Medical teleconsultation from the patient’s perspective. A demographic segmentation

Vol.:(0123456789) The European Journal of Health Economics https://doi.org/10.1007/s10198-024-01753-4 ORIGINAL PAPER Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation JorgeArenas‑Gaitán1 · PatricioE.Ramírez‑Correa2 · PabloLedesma‑Chaves1 · LuisJ.CallarisaFiol3 Received: 28 September 2023 / Accepted: 16 December 2024 © The Author(s) 2025 Abstract Medical teleconsultation is a tool that is here to stay among the services offered by health systems. Therefore, it is important to understand the process of adopting this technology. However, most studies have endorsed the point of view of health professionals. Our research adopts the patient’s point of view with a sample of 1500 patients who have used teleconsultation in Spain between May and November 2022, therefore, in a post-COVID-19 scenario. We started from a technology acceptance model, UTAUT, and applied a novel segmentation technique: Pathmox. As a result, we have obtained six segments of patients using teleconsultation with differentiated technology acceptance processes, and we also propose strategies adapted to each of them. Keywords Telemedicine· Teleconsultation· Demographic segmentation· Technology acceptation· Pathmox Introduction One of the priority objectives of the European Union is to ensure that its citizens have access to the health systems of its Member States. EU countries aim to ensure that their health systems provide affordable, equitable and high-quality medical care, emphasising that health care is a fundamental human right (European [26]). In order to achieve these goals, it is essential to advance the process of digitalisation of health systems [51]. From an economic perspective, the European Commission itself estimates that improved access to, and exchange of, health data could save €5.5 billion over the next 10years (European [27]). Healthcare systems have used advances in information and communication technologies since their inception. For example, the first steps in telemedicine date back to the mid-twentieth century, with the use of the telephone as a means of medical consultation. Later, in the 1970s, during the space race, NASA developed telemedicine tools to provide health care to its astronauts. However, the recent COVID-19 pandemic was a turning point in the development of these tools [62]. Today, in a society heavily influenced by technology at all levels, all kinds of devices such as mobile phones, cameras or wearable biosensors have been incorporated to obtain clinical information [1]. This technological advance affects individuals and society as a whole, implying a digital transformation [32] in the field of health systems. However, this digital transformation does not affect everyone equally. There are important differences between individuals and social groups when it comes to coping with the use of digital technologies [13, 39]. In this digital context, the patient plays an increasingly active role [85]. And among all the digital services available, a key service is the medical consultation as a moment of interaction between doctor and patient. The digital environment has given way to this [36]. The study by [22] suggests that direct-to-consumer telemedicine use might lead to increased medical service utilisation in the short and intermediate term. This trend could * Jorge Arenas-Gaitán [email protected] Patricio E. Ramírez-Correa [email protected] Pablo Ledesma-Chaves [email protected] Luis J. Callarisa Fiol [email protected] 1 Departamento de Administración de Empresas y Marketing, Facultad de Ciencias Económicas y Empresariales, Universidad de Sevilla, 41018Seville, Spain 2 Escuela de Ingeniería, Universidad Católica del Norte, Coquimbo, Chile 3 Departamento de Administración de Empresas y Marketing, Universidad Jaume I, 12071CastellódelaPlana, Spain J.Arenas-Gaitán et al. result from the ease of access and the availability of technology but could also encourage overuse or unnecessary use if a strong connection to primary care is not established. To address these risks, the authors underscore the need to identify which diagnoses and treatments are suitable for direct-to-consumer telemedicine, aiming to maximise its cost-effectiveness while curtailing its use where this could be inappropriate. This requires a thorough examination of behaviour patterns among telemedicine users, which could offer insights to refine public policies and foster efficient practices. Meanwhile, post-COVID-19, Europe is turning to telemedicine to bridge gaps in access to primary healthcare services. A key component of this strategy is the establishment of integrated primary care centres that harness technology to manage electronic medical records and promote collaborative clinical documentation. This model is particularly beneficial for healthcare professionals who practise telemedicine. These centres can substantially enhance coordination among professionals and ensure high-quality care for patients, even in remote or underserved areas [30]. To achieve this objective, a thorough understanding of teleconsultation users’ practices and preferences is essential, as this information could drive the effective management and optimisation of these centres. There are a significant number of studies on medical teleconsultation [31], most of them from the point of view of health care professionals. However, there are still few studies from the patients’ point of view [7]. Moreover, we know from other highly digitised sectors, such as electronic banking [79] or online social networks (Villarejo-Ramos, Peral-Peral, and Arenas-Gaitán 2019) that the process of acceptance of new technologies is not homogeneous among all users. There is heterogeneity, differences, between segments of users of online services. We believe that medical teleconsultation is no exception. The aim of this paper is to analyse the acceptance of teleconsultation technology by differentiating between different types of users. In order to achieve this objective, we will translate it into several research questions. The first research question examines the acceptance process of teleconsultation from the patient’s perspective. To address this, we will utilise the Unified Theory of Acceptance and Use of Technology (UTAUT) [77] allows for an adequate analysis of the technological acceptance process of medical teleconsultation. This model has been widely used in the field of telemedicine [31], although there are still few studies for the case of medical teleconsultation. Nonetheless, we understand that this acceptance process may be influenced by the characteristics of individuals. Therefore, the second research question is whether the socio-demographic characteristics of the users, such as their age, level of education or income, influence the process of acceptance of teleconsultation. To answer this question, we will use the Pathmox technique [44], which allows us to distinguish different groups, in the form of a tree, using distinct segmentation criteria at the same time. Given that each segment has its own characteristics, we will analyse the similarities and differences between the different segments obtained above with respect to their medical teleconsultation acceptance process. To test this, we will use the PLS-MGA multigroup analysis [65]. The paper is organised in a manner that first provides a comprehensive review of the relevant literature. Subsequently, the methodology employed in the study is explained. The main findings of the analysis are then presented, followed by a discussion of the results in comparison to other studies. The academic, managerial, and social implications of the research are also highlighted. The paper concludes with an assessment of the limitations of the research and suggestions for future research avenues. Theoretical framework Telemedicine andmedical teleconsultation Telemedicine, as defined by the World Health Organisation (WHO), is a healthcare practice that leverages interactive audio-visual and data communications technologies to provide medical services, including diagnosis, treatment, consultation, health education and the exchange of medical data [2]. The use of telemedicine has been widely accepted as an effective solution for remote healthcare, particularly in areas with limited or inaccessible healthcare facilities [56]. With the widespread adoption of technology, telemedicine has gained popularity by enabling medical specialists to offer their services to patients without the need for physical travel [56]. Telemedicine represents the progression of healthcare into the digital age and is poised to shape the future of medical practices [18]. Telemedicine, as it is practised today, utilises the computing devices of either the patient or healthcare professional, along with low-cost proprietary equipment such as smartphones, biosensors and laptops, to gather clinical data, thus obviating the need for extensive training [54]. This has resulted in a reduction of travel expenses and saved time, while also lowering medical costs and facilitating greater access to specialist medical practitioners for the general public without the need to interrupt their daily activities, thereby increasing overall productivity. Additionally, it has alleviated the workload of healthcare professionals by decreasing missed appointments and cancellations, which in turn has increased revenue and patient throughput, and has led to improved follow-up care and overall health outcomes [6]. The advent of telemedicine has sparked a migration of healthcare from traditional clinics and hospitals to the home environment [49]. As can be Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation seen, telemedicine is a broad concept. Therefore, Tables1, 2 summarise the main types of telemedicine according to different criteria and their main applications today. The implementation of telemedicine in healthcare has faced numerous challenges, hindering its widespread adoption. These challenges include a lack of awareness among patients, the high cost of implementation, operational inefficiencies, difficulties in conducting physical examinations, a general perception that virtual care is not as effective as in-person care, financial implications, legal and regulatory hurdles, and concerns about medical liability [52]. In this context, it is crucial to understand medical teleconsultation as a pivotal component within the realm of telemedicine. Advances in technology, such as computers, smartphones and tablets, have elevated the traditional doctorpatient and doctor-doctor communication methods beyond just auditory means [48]. The ability to transmit laboratory results, diagnostic images, videos, and even conduct video consultations to examine a patient’s own pathology, has been demonstrated through numerous studies to be highly effective. Researchers like Augusterfer etal. [5] and Mishkin etal. [55] have analysed teleconsultations in the fields of psychology and psychiatry, and despite acknowledging the Table 1 Telemedicine types Own elaboration based on Chellaiyan, Nirupama, and Taneja [18] Category Telemedicine type Description According to the timing of the information transmitted Real time or synchronous telemedicine Where the sender and receiver are online at the same time and information is transferred live Store-and-forward or asynchronous telemedicine Where the sender stores information in databases and sends it to the receiver to review at their convenience Remote monitoring type of telemedicine Uses various technological devices to remotely monitor a patient’s health and clinical signs According to the interaction between the individuals involved Health professional to health professional Provides easier access to speciality care, referral, and consultation services Health professional to patient Delivers healthcare to underserved populations by providing direct access to a medical professional Table 2 Telemedicine applications Own elaboration based on Chellaiyan, Nirupama, and Taneja [18] Category Telemedicine application Description Educational Tele-education Interactive long-distance learning programme for training and updates on recent medical advances Tele-conferencing Virtual discussions and interactions between doctors during workshops, conferences, and continuing medical education programmes Tele-proctoring Remote mentoring and evaluation of surgical trainees using advanced video conferencing equipment Healthcare delivery School-based health centres Manages chronic conditions like asthma, diabetes, and obesity by providing school nurses remote access to specialist medical opinions Correctional facilities Addresses inmates’ healthcare needs without the costs and risks of inmate transportation or requiring specialist visits Mobile health clinics Provides quick access to remote physicians or medical specialists Shipping and transportation Helps avoid evacuations and unscheduled diversions during medical emergencies Industrial health Offers on-site medical management and triage advice Healthcare management Tele-health care Uses ICT for preventive and promotive healthcare, including teleconsultation and tele follow-up Tele-home health care Monitors patients remotely with a Computer Telephone Integrated (CTI) system for 24-h vital signs monitoring Specialties Includes tele-ophthalmology, tele-psychiatry, tele-cardiology, tele-surgery, etc Diagnostic services Provides tele-radiology and tele-endoscopy services Emergency and preventive care Disaster management Ideal for disaster-stricken regions with disrupted connectivity, using satellite and customised telemedicine software Screening of diseases Uses telemedicine technologies for early detection and screening of diseases J.Arenas-Gaitán et al. need for professional training, they highlight the benefits and potential for growth. Similarly, Wright and Honey [84] studied the application of telemedicine in home care, while Vural and Ramadan (2019) explored its effectiveness in emergency situations. Furthermore, it has been applied to address the consequences of gender-based violence [80] and has been utilised in consultations related to the COVID-19 pandemic with high levels of patient satisfaction (Blanco [11]). Vranda and Cicil [80] posit that the success of consultation tools is contingent upon a number of factors, including feasibility, device size, enhancement of face-to-face interaction, capability of transmitting patient data, data type, user-friendly interface and portability, and a synchronous or asynchronous mode of operation. The interaction can occur through a variety of means, such as mobile applications, text-based methods using specialised smartphone applications or chat-based systems and platforms (e.g., WhatsApp, Google Hangouts, Facebook Messenger), video chat platforms (e.g., Skype, Facetime), and even asynchronous media like emails and faxes. While each consultation tool has its own unique advantages and disadvantages, it is however evident that the field of medical teleconsultation is undergoing continuous evolution due to technological advancements (Vural and Ramadan 2019b). From the patient’s perspective, medical teleconsultation has been studied through theories of technology adoption, with theoretical frameworks such as the Technology Acceptance Model (TAM) and UTAUT [8, 50, 61]. The TAM was developed to be applied to technologies at the workplace level [23, 25]. The TAM model is based on the idea that the acceptance of a technology depends on its usefulness and ease of use [24]. Subsequently, the UTAUT model [77] is a revision and extension of the TAM with contributions from other related theories, incorporating concepts such as social influence or facilitating conditions. We believe that UTAUT offers a suitable conceptual framework to study the process of adoption of teleconsultation. More specifically, UTAUT [77] proposed four latent variables that determine user acceptance and usage behaviour (USE): Performance Expectancy (PE), Effort Expectancy (EE), Social Influence (SI), and Facilitating Conditions (FCs). PE is defined as the degree to which a person using an application believes that this will help him or her achieve a better job performance. EE is defined as the degree of perceived ease of use of an application. SI is defined as the degree to which an individual perceives that important others believe that he or she should use an application. FCs are defined as the extent to which an individual believes that an organisational/technical infrastructure exists to assist the use of an application. These four constructs directly affect behavioural intention (BI). In addition, behavioural intention directly influences the use of the technology, and the facilitating conditions directly determining use behaviour of technology. It can be posited that an increase in perceptions of PE, EE, SI and FC will result in an increase in the utilisation of a technology through BI [77]. Based on this idea, we formulate the following hypothesis: H1: The UTAUT model provides a robust theoretical basis for examining the adoption processof medical teleconsultation from the user’s perspective. Exploring theintersection offamily structures, socio‑demographics, andtechnology adoption Traditionally, the application of models to analyse a reality has assumed that the data under analysis comes from a homogeneous population. However, in the field of social sciences, and more specifically in the study of human behaviour, this premise is often unrealistic, given that individuals are likely to exhibit heterogeneity in their perceptions and evaluations of a phenomenon [53, 64, 66]. This assertion applies to the use of medical teleconsultation services as well. Family characteristics, such as size and composition, as well as individuals’ level of technology literacy, play a crucial role in the adoption of these new technologies [21]. Education, training, the number of household members, and the geographic location also influence the uptake and usage of new technologies, and some studies even consider race and income [58]. For instance, families with children or older members tend to exhibit different behaviours compared to families without such members [37, 46, 72]. In this regard, we present some demographic variables that have been shown to affect the behaviour of technology users: Age Numerous studies have highlighted age as the most influential demographic factor affecting technology adoption [9, 10, 47]. This phenomenon often results from the natural decline in cognitive abilities or older adults’ self-perception of feeling aged. The belief that their cognitive skills have diminished becomes a barrier to adopting new technology [28]. Additionally, the literature review suggests that seniors’ perception of the complexity of interactive technology significantly impacts their adoption decisions. In this respect, telemedicine applications can be complex. Educational level Indeed, research consistently highlights the impact of the education level on technology adoption [9, 10]. The theory of learning suggests that individuals with lower education levels may struggle due to simpler cognitive structures, hindering their ability to adapt to new environments [57]. In contrast, those with higher education tend to approach technology with less apprehension [68]. Their awareness Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation and positive attitudes towards new technology contribute to higher adoption rates. Income There is a significant body of research indicating that income significantly affects the process of technology adoption [9, 10]. Lower-income consumers tend to be more cost-sensitive and resist investing in new technology [69]. Higher income levels are associated with greater self-confidence, leading to a higher perceived ability to use new technology. Lower-income individuals often view new technology as useless. Adopters of new technology generally have higher income levels than non-adopters. However, the analysis of these variables separately gives a partial view of the reality. There may be connections between these socio-demographic variables at different levels which need to be taken into account. For example, despite the growing interest in understanding the uptake of information technologies in the healthcare sector from the patient’s perspective, the majority of studies focus specifically on the elderly demographic. This is demonstrated by Kavandi and Jaana’s [37] findings which indicate that 51% of research on the adoption of Health Information Technologies (HIT) by older adults fails to consider an adoption model or framework. There are conflicting results regarding the impact of sociodemographic variables, such as age, gender, and education, on HIT adoption in older adults, suggesting that a broader age range should be studied to fully understand these effects [20]. On the other hand, Chimento-Díaz etal. [19] found, in their study of technology adoption in those over 64years old based on the TAM, that factors such as younger age, higher education, and zest for life positively influence acceptance of technology use in this socio-demographic. Lastly, Tsertsidis, Kolkowska, and Hedström [75] observed that technology acceptance post-implementation is influenced by multiple factors, including age, with views on technology changing for the better as older adults realise its various benefits in their daily lives. In summary, the current body of literature regarding the acceptance of health technologies, particularly teleconsultation, lacks sufficient exploration from the patient’s perspective, especially with regards to a more comprehensive analysis of the influence of socio-demographic factors. Based on the analysis of the above literature, we propose the following hypothesis: Hypothesis 2: The combination of socio-demographic characteristics, including age, education, and income levels, among users of medical teleconsultation allows for the identification of statistically distinct segments based on their technology acceptance behaviors. Methodology Scales ofmeasurement The measurement scales used in our investigation had been rigorously evaluated and validated in prior research. The scales used to assess Performance Expectancy, Effort Expectancy, Facilitating Conditions, Social Influence and Behaviour Intention were adapted from the UTAUTmodel proposed by Venkatesh etal. [77] and Venkatesh, Thong, and Xu [78]. Meanwhile, the scale for measuring the utilisation of teleconsultation was derived from [40]. The variables pertaining to the UTAUT model were quantified using five-point Likert scales. In addition, a comprehensive examination of various socio-demographic factors associated with the participants was carried out, including their age and level of education. Furthermore, inquiries regarding the composition of their household were conducted, including the size of the household, the presence of minors, and the number of individuals over 70years of age residing within the household. These socio-demographic variables may be either qualitative, such as educational attainment, or numerical, such as age. To ensure that the results of our analysis are free from Common Method Bias, we have employed the [38] test to assess the Variance Inflation Factor (VIF) of our variables. Our findings indicate that all the VIF levels are below 3.3, thereby guaranteeing the absence of Common Method Bias in our study. Sample We utilised a non-probability sample of telemedicine users. The data were collected for Spain as a whole between May and November 2022. This period followed the mandatory confinement in Spain due to the Covid-19 pandemic from March to June 2020. To ensure the suitability of participants, we required them to be over 18years old and to have used medical teleconsultation services within the past year. We engaged a company specialising in electronic questionnaire data collection. In total, we obtained a sample of 1500 individuals. Additionally, we implemented a filtering process to ensure sample quality. Specifically, we excluded questionnaires completed in less than three minutes, considering this duration insufficient for reliable responses. The average completion time for the questionnaire was just over five minutes. Similarly, we removed questionnaires with inconsistent responses. As a result of this process, we obtained a final sample of 1412 individuals for our analyses. A priori, this is a sufficient J.Arenas-Gaitán et al. sample [74, 83] for the proposed structural model, with an anticipated effect size of 0.11 and a desired statistical power level of 0.8. Of the respondents, 64.4% were male and 35.5% were female. The majority of the participants, 68.2%, resided in urban areas with a population of more than 50,000 while 31.8% lived in rural locations. In terms of education, 1.7% had no or basic education, 59.3% had secondary education, and 39% had a university education. Further demographic details on the sample can be found in the accompanying Table3. The Spanish health system follows the Beveridge model. This model is characterised by tax-based financing, universal access, salaried or capitated doctors, a minor role for the private sector and a strong state involvement in management. In the European context, other states that follow this model include Portugal, Italy, the United Kingdom, Ireland, Denmark, Finland, and Sweden. Statistical tools To conduct our study, we employed several statistical methods. Firstly, we used structural equation modelling, specifically PLS-SEM, to validate our UTAUT model. This model serves as a foundation for the creation of Pathmox [44], which was utilised to analyse the diversity in the responses, leading to a segmentation of non-face-to-face consultation users. The Pathmox technique relies on constructing a binary tree to detect population segments with diverse behaviours in relation to the proposed structural equation model, PLS. Finally, we employed multi-group analysis, MGA-PLS [17, 45], to assess and examine the behavioural differences between the identified segments. Results PLS‑SEM The Partial Least Squares Structural Equation Modelling (PLS-SEM) methodology encompasses two distinct stages: the examination of measurement scales, followed by the evaluation of the structural model. To present the outcomes of this analysis, in our reporting we shall adhere to the guidelines put forward by Hair etal. [33]. At the onset of our analysis of the measurement scales, it is imperative to emphasise that all the constructs were deemed to be reflective in our study. To verify the reliability of the individual items, we evaluated each item’s loading value and confirmed that they all exceeded the recommended 0.7. The reliability of the constructs was further established through Composite Reliability and Cronbach’s Alpha, both of which revealed values above 0.7, in line with the established literature [33]. The convergent validity was established through the Analysis of Variance Extraction (AVE), yielding values above the threshold of 0.5. These findings are presented in Table4. To ensure discriminant validity, we employed the Fornell and Larcker [29] and HTMT tests [35], as presented in Tables4, 5. Overall, our results demonstrate the suitability of the measurement scales used in the study. The objective of analysing the structural model is to determine the paths and R2 values. The paths reflect the magnitude of the correlation between the dependent and independent variables, while the R2 values demonstrate the proportion of variance explained by each dependent variable in the model. A bootstrapping approach with 10,000 subsamples was employed in this analysis. The results of these measurements are presented in Fig.1. Furthermore, the SRMR was used as a metric of the model’s goodness of fit, yielding a value of 0.064, which is lower than the benchmark of 0.08 established by Hair etal. [33]. The combination of the SRMR and R2 values suggests that the structural model has a satisfactory explanatory capability. It should be noted that when considering path values, Performance Expectancy and Social Influence are significant predictors of Behavioural Intention. There is a robust correlation between Behavioural Intention and Use. However, Effort Expectancy and Facilitating Conditions do not have a significant impact on either current use or the intention to use non-face-to-face consultations among the full sample of 1412 respondents. These overall results may not accurately depict the complex and diverse behaviours of different segments. Through heterogeneity analysis, we can study these unique behaviours, which may not be revealed in the overall results [4] To uncover these segments, Pathmox analysis will be applied. Pathmox Pathmox is an original idea by Gastón Sánchez [63] to discover heterogeneity in a structural equation model based on decision trees. Decision trees can uncover hidden decision rules and allow high interpretability to explain real applications [86]. Generally, heterogeneity is associated with a mixture of populations that form differentiated segments [43]. This Table 3 Sample demographic characteristics Min Max Average Age 18 74 38.6 Household size 1 10 3.6 Number of children under 18 in the household 0 6 1.1 Number of people over 70 in the household 0 5 0.2 Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation Table 4 Indicators of measurement scales AVE average variance extracted, CR composite reliability, CA Cronbach’s alpha Effort expectancy (EE) Global Nod 4 Nod 5 Nod12 Nod13 Nod14 Nod15 AVE 0.748 0.668 0.701 0.825 0.850 0.685 0.785 CR 0.922 0.889 0.903 0.950 0.958 0.896 0.936 CA 0.862 0.834 0.857 0.929 0.941 0.845 0.909 Learning how to use teleconsulting is easy for me 0.851 0.805 0.796 0.903 0.925 0.834 0.858 My interaction with teleconsulting is clear and understandable 0.891 0.858 0.865 0.919 0.933 0.909 0.911 I find teleconsulting easy to use 0.884 0.846 0.862 0.939 0.931 0.812 0.911 It is easy for me to become skilful at using teleconsulting 0.832 0.755 0.822 0.871 0.898 0.747 0.863 Performance expectancy (PE) AVE 0.830 0.772 0.800 0.852 0.890 0.885 0.851 CR 0.900 0.910 0.923 0.945 0.960 0.959 0.945 CA 0.830 0.853 0.875 0.914 0.938 0.935 0.913 I find teleconsulting useful in my daily life 0.915 0.891 0.908 0.929 0.936 0.932 0.927 Using teleconsulting increases my chances of achieving things that are important to me 0.903 0.843 0.902 0.901 0.942 0.934 0.921 Using teleconsulting helps me accomplish things more quickly 0.916 0.900 0.873 0.940 0.952 0.957 0.920 Social influence (SI) AVE 0.883 0.833 0.840 0.950 0.923 0.909 0.923 CR 0.958 0.937 0.940 0.983 0.973 0.968 0.973 CA 0.934 0.900 0.905 0.974 0.959 0.950 0.958 People who are important to me think that I should use teleconsulting 0.937 0.912 0.906 0.972 0.965 0.918 0.960 People who influence my behaviour think that I should use teleconsulting 0.939 0.909 0.911 0.983 0.959 0.970 0.957 People whose opinions I value prefer that I use teleconsulting 0.943 0.917 0.932 0.969 0.959 0.972 0.965 Facilitating conditions (FC) AVE 0.705 0.657 0.632 0.821 0.781 0.617 0.711 CR 0.905 0.884 0.873 0.948 0.934 0.866 0.908 CA 0.862 0.827 0.807 0.927 0.906 0.808 0.870 I have the resources necessary to use teleconsulting 0.853 0.854 0.770 0.922 0.894 0.771 0.834 I have the knowledge necessary to use teleconsulting 0.857 0.835 0.816 0.917 0.912 0.779 0.833 Teleconsulting is compatible with other technologies I use 0.839 0.763 0.805 0.903 0.902 0.827 0.863 I can get help from others when I have difficulties with teleconsulting 0.810 0.788 0.790 0.881 0.824 0.764 0.843 Behavioural intention (BI) AVE 0.812 0.762 0.790 0.816 0.853 0.861 0.839 CR 0.928 0.906 0.918 0.930 0.946 0.949 0.940 CA 0.884 0.844 0.867 0.887 0.914 0.919 0.904 I intend to continue using teleconsulting in the future 0.880 0.856 0.865 0.869 0.898 0.916 0.898 I will always try to use teleconsulting in my daily life 0.909 0.877 0.901 0.907 0.934 0.927 0.927 I plan to continue to use teleconsulting frequently 0.914 0.886 0.900 0.934 0.938 0.941 0.923 Use AVE 0.812 0.782 0.786 0.785 0.832 0.792 0.853 CR 0.928 0.915 0.917 0.916 0.937 0.920 0.946 CA 0.882 0.860 0.864 0.864 0.899 0.870 0.914 I tend to use teleconsulting frequently 0.906 0.871 0.883 0.907 0.921 0.914 0.940 I spend a lot of time on teleconsulting 0.885 0.898 0.866 0.836 0.886 0.857 0.895 I get involved a lot in teleconsulting 0.912 0.883 0.910 0.913 0.929 0.898 0.935 J.Arenas-Gaitán et al. characteristic between units is a significant problem in data analysis. When a sample is heterogeneous but is treated as homogeneous, the quality of the study can be affected, generating biases in the interpretation [15]. Specifically, Pathmox is oriented to look for differences at the structural level. That is, it searches for segments that differ in the effects between the latent variables of the model. For this purpose, Pathmox uses a recursive algorithm based on binary decision tree learning [59] to analyse the models associated with different data segments. These data segments are generated by dividing the sample based on additional analysis variables. The analysis variables are external to the structural equation model. Given the objective of Pathmox, the procedure uses the F-test to determine the discrepancy of the true coefficients in two linear regressions associated with different data segments. The Pathmox algorithm can be summarised as follows. All possible pairs of segments based on the analysis variables are generated, and then the pair of segments with the most significant discrepancy in their effects is selected. Next, a procedure like the previous one is executed for each segment of this pair, and so on, until the differences are not significant, the size of the segments is tiny, or the number of segmentation levels is many. Finally, the algorithm delivers the discovered segments and the variables based on this segmentation. As there are more analysis variables, it is possible to test more potential segment configurations. However, the time complexity of the algorithm increases [86]. In the present study, seven analysis variables were used to execute the Pathmox: educational level, place of residence, income, age, number of household members under 18years, and number of elderly household members. The algorithm was run in the R environment [60] using the genpathmox package [42]. The result of the procedure is shown in Fig.2. As can be seen, Pathmox discovered six segments with differences at the structural level, as detailed below. In general, Pathmox selects the analysis variables age, educational level, and income. Age is the first variable that discriminates. Then, the educational level determines the difference in individuals under 39years of age. In individuals aged 39 and over, the educational level and income determine this difference. Specifically, the node labelled four comprises 432 individuals between 18 and 38years old with a K-12 education level. This segment is the largest. Node five has 252 individuals between 18 and 38years old with a university education level. The node labelled twelve Table 5 Discriminant validity Note: Diagonal elements (bold) are the square root of the variance shared between the constructs and their measures (AVE). Under the diagonal elements are the correlations between constructs [29]. The elements above the diagonal are HTML test values Behavioural intention Effort expectancy Facilitating conditions Performance expectancy Social influence Use Behavioural intention 0.901 0.555 0.468 0.674 0.714 0.723 Effort expectancy 0.623 0.865 0.739 0.707 0.500 0.412 Facilitating conditions 0.528 0.847 0.840 0.566 0.436 0.333 Performance expectancy 0.756 0.789 0.637 0.911 0.600 0.540 Social influence 0.785 0.544 0.472 0.654 0.940 0.709 Use 0.812 0.455 0.365 0.600 0.777 0.901 Fig. 1 UTAUT global result Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation comprises 106 individuals between 39 and 74years old with a K-12 education level and low income. Node thirteen has 232 individuals between 39 and 74years old with a K-12 education level and non-low income. The node labelled fourteen is made up of 85 individuals between 39 and 74years old with a university education level and low income. This segment is the smallest. Finally, node fifteen is formed of 214 individuals between 36 and 74years old with a university education level and non-low income. We calculated the statistic power using G*Power software; the minimum statistic power calculated was 0.764. The 10-times rule (Hair etal. 2011) has been applied, see Table6. The main features of the segments are summarised in Table7, based on the supported relationships in the research model we proposed labels for. Following the Pathmox results, six segments have been identified (see Table7). On the one hand, there are two segments made up of young people. The first segment, Curious Explorers (Node 4), comprises young individuals below 39years of age with no university education and is the largest among the segments. Analysing the teleconsultation adoption process of this segment compared to the rest, we find that it is characterised by performance expectancy (PE) significantly affecting the intention to use this tool (BI) but less so than other segments. This result could be explained by the idea that in a segment made up of young people who are very familiar with new technologies in all aspects of their lives, they may have normalised the advantages of digital media, so their expectations of usefulness are lower when not compared to offline alternatives. These results are also applicable to Node 5. Dynamic Educated (Node 5), the second segment, is made up of young people below 39years with a university education. This segment, in addition to its significant but low PE to BI ratio (as in Curious Explorers), is particularly characterised by the fact that the facilitating conditions (FC) are significant. In other words, for this segment of university-educated young people, it is important to have suitable devices to carry out medical teleconsultation. This last result may be related to disposable income in this Fig. 2 Pathmox result Table 6 Estimation of the statistic power Global Node 4 Node 5 Node 12 Node 13 Node 14 Node 15 Sample size 1412 432 252 106 323 85 214 10-times rule 40 40 40 40 40 40 40 Statistic power 1.000 0.999 0.999 0.868 0.998 0.764 0.997 Table 7 Summary of Pathmox segment characteristics Node Label Age Educational level Income Differential characteristics compared to the UTAUT global model 4 Curious Explorers 18–38 No University PE- > BI significant, but low 5 Dynamic Educated 18–38 University FC- > BI significant 12 Experienced Experimenters 39–74 No University Very low PE- > BI not significant 13 Resilient Adapters 39–74 No University Low to High Similar to the global model 14 Insightful Graduates 39–74 University Low and very low PE- > BI significant, high 15 Critical Users 39–74 University Medium to High FC- > USE, significant and negative J.Arenas-Gaitán et al. 17 Cheah, J.H., Thurasamy, R., Memon, M.A., Chuah, F., Ting, H.: Multigroup analysis using smartpls: step-by-step guidelines for business research. Asian J. Bus. Res. (2020). https:// doi. org/ 10. 14707/ ajbr. 200087 18. Chellaiyan, V.G., Nirupama, A.Y., Taneja, N.: Telemedicine in India: where do we stand? J. Fam. Med. Prim. Care 8(6), 1872 (2019) 19. Chimento-Díaz, S., Sánchez-García, P., Franco-Antonio, C., Santano-Mogena, E., Espino-Tato, I., Cordovilla-Guardia, S.: Factors associated with the acceptance of new technologies for ageing in place by people over 64 years of age. Int. J. Environ. Res. Public Health 19(5), 2947 (2022). https:// doi. org/ 10. 3390/ ijer p h1905 2947 20. Chua, S.L., Chem, D., Wong, P.: Computer anxiety and its correlates: a meta analysis. Comput. Hum. Behav. 14, 559–577 (1999). https:// doi. org/ 10. 1016/ S07475632(99) 00039-4 21 Creaser, A.V., Hall, J., Costa, S., Bingham, D.D., Clemes, S.A.: Exploring families’ acceptance of wearable activity trackers: a mixed-methods study. Int. J. Environ. Res. Public Health (2022). https:// doi. org/ 10. 3390/ ijerp h1906 3472 22. Dahlgren, C., Spånberg, E., Sveréus, S., Dackehag, M., Wändell, P., Rehnberg, C.: Short-and intermediate-term impact of DTC telemedicine consultations on subsequent healthcare consumption. Eur. J. Health Econ. 25(1), 157–176 (2024). https:// doi. org/ 10. 1007/ s1019802301572-z 23. Davis, F. D. 1985. A technology acceptance model for empirically testing new end-user information systems: theory and results. 24. Davis, F.D.: Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 13(3), 319–340 (1989). https:// doi. org/ 10. 2307/ 249008 25. Davis, F.D., Bagozzi, R.P., Warshaw, P.R.: User acceptance of computer technology: a comparison of two theoretical models. Manage. Sci. 35(8), 982–1003 (1989). https:// doi. org/ 10. 1287/ mnsc. 35.8. 982 26. European Commision. 2024a. EU Health Policy. Overview. https:// health. ec. europa. eu/ euhealthpolicy/ overv iew_ en# prior itiesfor20212027. Retrieved 20 Sep 2024 (https:// health. ec. europa. eu/ euhealthpolicy/ overv iew_ en# prior itiesfor20212027) 27. European Commision. 2024b. Factsheet on European Health Data Space. Brussels 28 Farivar, S., Abouzahra, M., Ghasemaghaei, M.: Wearable device adoption among older adults: a mixed-methods study. Int. J. Inf. Manag. (2020). https:// doi. org/ 10. 1016/j. ijinf omgt. 2020. 102209 29. Fornell, C., Larcker, D.F.: Evaluating structural equation models with unobservable variables and measurement error. J. Mark. Res. 18(1), 39–50 (1981) 30. Garattini, L., Martini, M.B., Zanetti, M.: More room for telemedicine after COVID-19: lessons for primary care? Eur. J. Health Econ. 22(2), 183–186 (2021). https:// doi. org/ 10. 1007/ S1019802001248-Y 31 Garavand, A., Aslani, N., Nadri, H., Abedini, S., Dehghan, S.: Acceptance of telemedicine technology among physicians: a systematic review. Inf. Med. Unlocked (2022). https:// doi. org/ 10. 1016/j. imu. 2022. 100943 32. Gradillas, M., Thomas, L.D.W.: Distinguishing digitization and digitalization: a systematic review and conceptual framework. J. Prod. Innov. Manag. (2023). https:// doi. org/ 10. 1111/ jpim. 12690 33. Hair, J.F., Risher, J.J., Sarstedt, M., Ringle, C.M.: When to use and how to report the results of PLS-SEM. Eur. Bus. Rev. 31(1), 2–24 (2019) 34 Harst, L., Lantzsch, H., Scheibe, M.: Theories predicting enduser acceptance of telemedicine use: systematic review. J. Med. Internet Res. (2019). https:// doi. org/ 10. 2196/ 13117 35. Henseler, J., Ringle, C.M., Sarstedt, M.: A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. Acad. Mark. Sci. 43(1), 115–135 (2015) 36. Hernández-Zambrano, S.M., Castiblanco-Montañez, R.A., Guerrero, P.M.Á., Moreno, M.A.G., Sacristán, L.G.M., Rodríguez-Vargas, G.S., Villareal, L., Rubio, J.A., Quintero, E., Gómez, M.A., Chávez-Chávez, J., Triana, D.P.R., Casanova, R., Rojas-Villarraga, A., Santos-Moreno, P.: Experience of tele consulting care for health care professionals and patients with rheumatoid arthritis, in the health emergency due to covid-19. J. Clin. Rheumatol. 27, 53 (2021) 37. Kavandi, H., Jaana, M.: Factors that affect health information technology adoption by seniors: a systematic review. Health Soc. Care Commun. 28(6), 1827–1842 (2020). https:// doi. org/ 10. 1111/ hsc. 13011 38. Kock, N.: Common method bias in PLS-SEM: a full collinearity assessment approach. Int. J. E-Collab. 11(4), 1–10 (2015). https:// doi. org/ 10. 4018/ ijec. 20151 00101 39. Kraus, S., Schiavone, F., Pluzhnikova, A., Invernizzi, A.C.: Digital transformation in healthcare: analyzing the current state-ofresearch. J. Bus. Res. 123, 557–567 (2021). https:// doi. org/ 10. 1016/j. jbusr es. 2020. 10. 030 40. Kwon, O., Wen, Y.: An empirical study of the factors affecting social network service use. Comput. Hum. Behav. 26(2), 254–263 (2010) 41 Lakoma, S., Pasanen, H., Lahdensuo, K., Pehkonen, J., Viinikainen, J., Torkki, P.: Quality of the digital gp visits and characteristics of the users: retrospective observational study. Scand. J. Prim. Health Care (2024). https:// doi. org/ 10. 1080/ 02813 432. 2024. 23809 21 42. Lamberti, G., Lamberti, M. G. 2014. Genpathmox. 43 Lamberti, G., Aluja, T.B., Sanchez, G.: The pathmox approach for PLS path modeling segmentation. Appl. Stoch. Models Bus. Ind. (2016). https:// doi. org/ 10. 1002/ asmb. 2168 44. Lamberti, G., Banet-Aluja, T., Sanchez, G.: The pathmox approach for PLS path modeling segmentation. Appl. Stoch. Model. Bus. Ind. 32(6), 453–468 (2017). https:// doi. org/ 10. 1002/ asmb. 2168 45. Latan, H., Noonan, R.: Partial least squares path modelling. Basic concepts, methodological issues and applications. Springer International Publishing, Cham (2017) 46. Lindsay, J.S.B.: Types of family and family types. Fam. Process 7(1), 51–66 (1968). https:// doi. org/ 10. 1111/j. 15455300. 1968. 00051.x 47. Liu, C., Hung, K., Wang, D., Wang, S.: Determinants of selfservice technology adoption and implementation in hotels: the case of China. J. Hosp. Market. Manag. 29(6), 636–661 (2020). https:// doi. org/ 10. 1080/ 19368 623. 2020. 16892 16 48. Liu, C., Lim, R., Taylor, S., Calvo, R.A.: Students’ behavioural engagement in reviewing their tele-consultation feedback within an online clinical communication skills platform. Comput. Hum. Behav. 94, 35–44 (2019) 49. Lopez, A.M., Lam, K., Thota, R.: Barriers and facilitators to telemedicine: can you hear me now? Am. Soc. Clin. Oncol. Educ. Book 41, 25–36 (2021) 50. Lu, W., Wei, D., Li, C., Gao, P., Ma, R., Zhai, Y., Wang, C.: How to promote telemedicine patient adoption behavior for greener healthcare? J. Clean. Prod. 434, 139884 (2024). https:// doi. org/ 10. 1016/J. JCLEP RO. 2023. 139884 51. Majcherek, D., Hegerty, S.W., Kowalski, A.M., Lewandowska, M.S., Dikova, D.: Opportunities for healthcare digitalization in Europe: comparative analysis of inequalities in access to medical services. Health Policy 139, 104950 (2024). https:// doi. org/ 10. 1016/j. healt hpol. 2023. 104950 52. Makhni, M.C., Riew, G.J., Sumathipala, M.G.: Telemedicine in orthopaedic surgery: challenges and opportunities. JBJS 102(13), 1109–1115 (2020) 53. Matthews, L.M., Sarstedt, M., Hair, J.F., Ringle, C.M.: Identifying and treating unobserved heterogeneity with FIMIX-PLS. Medical teleconsultation fromthepatient’s perspective. Ademographic segmentation Eur. Bus. Rev. 28(2), 208–224 (2016). https:// doi. org/ 10. 1108/ EBR0920150095 54. Mazzuca, D., Borselli, M., Gratteri, S., Zampogna, G., Feola, A., Corte, M.D., Guarna, F., Scorcia, V., Giannaccare, G.: Applications and current medico-legal challenges of telemedicine in ophthalmology. Int. J. Environ. Res. Public Health 19(9), 5614 (2022) 55. Mishkin, A.D., Cheung, S., Capote, J., Fan, W., Muskin, P.R.: Survey of clinician experiences of telepsychiatry and tele-consultation-liaison psychiatry. J. Acad. Consult. Liaison Psychiatry 63(4), 334–344 (2022) 56. Nittari, G., Khuman, R., Baldoni, S., Pallotta, G., Battineni, G., Sirignano, A., Amenta, F., Ricci, G.: Telemedicine practice: review of the current ethical and legal challenges. Telemed. E-Health 26(12), 1427–1437 (2020) 57. Paas, F., van Merriënboer, J.J.G.: Cognitive-load theory: methods to manage working memory load in the learning of complex tasks. Curr. Dir. Psychol. Sci. 29(4), 394–398 (2020). https:// doi. org/ 10. 1177/ 09637 21420 922183 58. Porter, C.E., Donthu, N.: Using the technology acceptance model to explain how attitudes determine internet usage: the role of perceived access barriers and demographics. J. Bus. Res. 59(9), 999–1007 (2006). https:// doi. org/ 10. 1016/j. jbusr es. 2006. 06. 003 59. Quinlan, J.R.: Induction of decision trees. Mach. Learn. 1(1), 81–106 (1986). https:// doi. org/ 10. 1007/ bf001 16251 60. R Development Core Team. 2021. R: a language and environment for statistical computing. R Found. Statistical Comput. 61 Rouidi, M., Elouadi, A.E., Hamdoune, A., Choujtani, K., Chati, A.: TAM-UTAUT and the acceptance of remote healthcare technologies by healthcare professionals: a systematic review. Inf. Med. Unlocked (2022). https:// doi. org/ 10. 1016/j. imu. 2022. 101008 62 Geetanjali, S., Sharma, M., Kapur, A.: Evolution of smart healthcare: telemedicine during COVID-19 pandemic. J. Inst. Eng. India Ser. B 102(6), 1319–1324 (2021). https:// doi. org/ 10. 1007/ s4003102100568-8 63. Sanchez, G.: Pathmox approach: segmentation trees in partial least squares path modeling. Universitat Politècnica de Catalunya, Barcelona (2009) 64. Sarstedt, M.: A review of recent approaches for capturing heterogeneity in partial least squares path modelling. J. Model. Manag. 3(2), 140–161 (2008). https:// doi. org/ 10. 1108/ 17465 66081 08901 26 65 Sarstedt, M., Ringle, C.M., Hair, J.F.: Partial least squares structural equation modeling. In: Handbook of market research, pp. 1–40. Springer, Cham (2017) 66 Sarstedt, M., Ringle, C.M., Hair, J.F.: Treating unobserved heterogeneity in PLS-SEM. In: Partial least squares path modeling, pp. 361–81. Springer, Cham (2017) 67 Scheufele, E., Hodor, B., Popa, G., Wang, S., Kassler, W.: Population segmentation using a novel socio-demographic dataset. Online J. Public Health Inf. (2022). https:// doi. org/ 10. 5210/ ojphi. v14i1. 11651 68. Schönauer, A.L., Glanz, S.: Hydrogen in future energy systems: social acceptance of the technology and its large-scale infrastructure. Int. J. Hydrogen Energy 47(24), 12251–12263 (2022). https:// doi. org/ 10. 1016/j. ijhyd ene. 2021. 05. 160 69 Schuetz, S., Venkatesh, V.: Blockchain, adoption, and financial inclusion in India: research opportunities. Int. J. Inf. Manag. (2020). https:// doi. org/ 10. 1016/j. ijinf omgt. 2019. 04. 009 70. Sempere, M.M., Calatayud, C.R.: Grants programme and tuition fees policy in the Spanish public university system: is it really effective? Revista de Educacion 2022(398), 135–160 (2022). https:// doi. org/ 10. 4438/ 1988592XRE2022398555 71. Shannon, R., Sthienrapapayut, T., Moschis, G.P., Teichert, T., Balikcioglu, B.: Family life cycle and the life course paradigm: a four-country comparative study of consumer expenditures. J. Glob. Sch. Market. Sci. 30(1), 34–44 (2020). https:// doi. org/ 10. 1080/ 21639 159. 2019. 16139 13 72. Sharma, R.: The family and family structure classification redefined for the current times. J. Fam. Med. Prim. Care 2(4), 306 (2013). https:// doi. org/ 10. 4103/ 22494863. 123774 73. Shin, S.H., Yun, E.K.: Understanding the online health information user profiles in Korea: from a psychological perspective. Telemed. E-Health 17(5), 341–347 (2011). https:// doi. org/ 10. 1089/ tmj. 2010. 0170 74. Soper, D. S. 2023. A-priori sample size calculator for structural equation models. https:// www. danie lsoper. com/ statc alc 75. Tsertsidis, A., Kolkowska, E., Hedström, K.: Factors influencing seniors’ acceptance of technology for ageing in place in the postimplementation stage: a literature review. Int. J. Med. Informatics 129, 324–333 (2019). https:// doi. org/ 10. 1016/j. ijmed inf. 2019. 06. 027 76. Venkatesh, M., Davis, and Davis.: User acceptance of information technology: toward a unified view. MIS Q. 27(3), 425 (2003). https:// doi. org/ 10. 2307/ 30036 540 77. Venkatesh, V., Morris, M.G., Davis, G.B., Davis, F.D.: User acceptance of information technology: toward a unified view. MIS Q. 27(3), 425–478 (2003) 78. Venkatesh, V., Thong, J.Y.L., Xu, X.: Consumer acceptance and use of information technology: extending the unified theory of acceptance and use of technology. MIS Q. 36(1), 157–178 (2012) 79 Villarejo-Ramos, Á.F., Peral-Peral, B., Arenas-Gaitán, J.: Latent segmentation of older adults in the use of social networks and e-banking services. University of Boras, Sweden (2019) 80. Vranda, M.N., Cicil, V.R.: Tele-consultation for the survivors of intimate partner violence: guidelines for mental health professionals. Indian J. Psychol. Med. 44(4), 349–353 (2022) 81. Vuik, S.I., Mayer, E., Darzi, A.: A quantitative evidence base for population health: applying utilization-based cluster analysis to segment a patient population. Popul. Health Metrics 14(1), 44 (2016). https:// doi. org/ 10. 1186/ s129630160115-z 82. Vural, S., Ramadan, H.: A short review on the comparison of consultation systems and tools in the emergency department practice: tele-consultation. Cumhuriyet Med. J. 41(2), 239–243 (2019) 83. Westland, Ch.J.: Lower bounds on sample size in structural equation modeling. Electron. Commer. Res. Appl. 9(6), 476–487 (2010). https:// doi. org/ 10. 1016/J. ELERAP. 2010. 07. 003 84 Wright, Jane, Honey, Michelle: New Zealand nurses’experience of tele-consultation within secondary and tertiary services to provide care at a distance. Nurs. Prax. N. Z. 32(2), 30 (2016) 85 Yang, H., Guo, X., Peng, Z., Lai, K.H.: Patient empowerment in an online health platform: exploring the quadratic effects of patients’ conscious-competence on perceived health status. Comput. Hum. Behav. (2022). https:// doi. org/ 10. 1016/j. chb. 2022. 107346 86 Zhao, X., Nie, X.: Splitting choice and computational complexity analysis of decision trees. Entropy (2021). https:// doi. org/ 10. 3390/ e2310 1241 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.