1 INTRODUCTION

At the end of 2019, health organizations worldwide faced challenges after the spread of the coronavirus (COVID-19) in regard to providing healthcare services, ensuring social distancing, and limiting people's face-to-face communication at the same time in order to control the virus [1]. On this basis, the State of Kuwait has imposed a complete ban, closed all private clinics and hospitals (since it has been observed that healthy people become infected with the virus while visiting hospitals), and relies only on emergency rooms in public hospitals [2]. In terms of social responsibility, several private hospitals and clinics in Kuwait fully utilized telemedical services as an alternative to traditional medicine during the COVID-19 crisis [2, 3].

Furthermore, Information and Communication Technologies (ICT) have been utilized to enhance the traditional environment of healthcare services with the usage of telemedicine as the most prominent service in the medical field. According to Kamal et al. [4], telemedicine is known as utilizing telecommunication technologies to diagnose, treat, and monitor patients by healthcare physicians and specialists [4]. Telemedicine applications examples may include live video conferences, portals, remote monitoring, and mobile health [5].Telemedical services have proven to be successful in reporting and tracking patient records, delivering, real time monitoring, providing correct medications, and early detection of clinical decline [1]. Moreover, the World Health Organization (WHO) has confirmed that the use of telemedicine services is efficient and cost-effective for health control [1].

The State of Kuwait has a national development plan for all sectors for the year 2035 which is known as "New Kuwait”. One of the health development projects is the telemedical electronic portal lunched by the Kuwaiti ministry of health with the objective of achieving goals related to the New Kuwait development plan [2, 6]. Accordingly, Dar Al Shifa Hospital is one of the first private hospitals in Kuwait to use telemedicine services. In 2017, this hospital began providing telemedical service through an electronic service portal for patients entitled ‘Sehaty online’ [7]. As such, patients registered on ‘Sehaty online’ by calling the hospital help desk after identity validation via email. Registered patients can access the system to contact the healthcare team (physicians and help desk staff), check lab results, request medication refills and renewals, and schedule appointments [7]. Also, patients can consult their doctors to take a suitable online medical treatment and receive the appropriate prescribed medicine.

Despite strong encouragement and interest from governmental authorities and organizations in regard to the continued usage of telemedicine, there exists a generally limited and unclear research in this area, particularly in Kuwait [8, 9]. So far, the focus of a great part of the corresponding literature is on measuring and predicting the “acceptance” or “first-time use” of telemedicine in healthcare sector. However, these studies fail to shed light with respect to its long-term viability and continued usage after COVID-19 [10-12]. Moreover, the intention to continue using any technology in any field in Information Systems (IS), namely e-commerce, e-government, e-learning, e-health etc., at the individual level is essential for the survival of these systems [10]. Furthermore, the importance of continuance is evident from the fact that acquiring new users may cost as much as five times more than retaining existing ones, given the costs of searching for new users, setting up new accounts, and initiating new users in the IS [10]. Therefore, retaining current users and making them continue their usage is indispensable to the achievement of any information system's success for sustainability [8, 13]. In the context of this research, patients’ choice to employ telemedicine after the COVID-19 pandemic would truly reflect the system's value and influence the legitimacy of health institutions’ investments in such a program. Also, the continuity of telemedical service usage can help all stakeholders by reducing the pressure on overburdened hospitals and geography barratries, which can ultimately improve the healthcare sector's social responsibility [8, 9, 13].

In addition, it is known that user satisfaction has an important and sensitive role in users’ intention to continue utilizing any technology in the IS [8, 10, 14, 15]. In a study conducted by Hossain et al. [16], satisfaction was found to be the strongest predictor for driving users’ continuance intention to use [16]. Therefore, it can be concluded that when an individual is satisfied with a technology after the initial trial, he or she may have a higher intention to continue using that technology. This is due to the positive reinforcement of patients’ attitude toward the technology after using it [17]. Thus, it is assumed that higher user satisfaction leads to higher continuance intention to use in the context of this research. Moreover, this situation highlights why decision makers and developers are interested to determine factors that would affect patients’ decisions to continue using telemedicine. Therefore, this study attempts to fill this gap and investigate factors that influence patients’ continuance intention to use telemedicine after COVID-19 in Kuwait's medical sector.

2 LITERATURE REVIEW

There exist a large number of technology adoption theories and models used in the context of telemedicine, such as the Technology Acceptance Models (TAM 1 and TAM 2), the Unified Theory of Acceptance and Use of Technology (UTAUT), the Theory of Planned Behavior, the Post-Acceptance Model (PAM), and many more [10]. However, the updated Information System success model is widely used to assess the success of any system employed in the medical sector [18, 19]. Moreover, it is considered to be the most suitable model in regard to explaining the factors that influence patients’ continuance intention to use telemedicine [8]. This model has an advantage over other models and theories related to explaining user satisfaction considering the fact that user satisfaction is a strong predictor of continuance intention, according to many studies [8, 10]. In addition, it is suggested that information quality, system quality, and service quality determine user satisfaction; also, continuance intention to use is found in several studies to be determined by user satisfaction [8, 10, 16]. The continuance intention to use has rarely been studied with Information System success model [8, 12]. Thus, this study examines factors that influence continuance intention to use telemedical services among Kuwaiti patients after the COVID-19 pandemic.

Furthermore, corresponding literature has revealed that the successful usage of a system is dependent on its continuing utilization [16]. Moreover, continuance intention to use refers to patients' intention to continue using telemedicine [10]. Numerous studies in the field of telemedicine have proven that the success of any system in regard to its sustainability depends on retaining existing users and leading them to repeat using that system [8]. Patients will always want to keep using a particular system that can fill full their needs and help them improve their communication with the healthcare team. Thus, the goal of this paper is to identify factors influencing patients’ continuance intention to use the telemedicine which is fully utilized in several Kuwaiti private hospitals. Identifying these factors can help improve telemedical services provided by private hospitals as well as expand these services in public hospitals. Hence, the next section elaborates more on the proposed research framework and hypothesis for each relationship based on a review of the relevant literature (see Figure 1).

Research conceptual model

2.1 Information quality (IQ)

Information Quality (IQ) is one of the most popularly examined measures in the literature in general as well as in telemedical technology in particular [20]. Moreover, the Information Quality of telemedicine refers to the quality of outputs that are produced by the system and perceived by patients [21], which can be in the form of medical reports or online screens. The relationship between ‘Information Quality’ and ‘User Satisfaction’ has been supported in many studies [17, 21]. In the context in this study, Zhou et al. [20] have found that information quality is significantly and positively related to patients’ satisfaction [20]. Furthermore, they argued that the accessibility of medical records directly affects the level of patients’ satisfaction with telehealth services. Additionally, authors concluded that patients’ effective usage of the information displayed on telehealth services can led to a positive impact on their satisfaction and the acceptance of telehealth services [20]. Moreover, according to the study of Keikhosrokiani et al. [22], after patients easily accessed their medical information in the mobile healthcare system (iHeart), they showed an extremely high level of satisfaction toward the practical application of this system [22]. Thus, it is assumed in this research that higher information quality of telemedicine leads to higher patient satisfaction.

H1: Information quality has a significant and positive effect on telemedicine users’ satisfaction.

2.2 System quality (SQ)

From the perspective of this paper, System Quality (SQ) is described as a system that has specific characteristics desired by patients and telemedicine users such as usability, responsiveness, flexibility, reliability, and ease of use. In a study related to m-health, Oppong et al., [23] found that system quality has no significant effect on user satisfaction [23]. However, numerous recent studies have reached contradictory results. As an example, Lin, H. C. [24] discovered that System Quality has a strong influence on system users’ satisfaction in regard to e-health. As such, when users observe that the system features are more fitted to their needs in terms of responsiveness, flexibility, reliability, and ease of use, they gain more satisfaction [24]. Similarly, the empirical results of Kuo et al., [25] reveal that system quality is one of the most important system attributes in a virtual hospital environment, impacting the satisfaction of users. Moreover, the authors argued that systems utilized for critical tasks such as diagnosing monitoring patients should be reliable and have sufficient quality characteristics [25]. Thus, it can be hypothesized that:

H2: System Quality has a significant and positive effect on telemedicine users’ satisfaction.

2.3 Service quality (SerQ)

Service Quality (SerQ) has been defined in the literature in different ways based on what the service represents for the user. In the e-health context, service quality is defined as the patient's perception of the quality of service delivered by the healthcare team using electronic tools [21]. In this study, the healthcare team includes medical consultation and other supporting services, namely contacting the help desk for system registration, identity validation via email, and scheduling appointments.

In a number of prior studies, it was found that service quality has no significant effect on user satisfaction. In this regard, Kaium et al., [8] investigated factors affecting user satisfaction and the continuance usage intention of mHealth in developing countries. In their study, service quality represents the user evaluation of the provided services and support, which was concluded to be insignificant [8]. The authors explained that this result was due to the lack of adequate support, professionalization, and diversification of the provided services. On the other hand, many previous researches revealed that service quality's yielding satisfactory human relations and interaction between healthcare personnel and system users can enhance the assurance of the provided services and, therefore, increase user satisfaction [23, 26]. As a result, it is logical to hypothesize that:

H3: Service Quality has a significant and positive effect on telemedicine users’ satisfaction.

2.4 Patient satisfaction (P Sat.)

Patient Satisfaction (P Sat.) refers to the extent to which a patient is pleased or contented with telemedicine after having gained direct experience with this technology [16]. Also, it refers to the feeling of pleasure or displeasure that results from aggregating all the benefits that a patient hopes to receive from interaction with telemedicine services [17]. Many studies provide insights on the importance of user satisfaction on the continuance intention to use healthcare applications [14, 15]. In this regard, Zhou et al. [20] discovered that patients' satisfaction with medical services mainly includes comfort, professionalism, privacy, and waiting time. Moreover, authors demonstrated that patient satisfaction could significantly predict patient's continuance intention to use telehealth services [20]. For instance, Kaium et al. [8] concluded in a study related to mHealth in developing countries that patients become satisfied with using mHealth services regularly only when it fulfills their medical needs, which can lead them to have a constant intention to use this technology. Furthermore, Bhattacherjee, A., [10] found that the positive experience of patients and their satisfaction with telemedical services can be due to fast access to their medical information and good helping lines. Thus, when the expectations of the telemedicine are met by patients, their satisfaction grows, and this can give patients the intention to continue using telemedical services. Therefore, the following hypothesis can be formulated:

H4: Telemedicine users’ satisfaction has a significant and positive effect on patients’ continuance intention to use.

3 METHODOLOGY 3.1 Data collection

Primary data was collected from patients who used ‘Sehaty online’ telemedicine of Dar Al Shifa Hospital in Kuwait. For this purpose, patients were contacted through emails and social media and their permission to obtain data was acquired. As such, from 407 distributed questionnaires, 290 valid responses were collected, resulting in a response rate of 41 percent which is more satisfying than the response rate of 29.2 percent in previous studies [27]. The response rate of this study also follows the sufficient statistical sample of 120, as calculated by G-power.

3.2 Research instruments

The development of instruments was carefully made in order to reflect the nature of the study. Hence, the questionnaire was created and included 19 items for this study. Moreover, the variables were measured using the five-point Likert Scale, with five being ‘Strongly Agree’ and one being ‘Strongly Disagree’. Also, because the respondents were Arabic speakers, it was vital for the questionnaire to be precisely translated from English to Arabic. Therefore, a back translation was performed, which is a procedure extensively applied to test the precision of the translation in a cross-cultural survey. The validated instruments shown in Appendix C were adapted from related previous studies to measure the variables of this research.

4 FINDINGS AND DISCUSSION 4.1 Respondents’ profile

In the demographic information section, respondents were categorized by their age, gender, marital status, and level of education, as displayed in Table 1.

TABLE 1. Respondents’ profile Item Frequency Percentage Age 20–30 82 28.2 31–40 104 35.8 41–50 82 28.2 51–55 22 7.5 Gender Male 172 59.3 Female 118 40.7 Marital status Single 117 40.3 Married 158 54.4 Divorced 15 5.1 Education High school 12 4.1 Degree 143 49.3 Master 73 25.1 PhD 62 21.3 4.2 Measurement model

The research model of this study was tested using SmartPLS 3.3. In addition, an examination was conducted in regard to the measurement model (validity and reliability of the measures) and the structural model (testing the hypothesized relationships). As a result, all variables scored satisfactory values of Cronbach's Alpha, Composite Reliability, and Average Variance Extracted (AVE) and were above the cutoff points of 0.7, 0.7, and 0.5, respectively, as recommended by Hair et al. [28]. However, SerQ5 scored −0.060 for factor loading, which is below 0.4, as recommended by Thurasamy et al. [29]. Table 2 illustrates the convergent validity scores.

TABLE 2. Convergent validity Variables Items Factor loading Cronbach's alpha Composite reliability Average variance extracted (AVE) Information Quality IQ1 .900 .943 .959 .855 IQ2 .930 IQ3 .929 IQ4 .939 Patient continuance intention to use telemedicine PCIU1 .883 .839 .904 .759 PCIU2 .925 PCIU3 .800 Patient satisfaction Psat1 .874 .898 .937 .831 Psat2 .937 Psat3 .924 Service quality SerQ1 .912 .909 .932 .775 SerQ2 .860 SerQ3 .844 SerQ4 .902 System quality SQ1 .802 .877 .916 .732 SQ2 .893 SQ3 .895 SQ4 .828

Second, the discriminant validity was examined in order to assess how truly distinct a construct is from other constructs. In the area of distinguishing validity, the correlations between variables in the estimation of the model did not exceed 0.95, as suggested by Kline [30] and the validity was tested based on measurements of the correlations between constructs and the square root of the average variance derived for a construct [31, 30]. Hence, Table 3 contains the results of the Fornell and Larcker Criterion and shows no value above the recommended cutoff point of 0.95 [31]

TABLE 3. Fornell and Larker criterion IQ PCIU Psat SQ SerQ IQ .925 PCIU .343 .871 Psat .583 .398 .912 SQ .686 .380 .555 .856 SerQ −.098 −.044 −.105 −.056 .880

Moreover, the Heterotrait-Monotrait ratio (HTMT) is an estimate of what the true correlation between two constructs would be if they were perfectly measured (i.e. if they were perfectly reliable). Furthermore, HTMT is the mean of all correlations of indicators across constructs measuring different constructs (i.e. the Heterotrait-Monotrait correlations) relative to the (geometric) mean of the average correlations of indicators measuring the same construct (i.e. the Heterotrait-Monotrait correlations). Moreover, it can be used for discriminant validity assessment [28]. As such, the accepted level of HTMT is 0.90, as recommended by Gold et al. [32] (see Table 4).

TABLE 4. HTMT ratio IQ PCIU Psat SQ SerQ IQ PCIU .385 Psat .630 .454 SQ .752 .438 .623 SerQ .098 .054 .102 .065 4.3 Structural model

The structural model represents the theoretical or conceptual element of the path model. Also referred to as the inner model in PLS-SEM, the structural model includes latent variables and their path relationships [28]. The next step after the evaluation of the measurement model is to assess the structural model. In sync with PLS-SEM, there are five steps required to assess the structural model [28], including the assessment of collinearity (step one), assessment of the path coefficients (step two), coefficient of determination (R2 value) (step three), blindfolding and predictive relevance Q2 (step four), and effect size f2 (step five).

Furthermore, Table 5 illustrates the results of PLS bootstrapping consisting of the Beta value, t-values, p-values, hypothesis results (whether supported or not) BCILL, BCIUL, f2, and VIF scores. Additionally, Appendix A summarizes the results of the structural model and PLS bootstrapping.

TABLE 5. Summary of the structural model (PLS bootstrapping) H Path Std. Beta Std. Error T-value P values Decision BCILL BCIUL f2 Effect size VIF H1 IQ → Psat .377 .067 5.612 P < .001 (.000) Supported .270 .487 .122 Small 1.902 H2 SQ → Psat .295 .067 4.397