. It highlights the importance of optimizing system performance to fit the users' needs. Generally, the findings in our research provide clinical technology designers and administrators with valuable information to better meet users' requirements and promote the implementation of VTE CDSS.

clinical decision support system

nurses

acceptance

UTAUT

What was already known on the topic?

Uses of VTE CDSS increase the proportion of adequate VTE prophylaxis and correlates with a reduction in VTE events.

UTAUT is a widely used model to access users' technology acceptance.

Adoption of nurses is of great significance to the successful implementation of VTE CDSS, however, no studies have ever used theory to investigate nurses' acceptance of VTE CDSS.

What this study added to the body of knowledge?

This study confirms the validity of the modified UTAUT to investigate the associated factors of users' acceptance of VTE CDSS in the context of developing country.

It is important to provide a user-friendly VTE CDSS to meet users' requirements and promote their satisfaction.

Venous thromboembolism (VTE), including pulmonary embolism (PE) and deep venous thrombosis (DVT), is estimated to be the third fatal cardiovascular event^{[1, 2]}. It is a common complication among inpatients, which causes perioperative mortality and unexpected death^[3]. During the past decade, the hospitalization rate of VTE has increased steadily from 3.2 to 17.5 per 100,000 population in China^[4]. Moreover, the complications of VTE, like post-thrombotic syndrome, chronic thromboembolic pulmonary hypertension and hemorrhage seriously affect patients' quality of life and cause a heavy disease burden^{[5, 6]}.

VTE is considered as a preventable event^{[7, 8]}. It is estimated that appropriate prophylaxis can reduce the relative risk of DVT and PTE by 50% and 66.7%, respectively^{[9, 10]}. However, the current prophylaxis rate is poorly low^[11]. A national, multi-center study^[12] revealed that only 14.3% of inpatients at risk of VTE received some form of thromboprophylaxis, among which just 10.3% received appropriate prophylaxis recommended by guidelines. The data emphasized the insufficient management of VTE and showed the necessity to improve the clinical practice of medical staff.

Nurses as the largest cluster of medical staff, play a critical role in identifying inpatients at risk of VTE, implementing prophylaxis measures, and making clinical decisions^[13–15]. Many hospitals have adopted VTE clinical decision support system (CDSS) to assist nurses to assess inpatients' individual risk and overcome the barriers in offering prophylaxis^{[16, 17]}. The VTE CDSS is a computerized application system based on artificial intelligence and clinical information storage technology, which aims to realize the functions of risk stratification with the embedded risk assessment models (e.g. Paduwa, Caprini, Geneva), electronic alert reminder, priority preventive measures recommended, and the record of prevention process^{[17, 18]}. Previous studies^{[17, 19]} have shown the introduction of VTE CDSS can significantly increase the rate of adequate prophylaxis and then decrease the incidence of VTE.

However, it was shown that there existed an apparent mismatch between the benefits and adoption of CDSS among nurses^{[20, 21]}. Nurses might become less likely to use the CDSS as they thought it brought them workload, work complexity and perceived threat to professional autonomy^{[20, 22]}. Several studies explored the factors related to nursing staff's use of clinical technology concluded that the value of technology was determined by the appraisal of users^{[20, 23]}. What's more, it has been reported that over 40% of information technology was failed or abandoned for the poor adoption of users^{[24, 25]}. Hence, it is important to understand nurses' attitude and use intention toward the VTE CDSS and seek the influential factors to help engineers improve the design of the system and then extend the implementation and utilization of the system^{[26, 27]}. Nowadays, in the field of technological nursing health care, little attention has been paid to the CDSS used by the nursing staff^[25]. The factors that influence the nurses' acceptance of VTE CDSS are still unknown. The purpose of this research is to explore the acceptance of VTE CDSS among nurses and investigate the associated factors.

Theoretical framework and research hypothesis

The original model

The Unified Theory of Acceptance and Use of Technology (UTAUT) is a widely used model to assess users' acceptance which was proposed by Venkatesh^[28], based on eight related psychological and social theories/models. It contains four key constructs, which are effort expectancy (EE), performance expectancy (PE), social influence (SI) and facilitating conditions (FC)^{[28, 29]}. According to the UTAUT, the first three constructs are the core determinants of the users' behavioral intention (BI) while the last construct directly influences the actual behavior use (UB)^{[28, 29]}.

UTAUT can explain 70% variance in technology use^{[30, 31]}. Since introduction, it has been applied to explore the critical factors related to the prediction of users' intention and actual use of the technology in various health care settings, such as the health information system^[32–34], mobile medical technology^[35–37] and other clinical information systems^{[38, 39]}. To our knowledge, UTAUT has not been applied to the field of VTE CDSS.

The modified UTAUT and research hypothesis

To provide a context-related understanding of technology acceptance, the theoretical model must be identified and tested for different technologies and different user groups under certain circumstances^[40]. A better understanding of the determinants of the constructs in UTAUT and the associations between them would enable us to design organizational interventions that would increase user' acceptance and usage of new systems^{[36, 41]}. Therefore, our research extended UTAUT to include additional key determinants based on antecedent researches and explore how these determinants affect the usage of the target system among nurses.

User satisfaction (US) is considered to be an important mediating variable influencing users' acceptance of the information technology^{[39, 42]}. The Wixom and Todd (WT) model^[43] combined US with technology acceptance, within the model, information satisfaction and system satisfaction represent a user's attitude toward the use intention of information technology. In addition, Abdrbo^[44] believed that US can be measured with respect to the EE and PE constructs. Thus, in our research, we added the US variable and proposed the following hypotheses:

H1: EE has a positive effect on nurses' US.

H2: PE has a positive effect on nurses' US.

H3: US has a positive effect on nurses' BI to use VTE CDSS.

Facilitating conditions (FC) measures whether there is an existence of the organizational and technical environment can help remove the barriers to implement CDSS^{[28, 29]}. It was reported that when both EE and PE played a role, the effect of FC on the intention to use would not be significant and thus in the original model, the direct relationship between FC and BI was not included^[28]. The review of 174 studies incorporating UTAUT^[29] found that 48 of these original studies investigated the direct relationship between FC and BI, and 32 of those studies reported significant positive effects. It is worth noting that most of the included studies also found a significant effect of EE and PE^{[29, 45]} which was contrary to Venkatesh's view^[28]. So, it is necessary to re-examine the relationship between FC and BI, and then we proposed the following hypothesis:

H4: FC has a positive effect on nurses' BI to use VTE CDSS.

In 1977, Bandura^[46] proposed the self-efficacy (SE) based on social cognitive theory. In the field of information technology, Davis^[47] firstly discussed the influence of SE on students' intention to use a word processing software and found it was a vital determinant of the BI. Compeau^[48] defined computer SE as an individual's judgment on their ability to use information technology, which means SE is an individual's level of confidence in using information technology to complete specific tasks. Several studies^[49–51] demonstrated that computer SE played an essential role in predicting users' BI. Hence, we added the SE as a variable and proposed the following hypothesis:

H5: SE has a positive effect on nurses' BI to use VTE CDSS.

Additionally, we set up 5 hypotheses based on the original model:

H6: EE has a positive effect on nurses' BI to use VTE CDSS.

H7: PE has a positive effect on nurses' BI to use VTE CDSS.

H8: SI has a positive effect on nurses' BI to use VTE CDSS .

H9: FC has a positive effect on nurses' UB.

H10: BI has a positive direct effect on nurses' UB.

The modified UTAUT is presented in Fig. 1.

All methods were carried out in accordance with relevant guidelines and regulations.

Research settings and participants

The target participants were recruited from the First Affiliated Hospital of Nanjing Medical University where the VTE CDSS has been implemented since 1 year ago. Registered nurses who took care of patients directly were qualified to participate in this research. Nurses in the internship and advanced training period were excluded.

Research instrument

A questionnaire designed based on the previous research^[32–39] was utilized to investigate the determinant factors of nurses' acceptance of VTE CCDSS. We did a brief introduction of the research and obtained informed consent from participants at the beginning. The questionnaire consisted of two parts. The first part was to collect basic information of participants, including gender, age, education, work ward, organizational position, length of the use, and training experience. No personal identification information of participants was involved. The second part included 29 questions covered 7 constructs based on the modified UTAUT model. Each item was evaluated through a 5-point Likert scale ranging from 1 (representing strongly disagree) to 5 (representing strongly agree).

The structured questionnaire was pilot tested for the expression and comprehension of each item. We recruited 10 nurses to conduct a preliminary investigation. Then the questionnaire was adapted to their feedback. The final questionnaire is presented in Table 1.

Table 1

Construct with its measurement items
Constructs	Corresponding Items
Effort expectancy (EE)	EE.1 The operating interface of VTE CDSS is clear.
	EE.2 The contents of VTE CDSS are easy to understand and easy to use.
	EE.3 Learning to use VTE CDSS is ease for me.
	EE.4 It is easy and convenient for me to use VTE CDSS
	EE.5 It is easy for me to become skillful at using CDSS.
Performance expectancy (PE)	PE.1 Using CDSS helps me dynamically assess and monitor the risk of VTE in patients
	PE.2 Using CDSS helps me make clinical decisions on VTE prevention (different measures according to the risk stratification)
	PE.3 Using CDSS helps me promote my work efficiency.
	PE.4 Using CDSS helps me improve the quality of my work
Social influence (SI)	SI.1Hospital administrator (eg,. nursing department, special nursing unit, head nurse) think that I should use CDSS.
	SI.2 Colleagues around me (including doctors) think that I should use CDSS.
	SI.3 The surrounding leader or colleague who is a member of the hospital VTE group, think that i should use CDSS.
	SI.4 The surrounding leader or colleague who participated in the design of the CDSS, think that i should use CDSS.
Facilitating conditions (FC)	FC.1 I can get help from others when i have trouble in using the CDSS.
	FC.2 During my work, CDSS works steadily.
	FC.3 Hospital provides adequate training on the use of CDSS.
	FC.4 I have the resources necessary to use CDSS.
	FC.5 Hospital’s quality control results for CDSS promoted my use of the system
Self-efficacy (SE)	SE.1 I have a comprehensive knowledge of VTE prevention involved in CDSS.
	SE.2 I am confident that I can use the system correctly.
	SE.3I can skillfully use every medical information system in hospital.
User Satisfaction (US)	US.1 Information satisfaction: I am satisfied with the information provided by the modules of CDSS.
User Satisfaction (US)	US.2 System satisfaction: I am satisfied with the overall operating process of CDSS.
Behavior intention (BI)	BI.1 I intend to use CDSS in the future.
	BI.2 I would like to continue to learn more about CDSS.
	BI.3 I would like to recommend CDSS to others.
User behavior (UB)	UB.1 I'm used to using CDSS.
	UB.2 I will continue to use CDSS.
	UB.3 I have recommended CDSS to others.

Data collection

The ethical approval of this study was obtained through the ethics committee of the First Affiliated Hospital of Nanjing Medical University (2020-SR-373). A cross-sectional survey was conducted from 25th September to 15th October 2021. The questionnaires were distributed by the medical and nursing departments through the Questionnaire Star Platform online.

We did a brief introduction of the research and obtained informed consent from participants at the beginning. The researchers predetermined that only participants completed all options in the questionnaires would they be able to submit it. In addition, participants could withdraw from the research at any time during the filling out process not until they clicked the submitted button, and these incompleted questionnaires would be dismissed by the system. The completed questionnaires would be automatically sent back to our website (https://www.wjx.cn/weixinlogin.aspx). Any information participants provided were treated confidentially.

Statistical analysis

The first part of the statistical analysis was to test the reliability and validity of the measurement model. The second part was to validate the associations between the hypothesized constructs. We employed a Structural Equation Modeling (SEM)-Partial Least Squares (PLS) path modeling approach for data analysis^[52]. SEM is a widely accepted paradigm to gauge the validity of theories with empirical data. SmartPLS software version3.0, one of the widely used software applications for PLS-SEM, was used to test and validate the proposed model^[53]. Survey data were exported from the website online database in the excel form and then imported into SmartPLS software for the analysis. P value ≤ 0.05 (two-tailed) was considered to be statistically significant.

Descriptive statistical analysis

A total of 1100 questionnaires were recycled. The sample comprised 1081 women (98.27%) and 19 men (1.73%). The average age of nurses was 32.70 ± 6.978 and the length of working experience was 8.95 ± 7.608 year on average. The respondents worked in different departments in the hospital, for instance, respiratory medicine department, orthopedics department and intensive care unit. With respect to the education level of participants, 937 nurses (85.18%) had a bachelor’s degree, 11 nurses had a master or a doctor’s degree (1%) and only 152 nurses had a Junior College’s degree or below (13.82%). The descriptive statistics result is shown in Table 2.

Table 2

Demographic characteristics of participants
Variable	Description	Frequency(n%)
Gender	man	19 (1.73%)
Gender	woman	1081 (98.27%)
Age	32.70 ± 6.978
Education	Junior College and below	152 (13.82%)
	Bachelor	937 (85.18%)
	Master or above	11 (1%)
Department	Sugical Department	342 (31.09%)
	Medical Department	306 (27.82%)
	Geriatric Department	149 (13.55%)
	Emergency Department	8 (0.73%)
	Maternal and Child Department	84 (7.64%)
	Intensive Care Unit	129 (11.73%)
	operating theatre	26 (2.36%)
	Medical technologic Department	43 (3.91%)
	Infectious Disease Department	13 (1.18%)
Length of working (year)	8.95 ± 7.608
Professional title	Nurse	262 (23.82%)
	Nurse practitioner	502 (45.64%)
	Nurse-in-charge	263 (23.91%)
	Associate senior nurse	61(5.55%)
	Full senior nurse	12(1.09%)
Position	department head nurse	0
	Head nurse	69 (6.27%)
	Member of VTE special group	14 (1.27%)
	None	1017 (92.45%)
Length of CDSS usage	<6 month	191 (17.36%)
	6 to 12 month	308 (28.00%)
	12 to 18 month	298 (27.09%)
	>18 month	303 (27.55%)
Training frequency	Never	248 (22.55%)
	1 to 2 per year	703(63.91%)
	3 to 5 per year	104 (9.45%)
	>5 per year	45 (4.09%)

The measurement model

In terms of the measurement model, we mainly examined internal reliability, convergent and discriminant validity^{[53, 54]}. Cronbach's alpha and composite reliability (CR) were used to evaluate the internal reliability of the UTAUT constructs^[53]. The level of 0.70 is the threshold for acceptable internal consistency^{[33, 53]}. As shown in Table 3, all constructs' Cronbach's alpha and CR are above 0.7. Convergent validity can be estimated using average variance extracted(AVE)^{[33, 53]}. The result demonstrated that AVE values ranged from 0.847 to 0.928, greater than the threshold value of 0.50^[53]. In addition, depicted in Table 4, we found that the factor loading of each item on it respective associated construct was over 0.4, greater than the acceptable levels^{[53, 55]}, so there was no need to delete any item.

The discriminant validity was accessed by the cross-loading matrix and the square root of the AVE. From Table 4 matrix, we can see that all the factor loadings (bold figures) are larger than the cross-loadings, which means a satisfactory discriminant validity. Besides, as Fornell-Larcker criterion^[52], good discriminant validity must possess that the AVE square root of a construct be greater than its correlation with other constructs. As shown in Table 5, the diagonal elements are all greater than the figures in corresponding rows and columns, which meets the above criterion.

Table 3

Construct Reliability and Validity
Constructs	Cronbachcs Alpha	Composite Reliability	Average Variance Extracted (AVE)
BI	0.961	0.975	0.928
UB	0.908	0.943	0.847
EE	0.969	0.976	0.891
FC	0.960	0.969	0.861
PE	0.965	0.973	0.877
US	0.955	0.978	0.957
SE	0.932	0.957	0.881
SI	0.953	0.966	0.876

Table 4

Results of items loadings and the cross-loadings
Constructs	Item	EE	PE	SI	FC	SE	US	BI	UB
BI	BI.1	0.784	0.844	0.832	0.858	0.842	0.838	0.964	0.877
	BI.2	0.753	0.825	0.818	0.847	0.815	0.822	0.966	0.880
	BI.1	0.728	0.829	0.822	0.844	0.810	0.830	0.960	0.892
UB	BU.1	0.756	0.789	0.791	0.848	0.826	0.804	0.881	0.952
	BU.2	0.770	0.827	0.819	0.855	0.831	0.823	0.917	0.956
	BU.3	0.576	0.656	0.691	0.699	0.695	0.700	0.716	0.848
EE	EE.1	0.938	0.748	0.722	0.755	0.757	0.723	0.724	0.703
	EE.2	0.948	0.777	0.744	0.777	0.751	0.755	0.731	0.717
	EE.3	0.955	0.796	0.762	0.787	0.763	0.767	0.746	0.721
	EE.4	0.961	0.804	0.765	0.808	0.788	0.779	0.765	0.743
	EE.5	0.917	0.801	0.740	0.775	0.770	0.736	0.732	0.742
FC	FC.1	0.758	0.834	0.862	0.908	0.800	0.826	0.799	0.783
	FC.2	0.797	0.835	0.836	0.924	0.836	0.826	0.840	0.824
	FC.3	0.767	0.804	0.815	0.929	0.850	0.826	0.809	0.810
	FC.4	0.765	0.806	0.822	0.944	0.861	0.845	0.816	0.809
	FC.5	0.749	0.809	0.826	0.934	0.869	0.844	0.829	0.831
PE	PE.1	0.810	0.933	0.811	0.813	0.783	0.792	0.798	0.779
	PE.2	0.787	0.944	0.830	0.824	0.802	0.791	0.806	0.781
	PE.3	0.769	0.936	0.829	0.815	0.784	0.797	0.803	0.779
	PE.4	0.774	0.936	0.841	0.838	0.775	0.794	0.834	0.772
	PE.5	0.757	0.933	0.853	0.835	0.783	0.798	0.806	0.765
US	SA.1	0.770	0.826	0.830	0.877	0.884	0.978	0.836	0.819
US	SA.2	0.789	0.833	0.832	0.880	0.884	0.979	0.850	0.835
SE	SE.1	0.761	0.805	0.814	0.897	0.929	0.838	0.813	0.817
	SE.2	0.782	0.785	0.789	0.852	0.953	0.845	0.815	0.816
	SE.3	0.740	0.771	0.766	0.808	0.934	0.862	0.775	0.775
SI	SI.1	0.775	0.841	0.916	0.842	0.790	0.795	0.819	0.794
	SI.2	0.725	0.849	0.935	0.821	0.767	0.781	0.790	0.767
	SI.3	0.745	0.831	0.949	0.851	0.799	0.805	0.809	0.799
	SI.4	0.714	0.809	0.943	0.843	0.794	0.799	0.783	0.769

Table 5

Discriminant Validity
Construct	EE	PE	SI	FC	US	SE	BI	UB
EE	0.944
PE	0.832	0.936
SI	0.791	0.890	0.936
FC	0.827	0.881	0.897	0.928
SE	0.811	0.839	0.842	0.909	0.938
US	0.797	0.848	0.850	0.898	0.904	0.978
BI	0.784	0.864	0.856	0.882	0.854	0.862	0.963
UB	0.768	0.828	0.836	0.875	0.856	0.846	0.917	0.920

Hypothesis testing

The structural model was then developed to investigate the relationships and the path coefficients between the constructs in the research model. The PLS statistical results are shown in Table 6. The results showed that only one hypothesis that EE has a positive effect on BI was rejected (r = 0.010, p = 0.785). The relationship between PE and BI (r = 0.254, p = 0.000), SI and BI (r = 0.136, p = 0.047), FC and BI (r = 0.245, p = 0.000), SE and BI (r = 0.121, p = 0.048), US and BI (r = 0.193, p = 0.001) were all significant. Among these, PE is the strongest predictor of BI, with the path coefficients of 3.801. Although the relationship between EE and BI is unsupported, it was statistically shown that EE (r = 0.296, p = 0.000) indirectly influenced nurses' intention with the US as the mediator. And PE (r = 0.602, p = 0.000) played the same indirect role on BI as EE did. Partial least squares path modeling revealed that the variance of BI explained by the modified UTAUT model was 83.0% which is higher than the original test result by the Venkatesh^[28] (the variance of BI explained by the model was almost 70%). Additionally, FC and BI directly influenced the UB and all constructs in the model explained 86.0% variance of the UB.

Table 6

Results of hypothesis tests and effect sizes
Hypothesis	path	t-value	R	P	Hypothesis results
H1	EE-> US	5.650	0.296	0.000	Accepted
H2	PE-> US	11.743	0.602	0.000	Accepted
H3	US-> BI	3.238	0.193	0.001	Accepted
H4	FC -> BI	3.735	0.245	0.000	Accepted
H5	SE-> BI	1.975	0.121	0.048	Accepted
H6	EE-> BI	0.272	0.010	0.785	Rejected
H7	PE -> BI	3.801	0.254	0.000	Accepted
H8	SI -> BI	1.989	0.136	0.047	Accepted
H9	FC -> UB	6.070	0.298	0.000	Accepted
H10	BI-> UB	14.179	0.654	0.000	Accepted

This research utilized UTAUT as a theoretical basis and integrated related variables to investigate the influential factors associated with the acceptance of VTE CDSS among nurses. The aforementioned statistical results all showed the effectiveness of the modified UTAUT model in predicting nurses' acceptance of VTE CDSS. The findings of the research indicated that: 1. EE and PE with US serving as a mediator significantly affected nurses' BI to use VTE CDSS; 2. nurses' intention to use CDSS was associated with their perceived ease and usefulness of use, their significant others concerning about VTE CDSS, their perception about the organizational and technical environment, their confidence and satisfaction over the usage process; 3. lastly, FC and BI had a significant direct effect on nurses' actual use of VTE CDSS. As we know, this is the first study to investigate the nurses' acceptance of VTE CDSS.

PE was the strongest factor in predicting nurses' intention to use VTE CDSS. This result is consistent with the antecedent researches^{[36, 38]}. CDSS can positively improve nurses' knowledge and enhance their awareness of VTE prevention as it provides instruction as soon as patients' information is recorded into the system^{[23, 25]}. The embedded electronic alert can also remind nurses to reassess the VTE risk within the specified time or when the patients' condition changes^{[10, 17]}. Porat^[55] stated that medical staff could give patients more personalized education while using the CDSS. What's more, previous research^[56–58] indicated that CDSS could help improve evidence-based clinical practice overall by improving the medical staff's performance and work efficiency. However, there still exist research suggesting that medical staff considered it as a threat to their clinical autonomy since they used to depend on their experience and preference to provide patient care^{[59, 60]}. They also reported that it took them too much time to document in the CDSS and thus their attention to patients decreased^{[25, 59, 60]}. During the literature review of VTE CDSS, we also noticed a phenomenon called alert fatigue^[61]. This means users have a tendency to ignore or over-ride the triggered tool due to the frequent alerts^{[18, 61]}. In general, most current studies showed that the advantages of the CDSS outweigh the disadvantages^[18,59−61]. Since all users tend to agree PE construct is of great importance in the nurses' intention to use and maintain use in the future^{[36, 54]}, the VTE CDSS still needs continuous improvement to remove the barriers of usage and increase its perceived usefulness.

However, our study did not find a significant relationship between EE and BI which may be apparently surprising, since EE is a core construct of UTAUT and is considered as a determinant of BI based on many antecedent empirical tests^[28–30], especially in the elderly group^[54]. Several studies also found a similar result^{[37, 46, 47]}, and one possible explanation for this is that our VTE CDSS is easy to use and thus the effect of EE is not salient^[47]. Another explanation is our participants were relatively young(with the mean age of 30) and well educated (85.18% with a bachelor's degree), so using an information system may not be a complicated job. Nonetheless, these statistical results do not mean that the EE construct does not help explain the adoption of VTE CDSS. EE has an indirect effect on BI, and this effect is mediated by users' satisfaction. Both PE and EE correlate with users' satisfaction (US), and then the US directly influences BI. Jessica^[62] have ever emphasized the importance of user-centered design CDSS, which means users could easily learn without training and use it correctly. To promote users' satisfaction, the system must be user friendly and useful with both satisfying information and system quality^{[59, 62]}. With increased satisfaction, the users' intention will be enhanced^{[59, 60, 62]}.

SI had a significant relationship with BI. A survey conducted by Lu^[32] found a prominent effect of SI on the adoption of hospital information system among nurses. What the two studies have in common is that both hospitals have encouraged and supported the nurses to adopt the systems^[32]. It is assumed that medical staff refused to use the system for the lack of policy regulations by the hospital administrators^[63]. Hence, the regulatory issues are vital to the nurses' adoption of the VTE CDSS^{[33, 63]}. Additionally, the VTE CDSS generated documentation can be imported into the clinical note and the successor can view the previous worker's documentation, which is conducive to hand-over work and fully grasp the patient's condition^[64]. Gradually, consistent consensus on the use of the system was achieved among nurses and they became more familiar with it.

The results established the link between FC with BI and UB. FC is considered as a highly rated requirement by all medical staff^{[32, 64]}. These organizational and technical factors are significant for the nurses, for instance, whether the infrastructure is sufficient to support the use of CDSS (e.g. network with good signals, fully equipped computers) or whether the users receive adequate training^{[18, 63]}. As we can see, among our participants, there still left 22.55% of participants reported that they had never received any training. What's more, among the training people, more than half stated that they only received 1 or 2 times training. Both hospital administrators and manufacturers should jointly provide nurses with adequate training and continuous assistance services to support the implementation of VTE CDSS^{[18, 32, 63]}.

The high correlation coefficient between BI and SE highlight the importance of SE. SE here refers to the computer self-efficacy which has been proved to play an essential role in predicting users' intention and behavior when using computer^{[33, 49]}. It is found to be associated with the computer literacy and computer anxiety^{[48, 49]}. People exhibit higher computer anxiety tend to show lower computer literacy, which means the two concepts play a counterproductive role^[48–50]. It is believed that the use of CDSS is dependent on computer literacy and lack of technical proficiency can be hindering^[18]. Computer anxiety is an unpleasant feeling which contains negative emotional states during the interaction with computer^{[48, 49]}. Women exhibit lower computer literacy and higher computer anxiety than men, so women ultimately have lower computer SE^[53]. In our survey, women occupied 98.27% of all participants, so we should pay more attention to such users' computer SE. Several studies indicated that the increased level of education would enhance the level of SE^[64]. Providing adequate and related training courses on the new technology will help nurses decrease their anxiety and increase their confidence to use the system^{[64, 65]}. For better acceptance of VTE CDSS, they should not only improve their knowledge of the system, but also increase their professional knowledge since many modules in VTE CDSS need their own expertise.

There are some limitations of this study. First, our study is a cross-sectional survey, therefore we can only determine association but not establish a causal relationship between the variables and outcomes. Second, our study used a web-based survey, all constructs were measured as self-report measures. The participants completed the questionnaires with respect to their working experiences and perceptions. This may lead to common variance bias and the results must be interpreted with caution. Third, nowadays in China, the VTE CDSS has not been widely used. Implementing a CDSS is a complex undertaking of successive stage and the length of use is inconsistent among various hospitals and departments, so this survey was carried out in a single site. We did not take the characteristic differences between different hospitals and different types of VTE CDSS into account. Hence, it is necessary to conduct further studies covered more hospitals in more regions in the future.

Our study supports the use of the modified UTAUT to predict the factors associated with nurses' acceptance of VTE CDSS. The findings indicated that performance expectancy is the most important determinant of nurses' intention to use CDSS. Users' needs and expectations need to be fully considered during the development and improvement of VTE CDSS. The electronic alerts should be set up in a way that nurses do not find them obstructive or threatening. Besides, hospital administrators should provide adequate training and infrastructure support. This research makes significant contributions to the VTE CDSS implementation and thromboprophylaxis research. Nowadays, CDSS is in its initial stage in developing country. Future research can assess the users' acceptance of the system in a quantitative way and if the users' acceptance do not reach the predefined threshold, the system should be subjective to the adaptive redesign.

1. Ethics approval and consent to participate

The ethical approval of this study was obtained through the ethics committee of the First Affiliated Hospital of Nanjing Medical University (2020-SR-373).

Informed consent was obtained from all individual participants included in the study. Details that might disclose the identity of the subjects under study should be omitted.

2. Consent for publication

Not applicable.

3. Availability of data and materials

All data generated or analysed during this study are included in this published article.

4. Competing interests

The authors declare that they have no competing interests.

5. Funding

This work was supported by the University's Advantageous Subject Construction Project "Nursing", Jiangsu (Suzheng award〔2018〕87), Medical Innovation Team of Jiangsu Province（CXTDA2017019）and University's philosophy and social science research in Jiangsu Province（2019SJA0302）.

6. Authors’ contributions

All authors have made substantial contributions to the study.

Study conception and design: Huixian Zha; Kouying Liu

Acquisition of data: Huixian Zha; Kouying Liu; Rong Wang; Ling Jiang, Hongyuan Yan; Weiping Xie

Analysis and interpretation of data: Huixian Zha; Ting Tang; Bei Dou; Yueheng Yin

Drafting the article: Huixian Zha

Critical revision: Huixian Zha; Kouying Liu; Ting Tang; Yueheng Yin; Bei Dou; Rong Wang; Ling Jiang; Hongyuan Yan; Weiping Xie

Final approval of the version to be submitted: Huixian Zha; Kouying Liu; Ting Tang; Yueheng Yin; Bei Dou; Rong Wang; Ling Jiang; Hongyuan Yan; Weiping Xie

7. Acknowledgments

We appreciated the nurses participated in the study for their assistance.

Schulman S, Ageno W, Konstantinides S V. Venous thromboembolism: Past, present and future[J]. Thromb Haemost, 2017,117(7): 1219-1229. doi: 10.1160/TH16-10-0823
Bruni-Fitzgerald K R. Venous thromboembolism: An overview[J]. J Vasc Nurs, 2015,33(3): 95-99. doi: 10.1016/j.jvn.2015.02.001
Brenner B, Hull R, Aray R, et al. Evaluation of unmet clinical needs in prophylaxis and treatment of venous thromboembolism in high-risk patient groups: cancer and critically ill[J]. Thromb J, 2019,17: 6. doi: 10.1186/s12959-019-0196-6
Zhang Z, Lei J, Shao X, et al. Trends in Hospitalization and In-Hospital Mortality From VTE, 2007 to 2016, in China[J]. Chest, 2019, 155(2): 342-353. doi: 10.1016/j.chest.2018.10.040
Bartholomew J R. Update on the management of venous thromboembolism[J]. Cleve Clin J Med, 2017,84(12 Suppl 3): 39-46. doi: 10.3949/ccjm.84.s3.04
Suh J, Desai A, Desai A, et al. Adherence to thromboprophylaxis guidelines in elderly patients with hospital acquired venous thromboembolism: a case control study[J]. J Thromb Thrombolysis, 2017, 43(2): 172-178. doi: 10.1007/s11239-016-1432-6
Hunt B J. Preventing hospital associated venous thromboembolism[J]. BMJ, 2019,365: l4239. doi: 10.1136/bmj.l4239
Dawoud D, Lewis S, Glen J, et al. Venous thromboembolism in over 16s: reducing the risk of hospital-acquired deep vein thrombosis or pulmonary embolism. NICE guideline [NG89]. 2018.
Haut E R, Lau B D, Kraus P S, et al. Preventability of Hospital-Acquired Venous Thromboembolism[J]. JAMA Surgery, 2015, 150(9): 912-915. doi:10.1001/jamasurg.2015.1340
Kahn S R, Morrison D R, Diendéré G, et al. Interventions for implementation of thromboprophylaxis in hospitalized patients at risk for venous thromboembolism[J]. Cochrane Database of Systematic Reviews, 2018(4). doi:10.1002/14651858.CD008201.pub3
Chaudhary R, Damluji A, Batukbhai B, et al. Venous Thromboembolism Prophylaxis: Inadequate and Overprophylaxis When Comparing Perceived Versus Calculated Risk[J]. Mayo Clin Proc Innov Qual Outcomes, 2017, 1(3): 242-247. doi:10.1016/j.mayocpiqo.2017.10.003
Zhai Z, Kan Q, Li W, et al. VTE Risk Profiles and Prophylaxis in Medical and Surgical Inpatients: The Identification of Chinese Hospitalized Patients' Risk Profile for Venous Thromboembolism (DissolVE-2)-A Cross-sectional Study[J]. Chest, 2019, 155(1): 114-122. doi: 10.1016/j.chest.2018.09.020
Yan T, He W, Hang C, et al. Nurses' knowledge, attitudes, and behaviors toward venous thromboembolism prophylaxis: How to do better[J]. Vascular, 2020: 130584982. doi:10.1177/1708538120933782
Ma Y F, Xu Y, Chen Y P, et al. Nurses' objective knowledge regarding venous thromboembolism prophylaxis: A national survey study[J]. Medicine (Baltimore), 2018,97(14): e338. doi:10.1097/MD.0000000000010338
Lloyd N S, Douketis J D, Cheng J, et al. Barriers and potential solutions toward optimal prophylaxis against deep vein thrombosis for hospitalized medical patients: a survey of healthcare professionals[J]. J Hosp Med, 2012,7(1): 28-34. doi:10.1002/jhm.929
Rre Y, Chen D, Tuo B, et al. Study on venous thromboembolism control system based on clinical decision support system[J].Chin J Hosp Admin, 2019(07):536-539. doi: 10.3760/cma.j.issn.1000-6672.2019.07.003
Borab Z M, Lanni M A, Tecce M G, et al. Use of Computerized Clinical Decision Support Systems to Prevent Venous Thromboembolism in Surgical Patients: A Systematic Review and Meta-analysis[J]. JAMA Surg, 2017,152(7): 638-645. doi: 10.1001/jamasurg.2017.0131
Sutton R T, Pincock D, Baumgart D C, et al. An overview of clinical decision support systems: benefits, risks, and strategies for success[J]. NPJ Digit Med, 2020,3: 17. doi: 10.1038/s41746-020-0221-y
Adams P, Riggio J M, Thomson L, et al. Clinical decision support systems to improve utilization of thromboprophylaxis: a review of the literature and experience with implementation of a computerized physician order entry program[J]. Hosp Pract (1995), 2012,40(3): 27-39. doi: 10.3810/hp.2012.08.987
Khong P C, Holroyd E, Wang W. A Critical Review of the Theoretical Frameworks and the Conceptual Factors in the Adoption of Clinical Decision Support Systems[J]. Comput Inform Nurs, 2015,33(12): 555-570. doi:10.1097/CIN.0000000000000196
Dowding D, Mitchell N, Randell R, et al. Nurses' use of computerised clinical decision support systems: a case site analysis[J]. J Clin Nurs, 2009,18(8): 1159-1167. doi:10.1111/j.1365-2702.2008.02607.x
Esmaeilzadeh P, Sambasivan M, Kumar N, et al. Adoption of clinical decision support systems in a developing country: Antecedents and outcomes of physician's threat to perceived professional autonomy[J]. Int J Med Inform, 2015,84(8): 548-560. doi:10.1016/j.ijmedinf.2015.03.007
Khalifa M, Alswailem O. Clinical Decision Support Knowledge Management: Strategies for Success[J]. Studies in health technology and informatics, 2015, 213:67-70.
Sharifian R, Askarian F, Nematolahi M, et al. Factors influencing nurses' acceptance of hospital information systems in Iran: application of the Unified Theory of Acceptance and Use of Technology[J]. Health Inf Manag, 2014,43(3): 23-28. doi: 10.1177/183335831404300303
Aarts J, Gorman P. IT in health care: sociotechnical approaches "To Err is System"[J]. Int J Med Inform, 2007, 76 Suppl 1: S1-S3. doi: 10.1016/S1386-5056(07)00078-0
Ammenwerth E. Technology Acceptance Models in Health Informatics: TAM and UTAUT[J]. Stud Health Technol Inform, 2019,263: 64-71. doi:10.3233/SHTI190111
Hsieh P J, Lai H M, Ma C C, et al. An Extended Expectation-Confirmation Model for Mobile Nursing Information System Continuance[J]. Res Theory Nurs Pract, 2016,30(4): 282-301. doi:10.1891/1541-6577.30.4.282
Venkatesh V, Davis F D. A Theoretical Extension of the Technology Acceptance Model: Four Longitudinal Field Studies[J]. Management Science, 2000,46(2): 186.
Williams M D, Rana N P, Dwivedi Y K. The unified theory of acceptance and use of technology (UTAUT): a literature review[J]. Journal of Enterprise Information Management, 2015,28(3): 443-488.
Dwivedi Y K, Rana N P, Tamilmani K, et al. A meta-analysis based modified unified theory of acceptance and use of technology (meta-UTAUT): a review of emerging literature[J]. Curr Opin Psychol, 2020,36: 13-18. doi: 10.1016/j.copsyc.2020.03.008
Magsamen-Conrad K, Wang F, Tetteh D, et al. Using Technology Adoption Theory and a Lifespan Approach to Develop a Theoretical Framework for eHealth Literacy: Extending UTAUT[J]. Health Commun, 2019: 1-12. doi: 10.1080/10410236.2019.1641395
Lu, Lin, Zhou, et al. Assessment of the social influence and facilitating conditions that support nurses' adoption of hospital electronic information management systems (HEIMS) in Ghana using the unified theory of acceptance and use of technology (UTAUT) model[J]. BMC medical informatics and decision making, 2019, 19(1):230-230.
Bawack R E, Kala K J. Adequacy of UTAUT in clinician adoption of health information systems in developing countries: The case of Cameroon[J]. Int J Med Inform, 2018,109: 15-22. doi: 10.1016/j.ijmedinf.2017.10.016
Sheikhtaheri A, Malekzadeh S, Hashemi N S, et al. Effects of Using Hospital Information Systems on Nurses' Individual Performance: A Study on Influential Factors[J]. Stud Health Technol Inform, 2020,271: 161-167. doi: 10.3233/SHTI200092
Jewer J. Patients' intention to use online postings of ED wait times: A modified UTAUT model[J]. Int J Med Inform, 2018,112: 34-39. doi: 10.1016/j.ijmedinf.2018.01.008
Zhang Y, Liu C, Luo S, et al. Factors Influencing Patients' Intentions to Use Diabetes Management Apps Based on an extended Unified Theory of Acceptance and Use of Technology Model: Web-Based Survey[J]. J Med Internet Res, 2019,21(8): e15023. doi: 10.2196/15023
Liu D, Maimaitijiang R, Gu J, et al. Using the Unified Theory of Acceptance and Use of Technology (UTAUT) to Investigate the Intention to Use Physical Activity Apps: Cross-Sectional Survey[J]. JMIR Mhealth Uhealth, 2019,7(9): e13127. doi: 10.2196/13127
Kim S, Lee K H, Hwang H, et al. Analysis of the factors influencing healthcare professionals' adoption of mobile electronic medical record (EMR) using the unified theory of acceptance and use of technology (UTAUT) in a tertiary hospital[J]. BMC Med Inform Decis Mak, 2016,16: 12. doi: 10.1186/s12911-016-0249-8
Maillet É, Mathieu L, Sicotte C. Modeling factors explaining the acceptance, actual use and satisfaction of nurses using an Electronic Patient Record in acute care settings: an extension of the UTAUT[J]. Int J Med Inform, 2015,84(1): 36-47. doi: 10.1016/j.ijmedinf.2014.09.004
Boyle L, Grainger R, Hall R M, et al. Use of and Beliefs About Mobile Phone Apps for Diabetes Self-Management: Surveys of People in a Hospital Diabetes Clinic and Diabetes Health Professionals in New Zealand[J]. JMIR Mhealth Uhealth, 2017,5(6): e85. doi: 10.2196/mhealth.7263
Venkatesh V, Morris M, Davis G, et al. User Acceptance of Information Technology: Toward a Unified View[J]. MIS Quarterly, 2003,27: 425-478.
Ho K F, Ho C H, Chung M H. Theoretical integration of user satisfaction and technology acceptance of the nursing process information system[J]. PLoS One, 2019,14(6): e217622. doi: 10.1371/journal.pone.0217622
Todd W P A. A Theoretical Integration of User Satisfaction and Technology Acceptance[J]. Information Systems Research, 2005, 16(1):85-102.
Abdrbo A A , Hudak C A , Anthony M K , et al. Moderating and mediating roles of nurses' beliefs: information systems use among Ohio nurses[J]. West J Nurs Res, 2009,31(1): 110-127. doi: 10.1177/0193945908325078
Kijsanayotin B, Pannarunothai S, Speedie S M. Factors influencing health information technology adoption in Thailand's community health centers: applying the UTAUT model[J]. Int J Med Inform, 2009,78(6): 404-416.
Bandura A. Self-efficacy: toward a unifying theory of behavioral change[J]. Psychol Rev, 1977,84(2): 191-215. doi: 10.1037//0033-295x.84.2.191
Davis F D , Bagozzi R P , Warshaw P R . User Acceptance of Computer Technology: A Comparison of Two Theoretical Models[J]. Management Science, 1989, 35(8):982-1003.
Compeau D R, Higgins C A. Computer Self-Efficacy: Development of a Measure and Initial Test[J]. MIS Quarterly, 1995,19(2): 189-211.
Chung M H, Ho C H, Wen H C. Predicting intentions of nurses to adopt patient personal health records: A structural equation modeling approach[J]. Comput Methods Programs Biomed, 2016,136: 45-53. doi: 10.1016/j.cmpb.2016.08.004
Scott J E, Walczak S. Cognitive engagement with a multimedia ERP training tool: Assessing computer self-efficacy and technology acceptance[J]. Information & Management, 2009,46(4): 221-232.
Singh A, Bhadauria V, Jain A, et al. Role of gender, self-efficacy, anxiety and testing formats in learning spreadsheets[J]. Computers in Human Behavior, 2013,29(3): 739-746.
Hair J F, Hult G T M, Ringle C M, et al. A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM), 2nd edition[M]. 2016.
Oliver Götz, Liehr-Gobbers K , Krafft M. Evaluation of Structural Equation Models Using the Partial Least Squares (PLS) Approach[J]. handbook of partial least squares, 2009.
Hoque R, Sorwar G. Understanding factors influencing the adoption of mHealth by the elderly: An extension of the UTAUT model[J]. Int J Med Inform, 2017,101: 75-84. doi: 10.1016/j.ijmedinf.2017.02.002
Porat T, Delaney B, Kostopoulou O. The impact of a diagnostic decision support system on the consultation: perceptions of GPs and patients[J]. BMC Med Inform Decis Mak, 2017,17(1): 79. doi: 10.1186/s12911-017-0477-6
Wagholikar K B , Hankey R A , Decker L K , et al. Evaluation of the effect of decision support on the efficiency of primary care providers in the outpatient practice[J]. J Prim Care Community Health, 2015,6(1): 54-60. doi: 10.1177/2150131914546325
Silveira D V, Marcolino M S, Machado E L, et al. Development and Evaluation of a Mobile Decision Support System for Hypertension Management in the Primary Care Setting in Brazil: Mixed-Methods Field Study on Usability, Feasibility, and Utility[J]. JMIR Mhealth Uhealth, 2019,7(3): e9869. doi: 10.2196/mhealth.9869
Alexander, Goehler, Christopher, et al. Clinical Decision Support for Ordering CTA-PE Studies in the Emergency Department-A Pilot on Feasibility and Clinical Impact in a Tertiary Medical Center[J]. Acad Radiol, 2019,26(8): 1077-1083. doi: 10.1016/j.acra.2018.09.009
Kux B R, Majeed R W, Ahlbrand J, et al. Factors Influencing the Implementation and Distribution of Clinical Decision Support Systems (CDSS)[J]. Studies in Health Technology & Informatics, 2017,243: 127-131.
Khairat S, Marc D, Crosby W, et al. Reasons For Physicians Not Adopting Clinical Decision Support Systems: Critical Analysis[J]. Jmir Medical Informatics, 2018,6(2): e24. doi: 10.2196/medinform.8912
Press A, Khan S, Mccullagh L, et al. Avoiding alert fatigue in pulmonary embolism decision support: a new method to examine 'trigger rates'[J]. Evid Based Med, 2016,21(6): 203-207. doi: 10.1136/ebmed-2016-110440
Ray J M, Ahmed O M, Solad Y, et al. Computerized Clinical Decision Support System for Emergency Department–Initiated Buprenorphine for Opioid Use Disorder: User-Centered Design[J]. JMIR Human Factors, 2019,6(1). doi: 10.2196/13121
Handayani P W, Hidayanto A N, Pinem A A, et al. Hospital information system user acceptance factors: User group perspectives[J]. Inform Health Soc Care, 2018,43(1): 84-107. doi: 10.1080/17538157.2016.1269109
Grout R W, Cheng E R, Carroll A E, et al. A six-year repeated evaluation of computerized clinical decision support system user acceptability[J]. International Journal of Medical Informatics, 2018,112(APR.): 74-81. doi: 10.1016/j.ijmedinf.2018.01.011
Lee C L, Huang M K. The influence of computer literacy and computer anxiety on computer self-efficacy: the moderating effect of gender[J]. Cyberpsychol Behav Soc Netw, 2014,17(3): 172-180. doi: 10.1089/cyber.2012.0029

No competing interests reported.

Acceptance of Clinical Decision Support System to Prevent Venous Thromboembolism Among Nurses: an Extension of the UTAUT Model

Status:

Version 1

Abstract

Figures

Summary points

Introduction

Theoretical framework and research hypothesis

The original model

The modified UTAUT and research hypothesis

Methods

Research settings and participants

Research instrument

Data collection

Statistical analysis

Results

Descriptive statistical analysis

The measurement model

Hypothesis testing

Discussion

Limitations

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1