Examining Trust as a Key Determinant of eHealth Adoption in Malawi

Background There is value in having the general public take initiative in taking care of personal health. With the heavy burden of finances and shortage of healthcare personnel, patient-centered healthcare is increasingly becoming important especially through eHealth. The way technology is accepted and utilized may have significant hypothetical and concrete inferences. Thus, eHealth, like any other technology, has little value unless it is used. Nevertheless, there are many factors that potentially promote or hinder uptake and use of eHealth services. This study particularly focused on the role that trust plays in determining an individual’s decision to use eHealth services. Methods Using the Unified Theory of Acceptance and Use of Technology (UTAUT) model as the primary model of inference, two new constructs were introduced to determine whether trust is a significant contributor in consumers’ decision to use eHealth. Through convenience sampling, participant responses were collected over a period of 6 weeks and evaluated using Structural Equation Modeling (SEM) technique. Results A total of 400 responses were collected and outcomes of the analysis showed that Performance Expectancy (PE), Effort Expectancy (EE), Social Influence (SI), and Facilitating Conditions (FC), had an affirmative effect on Behavioral Intentions to use eHealth. Performance Expectancy indirectly had a positive effect on Behavioral Intentions to use eHealth services via Trust in Internet (ToI) and Trust in Online Healthcare Providers (ToH). However, ToI had an insignificant effect while ToH had a positive effect. Conclusion Trust is indeed an important element in a user’s determination to use

Conclusion Trust is indeed an important element in a user's determination to use eHealth services. However, it depends on what exactly it is that users place trust in.
Consumers are less trusting in the internet to facilitate accessing health related services but are more trusting in online healthcare service providers to ably assist with relevant services. The study also shows that it is essential for key stakeholders such as public policy actors and web designers to take into consideration specific target groups and user preferences which will enhance greater engagement of eHealth services.

Background
During the 1990s, as the internet penetrated the public sphere, a number of e-terms such as e-government, e-learning, e-mail, e-banking, and e-commerce began to appear offering new ways to conduct traditional tasks. The genesis of eHealth created an indispensable term and epitomized the potential for ICT to advance health and healthcare systems [1]. Today, Information and Communication Technology (ICT) is highly relevant to the healthcare sector through eHealth as it helps people have more insight into personal health through a personal digital healthcare environment [2]. Websites providing healthcare services present an important opportunity to achieve personal healthcare management. As the internet becomes increasingly accessible and available to consumers through phones or computers, patient-centered healthcare becomes a greater possibility [3].
However, user acceptance and use of eHealth is generally a challenge around the world [4,5]. With eHealth technologies becoming increasingly ubiquitous, research into determinants of effective recognition, reception, and utilization is becoming more important [6,7]. Generally, researchers have theorized technology acceptance as resulting from a psychological process that users experience in deciding to use a technology [8]. Furthermore, scholars affirm that consumer attitudes to technology, and particularly trust attitude, has an effect on how that technology is used [9,10,11]. Since health is a highly personal aspect of life, it is necessary to evaluate the technology comprehensively to enhance adoption and utilization. This evaluation relatively involves trust. Basically, research into trust in eHealth is actually an investigation into the association between individuals and technology [12].
In Malawi, use of the internet as a means of accessing healthcare remains in its infancy. Therefore, this research proposes to examine the relationship between determinants of eHealth use and consumers' intention to use eHealth via consumer trust. Scholars argue that there are two objectives to trust in technology; one objective is service provider offering services, the other objective is the device bridging the service provider and consumer which in this case is through the internet [13,14]. While it is important to believe that an online service is trustworthy, it does not lead to automatic adoption of the service. After all, as demonstrated by the UTAUT model, other factors influence adoption as well [15,4].
Findings indicate that in terms of variance (variance R 2 ), the strength of different models to explain intentions has resulted in a lowest score of 0.36 (Theory of Reasoned Action; Social Cognitive Theory) to the highest score of 0.69 (UTAUT). This therefore indicates the powerful explanatory ability of the UTAUT in relation to determinants of behavioral intentions [16]. This study therefore utilizes the Unified Theory of Acceptance and Use of Technology (UTAUT) as the primary model with slight alterations. Furthermore, two constructs of Trust in Internet (ToI) and Trust in Online Healthcare Providers (ToH) are incorporated.

Theoretical Model
The UTAUT is the primary model for this research with several alterations.

Behavioral intention to use (BI)
Researchers have proven that behavioral intention to use and actual technology usage are related. Thus, they propose that combining behavioral intention and use behavior to form "behavioral intention to use" makes sense [17]. Therefore, "behavioral intention to use" is used in this research as the dependent variable to demonstrate the concrete usage of eHealth services.

Performance Expectancy (PE)
Performance Expectancy is "the extent to which an individual believes that utilizing the technology will assist him or her to realize benefits in job performance" [18].
The value of an eHealth service is in the degree to which the facility affords a meaningful contribution to a patient's health. After all, perceived usefulness is a major part of trust and in this case health related usefulness. Thus, it is expected that an effective eHealth service will provide a healthier balance between outcome and effort required [4]. Consumers' positive attitude toward eHealth's ability to attain intended healthcare objectives can stimulate adoption. Therefore, we posit that:

Effort Expectancy (EE)
Effort Expectancy is "the extent of ease related to the system use" [18]. High ease of use is fundamental to effort expectancy and relates to how easily users can navigate a technological platform. For instance, how easily one can navigate an eHealth website. Difficulties in finding information due to vaguely structured menus and web pages means a waste of time and is one of the major risk factors for eHealth. This suggests that a user-friendly and navigable structures should supersede improved visual appeal. Then again, if the visual appearance decreases significantly, this may also decrease trust in the service [4]. Thus, the assumption here is that the more user-friendly, expedient, and easy-to-use the technology is, the greater the likelihood that consumers will intend to use the service. Therefore, we posit the following:

Social Influence
Social influence is "the extent to which an individual perceives that important others believe he or she should use the new system". It is dependent on an extensive range of provisional stimuli with three mechanisms affecting individual behavior namely; compliance, internalization, and identification. Changing an individual's intention in reaction to societal forces is a function of compliance mechanism while the latter two mechanisms concern altering and amending an individual's beliefs structure and/or initiating the individual's reaction to possible situational benefits. Thus, adoption or resistance to change me be as a result of stimulus from associates. [18]. This study posits that:

Facilitating Conditions
Facilitating Conditions is "the extent of belief an individual has in the existence of an organizational and technical infrastructure to facilitate system use". It combines three key variables derived from preceding theories: firstly, perceived behavioral control which is individual insight into the existence or lack of necessary means and prospects; facilitating conditions defined as impartial environmental features that participants approve enable easy accomplishment of an act; and compatibility defined as the extent to which a novelty is observed to be well-suited to prospective users' prevailing standards, desires, and previous involvements" [18]. The study therefore examines how facilitating conditions affect intention to use. The following is therefore hypothesized:

Trust in internet (ToI)
With reference to trust in internet, the expectation is that there should be minimal or no technical errors [4]. Reliability of technical artifacts is often deemed to be a preferable feature [19]. Thus in this case, technical reliability relates to whether the service will experience a systems fault or other failures. Researchers have identified technical failures as a significant risk for Dutch eHealth adoption, as they may lead to higher time consumption and higher costs [20]. Thus, in this study, we propose that the degree to which an individual has confidence in the benefits which using the technical artifact will provide positively affects their belief in the reliability of the technical artifact. Therefore, we posit the following:

Hypothesis 5: Performance expectancy positively affects Trust of internet
With reference to e-Government, scholars theorize that trusting beliefs in the internet will positively affect behavioral intentions [21]. This paper explores this relationship with reference to eHealth. Thus, we posit the following:

Trust in online healthcare providers (ToH)
Benevolence is defined as the degree to which the object warranting an attitude of trust, which is the trustee, is thought to have good intentions towards the trustor. It concerns the trustors assessment of the trustee's good intentions [22]. Illustrations of this will frequently relate to privacy issues such as providers' intentions when handling patient data for purposes the patient is inadequately aware of. In this case, the trustee is not the eHealth system itself because technology cannot have its own intentions. Therefore, this domain applies to the providers of the eHealth services such as technology developers and the healthcare institution [4]. In this study it is theorized that consumers' positive attitude toward eHealth's ability to attain intended healthcare objectives will positively affect trust in online healthcare service providers. The following is therefore posited:

Hypothesis 7: Performance expectancy will positively affect Trust of online e-health service providers
It is hypothesized that trust in government will positively affect behavioral intentions. However, results of the analysis of this relationship found that this relationship was not significant and therefore not supported. This paper explores this relationship in the context of eHealth and hypothesizes the following:  Performance expectancy will positively affect behavioral intention to use e-health services H2 Effort expectancy will positively affect behavioral intentions to use e-health services H3 Social Influence will positively affect behavioral intentions to use e-health services H4 Facilitating conditions will positively affect behavioral intentions to use e-health services H5 performance expectancy will positively affect Trust of internet H6 Trust of internet will positively affect behavioral intentions to use e-health services H7 Performance expectancy will positively affect Trust of online e-health service providers H8 Trust of online e-health service providers will positively affect behavioral intentions to use e-health service

Instrument development
The research questionnaire contained items primarily adopted from previous studies [21,23], while integrating the necessary authentication and phraseology in order to adapt it to the research context. All questions were rated on a five-point Likert-type scale spanning from "Strongly Agree" to "Strongly Disagree." Questions were clear and simple, intended to find out respondents' general attitude towards eHealth. The questions further sought to capture respondents' prospects of using eHealth services in the near future. Respondents' gender, age, occupation and education information was also collected. The questionnaire was presented to two PhD students who have knowledge in the area. Thereafter, a pretest was carried out with 20 respondents to ensure content validity. The data was then examined for completeness of responses, reviewed and modified in accordance with recommendations. To ascertain questionnaire reliability Cronbach Alpha analysis was conducted.

Sampling
A sample of Malawian citizens who use the internet was identified for this study.
Sampling is a method of choosing an adequate proportion of components out of the population. If the right sample is selected It is promising to apply features of the components to the population elements [24]. Convenience sampling was used because it is quick, inexpensive, and appropriate since researchers simply use participants who are available at that particular moment [25].

Data Collection
The study adopted a cross-sectional approach to data collection in which data was collected at once for 6 weeks. Online survey was the chosen method of data  [29] as is the case with this study. Convergent and discriminant validity to determine construct validity, model testing, and hypothesis testing was done. The subsequent sections provide details of the analysis and results.

Reliability test
A Cronbach Alpha test suggests four cut-off points that represent results ranging from excellent to low [21]. Results of the Cronbach analysis (  [21,30]. Outcomes of AVE indicated scores higher than the recommended 0.5 score [28]. This means the questionnaire had good convergent validity [30]. Results are presented Table 2. Secondly, discriminant validity test was applied by using AVE and it is confirmed when there is low correlation between a measure and another measure which it ought to be different [30]. Thus, Table 3 shows the scores of square roots of AVE in bold diagonal cells higher than correlations between constructs thus indicating a good discriminant validity [28, 30].   Table 4 indicates relatively fair response from both males and females with 50.5% of respondents being female. Over 70% are aged below 35 years supporting observations by other scholars stating that 93% of people between 18 and 29 years are the largest internet and social network users [30]. Results also show that the majority of respondents hold either a Bachelors or Master's degree. 54% of respondents are employed either on a full time or part-time basis while 24.8% indicated belonging to the "Other" category. This could possibly capture students which was a category excluded as an occupation. Nevertheless, this provides useful insights into those who are using the internet the most. Interestingly, over 74.8% of respondents showed over 5 years of internet use and 82.3% use it on a daily basis.

Descriptive statistics
This shows that the majority have experience using internet. respectively. However, scholars argue that adjusted 2 ( 2 /degrees of freedom) is an appropriate metric as 2 alone is not strong for sample size [21].  Table 5 and Table 6 respectively.  H5 indicates greatest significance (β = 0.55, p < 0.001); that means, consumers have a higher expectation of the internet to perform well in order to also be able to trust that eHealth services will accomplish its purpose. Furthermore, H7 indicates a high correlation (β = 0.39, p < 0.001) meaning that consumers' expectations for Online Healthcare Providers to offer reliable and safe healthcare services is significant and this influences their level of trust in using eHealth services. Generally, among all the determinants, PE was a significant predictor of behavioral intentions through consumers' ToI (β = 0.55, p < 0.001) and ToH (β = 0.39, p < 0.001) respectively. If the internet as a means through which one is to access eHealth services is slow, unreliable, and expensive it will be difficult to encourage eHealth use. Likewise, if service providers are dishonest, insensitive and unscrupulous, engaging customers or maintaining those that exist is a difficult task (Bansal, 2016). This implies that PE indirectly influences behavioral intention since it affects ToH positively (β = 0.10, p < 0.05). Effort Expectancy (EE) indicates statistical significance in relation to behavioral intention (β = 0.22, p < 0.001). This implies that citizens will use eHealth services if they believe it to be easy to navigate and explore [21]. This supports sentiments made by other scholars that difficulties in finding information due to vaguely structured menus and web pages means a waste of time and this is not what consumers want [4]. Therefore, userfriendly and navigable structures should be prioritized along with visual appeal [4].  [34,35]. Other studies in e-Government have proven that trust of internet and trust of government have an indirect positive effect on behavioral intentions through performance expectancy [14,36,21]. However, this study explored the indirect effect of PE on behavioral intentions through ToI and ToH. Results may have profound importance to national policy-makers and other relevant stakeholders.
Firstly, PE was highly significant meaning that in order to increase traffic to health websites, there is need to show the usefulness of the systems. Moreover, consumers expect to have a user-friendly system which will facilitate easy use of the eHealth system. This provides an opportunity for government to build reliable platforms for interaction with consumers with reference to health services and promote awareness about these facilities.
Furthermore, ToH has a positive effect on behavioral intentions. It is therefore essential for online healthcare service providers to make sure their platforms are user-friendly and relevant to the needs of consumers. In addition, ToI is not a significant factor for using eHealth services. investigate what health information consumers are particularly interested in on the internet will provide useful insights into adoption behavior.
Thirdly, considering the economic status of the country and consequent economic characteristics of the population, a further look into facilitating conditions through a resource-based view may add valuable information on determinants of adoption.
Finally, the research data was collected from a developing country in which limitations and challenges relating to availability of and access to information on the use and applicability of technology in relation to health is limited as compared to the developed world. A similar study using data from developed countries would provide comparative insights into the generalizability of the results using these constructs.  Trust in internet a  7  I trust eHealth services through the internet  8  I use the internet to search for health information  9 I look at social media sites or professional sites for health-related in 10 I think that the eHealth services' technical and legal infrastructure personal information and data 11

Appendix
In general, internet is a trusted tool I can use to interact with eHea 12 I trust internet security and protection protocols, which increase m eHealth services 13 In general, I don't trust e-health and its services through the intern Trust in online healthcare institutions a 14 I trust online healthcare institutions and departments 15 I trust online healthcare institutions and departments' abilities to p effectively and securely 16 I trust that citizens and their benefits have the highest priorities at institutions and departments Performance expectancy b 17 Using the internet for health-related matters enables me to accom 18 Using the internet improves my success about the subject of the h 19 If I use internet for health-related matters, I will increase my produ Effort expectancy b 20 My interaction with the internet for health matters would be clear a 21 It would be easy for me to become skillful at using the internet for 22 I would find the internet for health related matters easy to use 23 I will recommend others to use health-related websites Social influence b 24 People who influence my behavior think that I should use eHealth s 25 People who are important to me think that I should use eHealth ser

Funding
No funding received for this research

Availability of data and materials
The datasets and materials used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
This research investigated trust as a key determinant of eHealth adoption. Consent to participate in the study was sought from individual participants and further ethical approval was not required. No identifying personal or medical information was obtained or recorded for the purposes of research or any other use.

Consent for publication
Participants involved in this study provided consent for publication. Figure 1 Proposed research model SupTable.docx