The analysis of the effect of satisfaction on telehealth utilization with a structural equation model

2.1 Literature and research hypotheses

Telehealth has been used limitedly for decades, primarily to make healthcare available to people who live in rural or isolated communities and people with limited time, mobility, and transportation options [17]. Widespread use of Telehealth was often connected to factors such as cost, technology, and ambiguous reimbursement regulations [18, 19]. However, during the pandemic, Telehealth helped providers and patients avoid unnecessary contact and minimized the risk of COVID-19 transmission. Thus, the need for telehealth services has driven up the use of Telehealth, and people use it because they need it.

According to the previous literature, several factors influence the use of Telehealth, including areas like satisfaction, comfort, ease of use, and convenience [20–24]. It has been identified that satisfaction of Telehealth can powerfully increase telehealth utilization [20–24]. How comfortable, easy, and convenient it is for patients to use Telehealth services and express their concerns are also important factors influencing satisfaction and utilization. Studies have found that several factors could impact the satisfaction with telehealth services; patients have different barriers (language, knowledge, etc.) that make it difficult to express their concerns or lack of connection or relationship with providers [20, 21]. Other studies found that patients expressed dissatisfaction with telehealth services due to the lack of “hands-on” care, making them feel like their needs were not met [22]. Meanwhile, looking at non-English speaking and elderly patients, researchers found that they were dissatisfied with telehealth services and preferred in-person visits due to feelings of being misunderstood or generally not feeling confident in the quality of the virtual visit [23]. Further, poor user accessibility, such as technological limitations, may outweigh the need and affect overall satisfaction [24].

With telehealth use trending more after the COVID-19 pandemic, it seems more people are acquiring resources and preferring the convenience of telehealth services [18, 26]. For example, the American Psychiatric Association found that nearly four in ten Americans, or approximately 38%, have been using telehealth services –which increased by 7% from fall 2020 or the pandemic’s beginning [25]. In addition, patients report the convenience of Telehealth contributes to their satisfaction due to not having to be absent from work and/or travel to these appointments –saving individuals gas, time, parking, physical limitations, etc., particularly for individuals in rural communities or cities [18, 27]. Studying 1734 patients, Polinski et al. found that 95% of patients appreciated Telehealth because of its convenience [28]. Similarly, using a study of 84 participants, Ruelos et al. pointed out that 86.7% of patients were using Telehealth in place of in-person visits due to convenience, again a primary driver of telehealth satisfaction and utilization [26]. Therefore, access to Telehealth is crucial for telehealth utilization; increasing access by making care more convenient may increase telehealth utilization and health care spending [29–31].

Again, using Telehealth requires technological and language skills, as language could preclude some individuals from using the service. A study conducted in California found that patients with language barriers (i.e., those patients with limited English proficiency) had lower rates of telehealth use than proficient English speakers (4.8% versus 12.3%) [32]. In a retrospective cohort study, investigators found fewer rheumatology appointments were kept during COVID-19 and associated the preference for the Spanish language or other non-English languages with the decline in video use [33].

The technology and the user’s technical skill level are critical factors influencing telehealth utilization. While Telehealth ensured patients continued HIV care services in South Carolina during the COVID-19 pandemic, technological challenges, digital literacy, client/provider experiences, and factors were barriers to using Telehealth [34]. Similarly, while Telehealth improves health outcomes and quality of life indicators for older adults, using the technology was often an obstacle [35]. High rates of non-participation in Telehealth have been noted in aging populations. The population refused acceptance because they believed there were requirements for technical competence and operation of equipment [36].

Investigators have been interested in determining the link between telehealth usage and knowledge of Telehealth. One study found that interest in telehealth utilization was higher when patients had more confidence in using the associated technology and perceived more advantages than disadvantages when using Telehealth [37]. While different forms of Telehealth were more prevalent before 2020, most respondents were willing to use videoconferencing visits [38]. However, there was a caveat to the findings; participants who identified as Black, were 65 years or older, or had less education (high school or less) were less likely to be willing to use videoconferencing [38]. Even when the benefits of Telehealth are known, some indicated a preference for in-person visits. For example, families of kids with disabilities who use early intervention therapy prefer in-person visits, feeling that they have to compromise the relationship between family, child, and therapist and that treatment would not be as effective if done through Telehealth [39].

The knowledge of Telehealth among providers and the clinicians also has been shown to impact the utilization of Telehealth. While most physicians indicated that they would be interested in using the technology, some also expressed they had little understanding of Telehealth and perceived the advantages of Telehealth at a moderate level [40]. Non- or underutilization of Telehealth exists among clinicians as well. For example, while clinical video telehealth (CVT) enables care to be provided directly to veterans’ home, it remains underutilized despite training initiatives [41]. While the majority of the mental health providers endorsed the use of CVT, less than half used it for patient care because of a lack of training or administrative burden. Training with a multimodal approach to nurse practitioners found that they were more likely to embrace telehealth usage [42]. Thus, increased training would likely increase telehealth utilization overall.

Meanwhile, knowledge sharing significantly and positively affected organizational performance when organizations were going virtually during the COVID-19 pandemic [43]. As the nature of healthcare is, patients must rely on physicians to make healthcare decisions, such is the so-called physician-induced demand. Thus, the telehealth providers’ utilization is connected patients’ utilization.

Through the exiting literature, some of the previous literature suggest that access, need, technology knowledge, and skills influence telehealth utilization while some indicate that those factors could influence telehealth satisfaction; however, few indicate that those factors could influence both utilization and satisfaction. To date, no study examines that access, need, technology skill, and knowledge could directly influence both utilization and satisfaction with telehealth; meanwhile, those factors also indirectly influence utilization through satisfaction with telehealth services. Therefore, this research fills this gap by employing the structural equation modeling (SEM) approach to capture the direct and indirect effects on telehealth utilization.

Therefore, the hypotheses in the following research are:

As the access to telehealth services increases, telehealth utilization increases (H1d), and telehealth satisfaction may also increase (H1i), which in turn increases telehealth utilization (H5).

2.2 Self-determination theory

Self-determination theory (SDT) is a theory of human motivation and personality that suggests people can become self-determined when their needs for competence, relatedness, and autonomy are fulfilled [44]. Self-determination refers to a person’s ability to manage themselves, make confident choices, and think independently [45]. Deci and Ryan first introduced self-determination theory in their book Self-Determination and Intrinsic Motivation in Human Behavior [14]. Self-determination significantly impacts motivation; for example, if the individual believes he/she can manage himself/herself correctly, he/she would have more motivation to carry out some tasks.

Self-determination theory posits that three innate and universal psychological needs drive people and personal well-being is a direct function of satisfying these basic psychological needs [46]. First, autonomy, the autonomous motivation is crucial as the individuals feel that they have the choice and willingly endorse their behavior. Second, relatedness, the need for growth as a human being, drives behavior as people always actively seek to grow and improve [44]. And third, competence, mastery over challenges, knowledge, and skill, are essential for developing a sense of self. As a result, the quality of motivation is key to both satisfaction (telehealth satisfaction) and sustained success in achieving one’s goals (telehealth utilization). The more pleasure and satisfaction the individuals receive from telehealth services, the more they would use these service options.

Self-determination theory has been widely used in health behavior, physical, and psychology studies regarding the effects on motivation [47–49]. For example, during the COVID-19 crisis, researchers investigated emotion regulation and psychological needs based on Self-determination theory to examine the individuals’ mental health and students’ motivation to engage in online learning [50, 51].

In this study, the access to telehealth services is the autonomy factor; the need for telehealth services is precisely the relatedness motivation; then, knowledge of telehealth and technology skills necessary to use the telehealth services are factors to estimate the competence in SDT. As hypothesized, access, need, knowledge, and technology skills can affect telehealth utilization directly or indirectly through telehealth satisfaction. Based on SDT, the theoretical model and factors in this study were depicted graphically in Fig. 1. (See Fig. 1: Conceptual model based on Self-Determination Theory)

OPEN IN VIEWER

Fig. 1

Conceptual model based on Self-Determination Theory (SDT).

3.1 Research design and planning

This study aims to look at how often and how well people utilize telehealth in the United States. Many variables will contribute to how often and well these services are used, such as satisfaction, accessibility, type of healthcare, etc. With COVID-19 being an integration of everyone’s daily life, we wanted to analyze why people in the United States are not utilizing telehealth to its full potential. This existing gap could lead to better implementation and improvement of the telehealth services that are currently in place in the United States. Thus, existing literature were investigated for this research to see any similarities that this research and others may have curated. This study consists of distributing an online survey to consenting adult volunteers in the United States. The survey participants anonymously participated by answering questions related to telehealth services in the US. By distributing the survey link on social platforms and to colleagues, the data were collected anonymously and effectively via an electronic survey.

3.2 Sampling, recruitment, and instruments

The primary research tool administered for this study was an anonymous survey to ascertain customer satisfaction with telemedicine services. This instrument was based on survey instruments pulled from several studies on telehealth research before or during COVID-19 [33, 48]. The IRB protocol “Telehealth utilization in the US during COVID-19” was submitted and approved by Metropolitan State University Denver’s Institutional Review Board. Survey questions were modified to fit the current research focus but were influenced by telehealth literature examples. The survey was distributed to a sample of patients, primarily from two states in the US, Colorado (CO) and Texas (TX), through social media and email recruitment. In addition, the email provided potential participants with an anonymous survey link administered through the Qualtrics survey program.

3.3 Data

In the survey, participants were asked a series of questions regarding basic demographic information, information on health behaviors, and then perceptions of telehealth satisfaction and telehealth utilization. Participants were instructed that they could choose not to answer any of the questions or end their participation at any time for a reason. From February to March 2022, we received 175 observations, but some were incomplete and contained missing values. Regarding the physician-induced demand for healthcare services, survey participants include healthcare providers. After data cleaning, 142 observations remained as complete data. Using the structural equation modeling (SEM) approach, an analysis was conducted with the data to capture the mediator effect with the bootstrapping method. The skewness and kurtosis for each factor suggest a normal distribution for each individual variable; however, the Multivariate for Kurtosis and c.r. (>5) indicates that the multivariate normality may not hold for the dataset. With bootstrapping technique, Bollen-Stine Bootstrap, we could handle this issue. (See Table 1: Descriptive Statistics)

The dependent variable in this study is the utilization of telehealth services (use), with five exogenous indicators, including the possibility, frequency, convenience, variety, and repeated use of telehealth services. In addition, there are four endogenous variables, access to Telehealth (access), the need for Telehealth (need), knowledge about Telehealth (know), and necessary technical skills to use telehealth services (tech). Finally, the mediator, telehealth satisfaction (satis), refers to the satisfaction with the telehealth system and services, comfort, and ease of use of telehealth services. We hypothesize that access, need, know, and tech could directly affect use or indirectly through satis.

Each variable was measured by three to five indicators, as shown in Table 2. With the four indicators, the intermediate variable, satis, is a mediator with an intervention effect between the four independent variables and the dependent variable, use. Specifically, the endogenous variables (access, need, know, and tech) directly affect use and indirectly through affecting satis. First, individuals’ satisfaction with Telehealth will affect their possibility and frequency of telehealth utilization. Meanwhile, the access, needs, knowledge, and technical skills may also affect telehealth utilization. In addition, access, needs, knowledge, and technical skills can affect individuals’ satisfaction, which, in turn, affects their utilization of telehealth services. As shown in Fig. 1, we sketch the direct and indirect paths from the endogenous variables to the intermediate variable to telehealth utilization in the Structural equation modeling diagram. (See Table 1: Descriptive Statistics and Table 2: Constructs and measurement factors from the Questionnaire)

With the theoretical framework based on Self-Determination Theory (SDT) design, we analyzed the data using structural equation modeling (SEM). In addition, various statistical analyses, such as reliability, validity, exploratory, and confirmatory analyses, were undertaken using SmartPLS, SPSS, and AMOS to assess the applicability of the measurement model and scales utilized. To determine the convergent validity of each construct, we performed a factor analysis to ensure the survey items could produce the expected number of factors and that each item loaded on the relevant construct. In addition, using a reliability test, we test the internal consistency of the research instrument. Then, we also check the new model fit based on the criteria. At last, we computed the indirect, direct, and total effects using the bias-corrected bootstrapping method to examine further whether the mediating effect was significant.