Technostress and medical students’ intention to use online learning during the COVID-19 pandemic in Pakistan: The moderating effect of computer self-efficacy

Abstract This study’s aim is to determine whether or not medical students in Pakistan are willing to use online learning during the COVID-19 pandemic, as well as their perception of and satisfaction with online learning. The study also analyzed the relationship between technology-related stress and computer self-efficacy among medical students. This cross-sectional study selected 369 newly enrolled medical students as samples using convenience sampling. Medical students exhibited a significant negative association with technology-stress and the intention to use online learning. Further, computer self-efficacy decreased the effects of technology-stress and enhanced the intention to use online learning among medical students. Thematic analysis revealed three main themes: (1) engagement in studies, (2) management of time, and (3) challenges faced by students. Although students experienced challenges in terms of methodological, technological, and behavioral concerns during online classes, online classes also saved time and enhanced performance as a result of better time management.


Introduction
The terms "global pandemic" and "global village" reflect the absence of geographical limits, which has caused the COVID-19 to spread and take a disproportionately high death toll in wealthy Western countries. Africa, along with many areas in South Asia and the Far East, appear to have escaped the worst of the storm (Hashim, 2020). Developing countries lack the financial resources necessary to weather a pandemic, whereas the developed countries benefits from technological advancements. The academic mindset of those enrolling in educational institutions in Pakistan, along with the hasty conversion by the Higher Education Commission (HEC) and universities to online academic learning that shifted away from the traditional instructional approach, has frequently been cited as a cause for concern. According to a recent Pakistan Telecommunication Association (PTA) estimate, only 36.86% of the Pakistani population has access to broadband internet (PTA, 2019). To its credit, the HEC has maintained a realistic stance throughout the crisis, offering technical assistance to universities and making a series of online education policy recommendations (Banuri, 2020). Mandatory online instruction is not the same as an unprompted act during a pandemic, and it has increased the technology-related stress among educators (Mushtaque, Waqaset al., 2021).
The impact of information technology on human life is immense, and its importance in education cannot be overstated. Due to the closure of educational institutions, which poses impediments to student learning under the current COVID-19 pandemic scenario, the usage of information technology has increased. During this quarantine period, information technology provides a solution for continuous learning through creative and learning management systems (Ahshan, 2021;Ferri et al., 2020). Many people have benefited from cutting-edge technology; yet improper or excessive use of technology can have a negative impact that worsens technology-related stress (Dong et al., 2019). Indeed, technology has become an integral part of most people's daily lives. During the COVID-19 pandemic, colleges expand their use of social media to provide information about classes and how to access virtual learning tools (Sun et al., 2020a). However, as a result of the overload that extended use of social media can generate when students are spending more time on them, there might be interference to their academic obligations. This can lead to "technology-related stress" or "technostress," which has been linked to decreased sleep quality and academic performance (Hsiao et al., 2017;Qi, 2019). Thus, students who want to learn online must have computer self-efficacy and information technology abilities.
Self-efficacy is a crucial component of e-learning (Roca & Gagné, 2008). The influence of selfefficacy on students' involvement, well-being, attitude toward school, and academic success has been well-established. Self-efficacy is regarded as a correct evaluation of one's own ability and a drive to develop one's talents (Scoda, 2019). In other words, self-efficacy will determine whether or not people engage in tasks, how much effort they put in, and how persistent they are in completing them (Bandura & Watts, 1999). The review discovered that student performance and happiness with online learning were predicted by internet self-efficacy, although the results were mixed. Other studies examined the correlation between student satisfaction and their level of internet self-efficacy, and it was found that internet self-efficacy could not predict student contentment (Alqurashi, 2016;Kuo et al., 2014). In developing countries, online education is still in its infancy. A survey of 702 university undergraduates from various departments was carried out by Hamdan et al. (2021a). During the COVID-19 outbreak, online interaction, self-efficacy, self-regulation, and contentment of Jordanian university students were investigated. According to the research, students had the lowest mean scores in internet self-efficacy and satisfaction with online classes, respectively. Students' postacceptance conduct as a crucial educational cognitive decision is influenced by their intentions (Kim, 2010).  found that students in arts and humanities programs in a Pakistani university were excited about using the internet. The students also said they wanted to use it in the future, or when the pandemic is over. However, obesity, visual problems (Noor et al., 2020), sleep deprivation, anxiety, aggression, lack of interest in education (Badasyan & Silva, 2018), and behavioral problems in students can all be attributed to excessive time spent on technology (Gao et al., 2020), while ignoring technology-related stress in students. Students' self-efficacy may have suffered as a result of COVID-19ʹs rapid transition to online and electronic learning, and the effects may have been mixed. As a result of this unanticipated development, students in Pakistan, a developing country, are questioning their capacity for online academic success. Academic success and learning require a high level of self-efficacy (Zhao et al., 2022). Understanding the barriers that prevent students and educational organizations from continuing to use online learning is thus crucial in ensuring that students and educational organizations can survive in a highly competitive educational environment, especially in light of the global health crisis.

Purpose and objective
The study used the mixed method design comprising qualitative and quantitative methods. We employed the explorative study to analyze medical students' perceptions and levels of satisfaction concerning the use of online learning method during the COVID-19 pandemic in Pakistan. The second objective of the study is to determine the association between technology-related stress, computer self-efficacy, and the intention to use online learning among medical students during the COVID-19 pandemic. The following hypotheses were conceived: H1: Technostress has a negative association with the intention to use online learning among medical students.
H2: Computer self-efficacy has a positive moderating effect on the relationship between technostress and the intention to use online learning among medical students.

Design
Since data were obtained at a single point in time, the cross-sectional approach was the best fit for this inquiry. It is regarded as a valuable and cost-effective method of determining relationships between variables (Polit, 2021).

Sampling and data collection
The study enlisted 369 medical students from several medical colleges in Pakistan by using the convenience sampling design. The G power software was used to calculate the sample size. The F-test was run with the alpha level set at .05, effect size at .25, and power at 0.80. It was established that a sample size of at least 305 students was required. Only full-time medical college students who enrolled in at least one online course at the time of data collection, could read and comprehend English, and agreed to participate were included in the study. Participants in the survey were given a participant information sheet and the contact information of the lead researcher. Participants were reminded that their participation was completely voluntary and anonymous. The research was carried out in a number of Pakistani medical colleges. There are 114 medical colleges in Pakistan, of which 44 are governmental and 70 are private medical colleges. During the COVID-19 pandemic, all medical universities that offer online courses for medical students were closed. This study's data were obtained from three public medical colleges: Nishtar Medical College Multan, Services Medical College Lahore, and King Edward Medical College Lahore, as well as one private medical college, Multan Medical and Dental College. A web-based (online) self-administered questionnaire was used to collect the data. The Google Forms tool was used to build the questionnaire. First, a cover letter and participant information sheet were sent out to describe the goal of the study and what participation entails. Then, a questionnaire (covering technostress, computer self-efficacy, and willingness to use online learning) was used to collect participant responses.
Consequently, an open-ended question was asked to the students about their perception and satisfaction regarding the implemented online learning. Students were provided with a link to the survey for them to access and complete it. The survey was estimated to require between 7 and 10 minutes to complete. Data were collected between February 1 and 30 April 2021.

Research questionnaire
The first section of the questionnaire contained general information about the study (such as the title and purpose of the study, who is invited to participate, research ethics, and so on). The second section of the questionnaire requested specific demographic information about the participants, including their age, gender, and degree of education. The third section covered the main body of the research. The questionnaire assessed students' perceptions of technology-related stress, computer selfefficacy, and their intention to continue using online learning as a medium of instruction.

Technostress
The construct consisted of nine items, which were scored on a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree .

Computer self-efficacy
The construct consisted of 10 items, which were scored on a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree (Akbulut, 2009;Chuang et al., 2018).

Intention to use online learning
The construct consisted of four items, which were scored on a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree (Lee, 2010;Wu & Chen, 2017).

Data analysis
Statistical Package for the Social Sciences (SPSS) version 25 was used to analyze the data. To determine the demographic profile of the respondent, three types of interaction, technologyrelated stress, computer self-efficacy, and student intention to use online learning, this study calculated the descriptive statistics such as mean, percentage, and standard deviation and PLS SEM analysis. All hypotheses were evaluated twice with a p-value = 05 and 95% confidence intervals.
Thematic analysis was used for the open-ended questions (Kiger & Varpio, 2020). This approach helped reduce data into important elements in a systematic way. Data were organized around the discussion questions. Inductive analysis was done to code each significant sentence and group related codes into overarching sub-themes. Finally, related sub-themes were grouped into a major subject. The initial coding sets were then compared, discussed, and adjusted until all authors agreed on a more representational coding scheme, sub-themes, and themes. Table 1 displays the background information of the respondents. Male respondents outnumbered female respondents by a wide margin (62.6% vs 37.4 %). The average age of the students was 22 years. Approximately 63% of the respondents were first-year medical students, and 36% were in their second year of study. Regarding the time spent on the internet, most of the respondents spent 1-5 hours each day (65.0%), while some spent 9-10 hours (13%) using the internet. The majority of the respondents came from cities (61%).

Measurement model assessment
The present study employed the PLS-SEM technique for data analysis. It consists of two models: measurement model and structural model. The measurement model, also known as the outer model, was utilized to evaluate individual items' reliability, internal consistency reliability, convergent validity, and discriminant validity. The structural model or referred to as the inner model was used to calculate the significance of the path coefficients. The measurement model (Figure 1.1) and its components are discussed below.

Individual item reliability
Individual item reliability was checked by calculating the factor loadings of every construct (Duarte & Raposo, 2010;Hair et al., 2016;Hulland, 1999). Items with loading between .40 and .70 can be retained in the model (Hair et al., 2016). Items should be deleted if the deletion would increase the  values of composite reliability (CR) and average variance extracted (AVE; Hair et al., 2016). In the current study, items TS8, TS9, SE4, SE5, and SE7 were deleted to increase the values of CR and AVE. All other items, whose values are between .40 and .70, were retained in the study.

Internal consistency reliability
The extent to which every item of a construct measures the same construct is known as internal consistency reliability (J. Hair et al., 2014). It can be determined using Cronbach's alpha and CR (Hair et al., 2017). Cronbach's alpha and CR are quite similar, but CR is widely recognized in modern research (Barroso et al., 2010). Thus, the current study adopted CR to measure internal consistency reliability. CR values below .60 are not acceptable, CR values between .60 and .70 are considered average, and CR values between .70 and 0.90 represent adequate internal consistency reliability (Hair et al., 2011). Table 2 shows the internal consistency reliability of the present study, where all the values fall within an acceptable range.

Convergent validity
Convergent validity was determined by average variance extracted (AVE) in this study (Hair et al., 2010). The AVE of every construct must be greater than .50 for adequate convergent validity (Fernandes, 2012;J. Hair et al., 2014). Table 2 exhibits that each construct's AVE is above .50.

Discriminant validity
The degree to which a construct is unique and different from other constructs is called discriminant validity (Hair et al., 2017). In this study, three methods were used to measure discriminant validity, namely Fornell-Larcker criterion, cross-loading, and heterotrait-monotrait ratio of correlations (HTMT). The Fornell-Larcker criterion (Fornell & Larcker, 1981) is adapted to measure discriminant validity by using the values of AVE. For cross-loading, the outer loading of an indicator with its construct must be greater than its cross-loadings with other constructs (Grégoire & Fisher, 2006). HTMT is a factor correlation that distinguishes between two factors (Henseler et al., 2016).

Structural model assessment
The structural model was used to determine the associations among the constructs. Bootstrapping method was applied to test the hypotheses. In total, 369 samples were incorporated to measure the significance of the path coefficients. Figure 2 shows the structural model of the study.
The structural model represents the path coefficient results of the study. Hypothesis H 1 states that technostress has a negative association with the intention to use online learning among medical students. Figure 2 and Table 6 demonstrate the existence of a significant and negative relationship between technostress and the intention to use online learning among medical   students (β = −0.092; t = 1.977; p < 0.049). Further, the results for H 2 show that computer selfefficacy had a positive moderating effect on the relationship between technostress and the intention to use online learning among medical students. As shown in Figure 2 and Table 6, the interaction effect of technostress*computer self-efficacy on medical students' intention to use online learning (β = 0.103; t = 2.575; p < 0.010) was significant.

Discussion
Due to the COVID-19 pandemic, the majority of medical schools hastened to implement e-learning regardless of their preparation. Various unique educational technologies are being developed and integrated into instructional design procedures in order to engage students in an interactive learning environment. This study examined medical students' online learning experiences in terms of computer self-efficacy, technostress, and intention to use online modes of learning during the COVID-19 pandemic in Pakistan. Learning instructions shifted dramatically from classroombased to entirely online learning in a relatively short time as a result of the rapid spread of the pandemic. For the first time, students and educators were completely immersed in online education . The current study included 369 students in their first or second year of medical school from several medical institutes in Pakistan. The result shows that technologystress was negatively associated with the intention to use online learning among medical students (Table 5). This negative correlation means that increased technostress was associated with lower levels of contentment and pleasure among medical students. Additionally, students expressed frustration and a lack of contact with the material and lecturers. Initially, online education was heavily dependent on teachers, who played a key role in the learning process. The fact is that internet education is not well-established in developing countries such as Pakistan (Hamdan et al., 2021b). According to the information gathered from the respondents, 63% of the students were newly enrolled. This study supports the findings of Booker, Rebman & Kitchens (2014) and Alam (2016), who established a link between technostress and student performance among aviation professionals participating in an online program.  In Table 1, the respondents reported that 38% of them were from Pakistan's rural areas, which might have an impact on their inclination to use online learning modes. Infrastructure concerns, inadequate technical help, and a shortage of expertise in the technology of digital media platforms such as Zoom and others might all be factors affecting students' intention to use online learning. When the network is down for an extended period of time, such as when there are a large number of people using the internet at the same time or there is power outage or a technical issue with a digital platform, both instructors and students are affected (Mokh et al., 2021). With today's more sophisticated technologies and virtual learning environments, technological proficiency is more vital than ever for students. According to the findings of this study, there was a negative link between technological stress and the intention of medical students to use online learning. Furthermore, problem-focused coping had a direct protective effect on stress symptoms. By providing students with access to ICT services, training, and workshops, as well as simple online ICT instructions and resources, their technological comfort and problem-solving skills can be improved. Medical students at the undergraduate level today have unprecedented access to educational resources. There are currently both traditional and online (or e-learning) methods available for this form of training, including textbooks, lectures, and tutorials. The phrase "blended learning," devised to characterize this approach, is now widely used (Ball et al., 2015). The topic of computer self-efficacy has garnered much attention (Compeau & Higgins, 1995). In the current study, computer self-efficacy was examined as a moderating variable. The result revealed that the interaction between computer self-efficacy and technostress had a positive association with medical students' intention to use online learning modes (Table 5). According to Bandura's (1997) theory, self-efficacy has four primary consequences. First, self-efficacy has a direct effect on the circumstances and activities that affect how an individual makes decisions. Second, when confronted with adversity, an individual's self-efficacy has an effect on the level of effort they expend to overcome obstacles and endure. Third, an individual's perception of self-efficacy has an effect on their stress and anxiety levels. Finally, self-efficacy is a component that impacts performance and coping abilities. Computer self-efficacy refers to the ability to utilize a computer effectively (Ebijuwa & Mabawonku, 2019). Self-efficacy is a trait that differs from person to person. Someone with high self-efficacy views pressure as a task to be overcome, whereas someone with low self-efficacy views work as stressful. Individuals with low computer self-efficacy, according to current research, face more technical stress . Computer effectiveness effects attendance, career interests, and academic success (Okon & Dijeh, 2022;Rhew et al., 2018). Computer self-efficacy influences people's assessments of their computer skills (Zelalem et al., 2022). Students with high levels of computer self-efficacy are more likely to experience a drop in perceived technostress than students with low levels of computer self-efficacy. By examining the moderating effect of computer self-efficacy on technostress and the intention to use online learning modes among newly enrolled medical students, we can conclude that increasing computer self-efficacy significantly reduces the technostress caused by technology complexity and increases the intention to use the online medium (Table 5). Based on the findings, students who are more confident in their abilities to utilize computers have lower levels of technostress, whereas students who rely on technology have higher levels of technostress. Individuals may report different levels of technostress in different circumstances (Shu et al., 2011;Toto & Limone, 2021). Students must have trustworthy equipment and be familiar with the technology used in the course in order to succeed in an e-learning course. Students who demonstrated a high level of self-efficacy with computers were more confident in their online education and achieved higher grades (Shen, 2017). According to (Mustafa, 2019), computer self-efficacy is critical for user retention. Even if someone has exceptional computer skills, inefficient computer use may cause them to avoid using them. General computing had stronger self-efficacy than sophisticated computing. According to (Brashi, 2022), only general computer experience has an effect on general computer self-efficacy. This could imply that university students have more general computing abilities rather than specialized ones. The computer self-efficacy scale was found to be a significant moderator of students' intentions to use online learning in this study.
The current study also investigates medical students' perception and satisfaction with online classes (Table 7). While this study's themes support the idea that "online learning works for medical students," it does not mean that online modes can totally replace in-person live sessions. Our study participants faced many issues in adopting online learning. One of the most common obstacles was a lack of basic computer skills and insufficient internet connectivity. Our findings confirm many prior findings (Dyrbye et al., 2009;Khalil et al., 2020;Niebuhr et al., 2018). The level of technology usage in medical education depends on a faculty's preparation and expertise. Traditional education methods must be modified to teach these abilities to physicians. This is in view that training in educational technology mastery seems to be a neglected ability in faculties that should be obligatory for medical colleges (Goldberg, 2014).
Personal variables such as learning style, acceptance of novel learning techniques, and level of involvement in online classrooms influenced the experiences of our study participants. Meanwhile, institutions' use of online learning is fraught with quality assurance issues (Omar et al., 2012;Terrell & Dringus, 2000). The Bediang et al., claimed that increasing collaboration among all departments and stakeholders is crucial to the adoption of online programs. When adopting an online learning module, faculty members must apply standardized approaches in order to produce a well-regulated and efficient system (Bediang et al., 2013). Instructors' lack of non-verbal communication was also noted as a major issue by our study participants. According to communication theorists, verbal messages are conveyed via words, whereas nonverbal messages are conveyed via nonverbal cues. A student's psychological closeness to their instructor is based on the instructor's nonverbal cues. Nonverbal communication includes eye contact, gestures, and posture (Manusov, 2016).

Conclusion
This article makes a number of noteworthy contributions. This study provides empirical evidence that computer self-efficacy is adversely associated with technostress and increases the desire for online modes of learning among medical students. Our research also found that medical students preferred online learning. Technological issues, individual behavioral peculiarities, institutional technique restrictions, and the lack of nonverbal cues all posed difficulties for our study participants. Apart from technological issues, personal variables such as learning style, acceptance of novel modes of instruction, and level of engagement in online classrooms also influenced their experiences.

Implication of the study
The findings of our study can help educators and policymakers build and implement an online medical learning system. Students value the adaptability, accessibility, engagement, information obtained, and convenience of online learning. An online learning system must have these properties in order to boost student confidence and academic accomplishment. These elements should be considered while creating and implementing an online learning system. Medical schools should locate their students in community health centers and hospitals near where they live, allowing them to mix theory and practice. Medical students learning online face hurdles such as a lack of enthusiasm, comprehending difficulties, a lack of emphasis on practical orientation, and a lack of computer abilities. The main barriers to online education in Pakistan are poor connectivity and a lack of technical support at home. Online learning modules, educators say, must be convenient, accessible, and participatory. Academics, according to administrators, should be encouraged and compensated for developing online teaching resources that benefit students. They should prepare educators and students to use online learning tools in these challenging circumstances.
Our research is distinctive due to its emphasis on medical education technology. This study lays the framework for building and implementing an online learning system that emphasizes both theory and practice to meet the needs of medical students. Given the importance of internal motivation in adopting an online learning environment, the synchronous teaching mode can be supplemented with an asynchronous mode, such as providing students with access to technologically-based, self-paced learning resources, such as surgical videos, telehealth, and telemedicine, as well as online practice questions, to facilitate the attainment of learning outcomes. Medical schools should provide students with unique learning opportunities, such as volunteer work, online electives, and educational resources for patients.