Adoption of mobile ICT for health promotion: an empirical investigation

Electron Markets (2010) 20:241 250 DOI 10.1007/s12525-010-0042-y GENERAL RESEARCH Adoption of mobile ICT for health promotion: an empirical investigation Mihail Cocosila & Norm Archer Received: 28 November 2009 / Accepted: 22 October 2010 / Published online: 13 November 2010 # Institute of Information Management, University of St. Gallen 2010 Abstract This research is an unbiased empirical evaluation of user reasons to accept or resist a mobile information and communication technology (ICT) application for health promotion. This innovative use of mobile ICT consists of developing services that educate people to stay healthy, with clear benefits for both individuals and society. Receiving customized health advice through mobile devices may be an attractive service. However, despite their ability to support users, mobile services may sometimes irritate by being too intrusive. A 1-month experiment exposed participants to a health promotion application, delivered through their cell phones. This was the framework for the evaluation of an adoption model that included both positive and negative user adoption factors. Findings revealed intrinsic motivation to be a sufficient reason for adoption and a multi-faceted perceived overall risk factor as the main obstacle. Accordingly, when usefulness is less apparent, enjoyment may be a key factor for the adoption of mobile ICT for health promotion. Keywords User adoption. Motivation. Perceived risk. Mobile ICT. Health promotion Responsible editor: Hans-Dieter Zimmermann M. Cocosila (*) Faculty of Business, Athabasca University, 1 University Drive, Athabasca, Alberta T9S 3A3, Canada e-mail: mihailc@athabascau.ca URL: http://business.athabascau.ca N. Archer DeGroote School of Business, McMaster University, Hamilton, Ontario, Canada e-mail: archer@mcmaster.ca JEL Classification M150 Introduction Although relatively new, the use of mobile information and communication technology (ICT) services in healthcare has gained additional dimensions in recent years. An interesting approach that takes advantage of these services is to deliver targeted health promotion education through cell phones. Keeping people healthy through customized education has obvious benefits for individuals, healthcare systems, and society. One of the approaches considered in several settings worldwide has been the use of the most popular mobile ICT application, short message service (SMS), for sending people various health reminders on their cell phones. The literature has reported early applications where SMS has been used in a broad range of interventions extending from health promotion to disease treatment: preventive vaccinations (Anna et al. 2004), reminding about upcoming medical appointments (Downer et al. 2005) or taking medications for long-term self-management treatments (Bauer et al. 2003). Virtually all publications reporting such initiatives did not evaluate scientifically one key aspect of the picture: user perceptions regarding this innovative use of mobile ICT. Beyond potential medical and social gains, more-or-less complicated technical issues, and business models regarding the feasibility of such mobile applications, a major question remains: what would be the main factors driving people to use or not to use such services? This question is of utmost importance in today s context where there is an emphasis on trying to educate people to use health promotion interventions in order to improve health and reduce cost. A focus on

242 M. Cocosila, N. Archer consumers or acceptors (Mantzana et al. 2007) of healthcare services is critical since their perceptions will ultimately determine the success or failure of any such initiative. Further, investigating user acceptance of any ICT at an early stage is a factor contributing to its later overall success (Venkatesh et al. 2002). Few studies have investigated technology acceptance of mobile ICT in healthcare in general. Further, most technology acceptance study in healthcare has been limited to health care practitioners (Schaper and Pervan 2004). Moreover, there is a lack of research about patient acceptance of any technology in general. This contrasts, surprisingly, with the patient-centric view of today s healthcare. It is always important to remember that not the software but the human side of the implementation cycle... will block progress in seeing that the delivered systems are used effectively (Keen 1991, p. 220). Consequently, this research tries to fill the void caused by the lack of a scientific perspective on user viewpoints regarding the use of mobile ICT for health promotion interventions, and adopts a virtual outpatient viewpoint which concurs with the contemporary concept of patient centered health provision. The scope of this work is to propose and validate scientifically a pro versus con adoption model. To accomplish this, an empirical study involving user exposure to a mobile ICT health application for 1 month was undertaken. After the trial period, user perceptions regarding reasons for and against using such an application were captured and analyzed. In the following two sections we present the theoretical background and research model and hypotheses. Next, we report on the experimental methodology and the main results. Finally, we present a discussion and draw some conclusions and ideas for future research. Theoretical background Investigating user reasons for adopting and using ICT has a long standing tradition in information systems (IS) research. Several models and theories have been proposed and successfully tested by various researchers (Venkatesh et al. 2003). One of these models is the motivational model (MM) validated in several prominent studies (Davis et al. 1992; Venkatesh et al. 2002). The motivational model posits that user behavioural intention (BI) to use an ICT application has two key determinants: intrinsic motivation (IM), representing inherent satisfaction or enjoyment associated with using the technology, and extrinsic motivation (EM) or perceived usefulness, which is related to attaining a goal or obtaining a reward through the use of the technology (Ryan and Deci 2000). Further, perceiving an activity as enjoyable has a positive impact on the user perceptions of usefulness regarding that activity (Deci and Ryan 1985). Although less used than the very popular Technology Acceptance Model (TAM) (Venkatesh et al. 2003), the motivational model is considered more suitable in investigating the innovative use of a new technology in a sensitive domain like healthcare. Whereas TAM is suitable for an established ICT application, the motivational model captures user perceptions regarding goals that are reachable through technology use, as well as the technology itself, and was considered more appropriate for a prospective use of the technology. The motivational model, as other popular models and theories in IS research, examines factors that would positively influence users to adopt a new ICT application. However, in recent years researchers have acknowledged the existence of other factors that influence users to resist or reject a new technology. To capture this obstruction phenomenon, researchers have used a multi-faceted perceived risk construct borrowed from consumer behaviour research (Pavlou 2003). More refined analyses reveal that these negative perceptions usually have several aspects: e.g., apprehension about possibly wasting time, fear about wasting money, etc. (Lim 2003). In line with this perspective, perceived risk has become an increasingly popular construct in IS, being mostly associated with online shopping and its intangibility. To capture the above issues, especially when referring to risk perceptions in the e-commerce context, IS studies have added other facets to those traditional in consumer behaviour research: fear of threat to privacy, apprehension of buying online, etc. (Jarvenpaa and Todd 1996). Consequently we considered that, in the innovative use of a new mobile ICT for health promotion, there may be motivational factors favouring user adoption, and risk perceptions disfavouring it. Accordingly, we pose the following research questions: What is the influence of a multi-dimensional perceived risk on the multi-dimensional motivation to use mobile ICT for a health promotion intervention? What is the combined influence of multi-dimensional risk perception and motivation on the intention to use mobile ICT for a health promotion intervention? How appropriate is the theoretical model we propose for explaining the intention to use mobile ICT for a health promotion intervention? Research model and hypotheses We propose a theoretical model that would reflect positive and negative user perceptions regarding technology adoption. These two opposite points of view are best captured

Adoption of mobile ICT for health promotion 243 through the integration of multi-sided perceived risk into the motivational model. A significant body of consumer behavior research (Jacoby and Kaplan 1972; Stone and Grønhaug 1993; Stone and Mason 1995; Lim 2003) and more recent information systems research (Featherman and Pavlou 2003; Cocosila et al. 2009) consistently showed that individuals risk perceptions have generally six dimensions: 1. Psychological risk (i.e., fear of making a bad choice); 2. Financial risk (i.e., fear of wasting money); 3. Time risk (i.e., fear of wasting time); 4. Performance risk (i.e., fear of the product/service not working properly); 5. Social risk (i.e., fear of negative opinions from significant other individuals on the product/service); 6. Physical risk (i.e., fear of the product/service being a threat to the individual s health). Previous research seeking for relevant and, at the same time, parsimonious theoretical models has usually included only some of the above facets, depending on the research context (Laroche et al. 2004; Cocosila et al. 2009). As the framework of this research regards the use of SMS for a new purpose by individuals already accustomed with SMS on cell phones, it was considered that performance risk and time risk would be insignificant (since individuals are already SMS users). Social risk was not expected to have a significant effect either since the usual utilization of SMS is in a social context (i.e., exchange messages with friends). Previous research demonstrated that perceived physical risk caused by cell phones is not an issue for users (Cocosila et al. 2007). In these conditions, from the six facets of perceived risk only two were considered relevant for the context of this research: perceived psychological risk (i.e., doubt about subscribing for a new and unknown mobile service) and perceived financial risk (i.e., fear of wasting money to subscribe for an unnecessary service). In addition to these, an additional facet of risk called perceived privacy risk was considered. This risk expresses the fear of revealing private information online to third parties and has gained increased popularity in research on e-commerce applications (Featherman and Pavlou 2003; Shareef et al. 2008). Therefore we captured risk perceptions through a second-order perceived overall risk construct formed by the individual risk dimensions. Since risk perception is context dependent (Conchar et al. 2004), the individual risk dimensions forming the overall risk may vary from one case to another. However, even if the relative importance of the diverse risk dimensions are not identical, their aggregate influence on the overall risk is expected to be about the same (Stone and Grønhaug 1993). All these risks express real or virtual threats that users would perceive when subscribing for and using a mobile service for health promotion. Consequently, we hypothesized that: H1-1: Perceived financial risk is positively associated with perceived overall risk of using mobile ICT for a health promotion intervention. H1-2: Perceived psychological risk is positively associated with perceived overall risk of using mobile ICT for a health promotion intervention. H1-3: Perceived privacy risk is positively associated with perceived overall risk of using mobile ICT for a health promotion intervention. This work proposes as an innovative theoretical approach the integration of perceived overall risk into the motivational model. This is similar to the work of Featherman and Pavlou (2003) who integrated a perceived risk multidimensional construct into the technology acceptance model (TAM). Extrinsic motivation and perceived usefulness represent a single construct which captures the performance expectancy of an activity (Venkatesh et al. 2002). Previous studies showed empirically that perceiving a service as riskier reduces its utilitarian value (Pavlou 2003; van der Heijden et al. 2005). Consequently, perceiving a mobile service as risky for various reasons (e.g., too expensive, or stressful) is likely to decrease the extrinsic motivation for using that service. Therefore, we hypothesized that: H2: Perceived overall risk will have a negative effect on extrinsic motivation of using mobile ICT for a health promotion intervention. Similarly to the influence of perceived ease of use on perceived usefulness that has been consistently demonstrated by TAM studies, evidence show that there is also a relationship between intrinsic motivation (IM) and extrinsic motivation (EM) (Davis et al. 1992). Empirical studies suggest that there is a positive link from IM to perceived usefulness: increasing the enjoyment regarding the fulfilment of a task results in higher quality and productivity. Therefore, it is hypothesized that: H3: Intrinsic motivation is positively associated with extrinsic motivation of using mobile ICT for a health promotion intervention. Previous studies in technology acceptance have established that in most cases perceived usefulness is the key determinant of the behavioural intention (BI) to use a technology, while perceived ease of use and enjoyment are secondary antecedents (Igbaria 1993; Venkatesh 1999). However, as some studies report, the more immersive, hedonic aspects of new media play at least an equal role (Childers et al. 2001). Therefore, fun s positive effect on the use of a technology should not be underestimated. Higher enjoyment leads to higher acceptance (even for unproduc-

244 M. Cocosila, N. Archer tive systems), but enjoyment has been seen to have an increased effect on acceptance for systems that are also high in perceived usefulness. An increase in enjoyability increases acceptability but has less of an effect on the acceptance of useless systems (Davis et al. 1992). From the opposite side, in general it is habitual for consumers to adopt strategies to decrease risk such as relying on product quality and performance features before making a decision to purchase (Shimp and Bearden 1982). In particular, for studies in the IS field, perceived risk has been found to play an important role by adversely influencing user intent to adopt ICT-related services (Doolin et al. 2005). Consequently, we hypothesized that: H4-1: Intrinsic motivation will have a positive effect on the behavioural intention of using mobile ICT for a health promotion intervention. H4-2: Extrinsic motivation will have a positive effect on the behavioural intention of using mobile ICT for a health promotion intervention. H4-3: Perceived overall risk will have a negative effect on the behavioural intention of using mobile ICT for a health promotion intervention. The theoretical model and hypotheses proposed by this research are described in Fig. 1. Methodology The experiment described in this study consisted of using SMS on cell phones for a health promotion intervention: reminding people to take a vitamin C pill daily for preventing flu and cold. The study involved 52 participants Note: PFR - perceived financial risk, PYR - perceived psychological risk, PPR - perceived privacy risk, POR - perceived overall risk, IM - intrinsic motivation, EM - extrinsic motivation, BI - behavioural intention to use Fig. 1 Theoretical model of user acceptance of mobile ICT for health promotion. PFR perceived financial risk, PYR perceived psychological risk, PPR perceived privacy risk, POR perceived overall risk, IM intrinsic motivation, EM extrinsic motivation, BI behavioural intention to use recruited online from the entire student and non-student population of a North-American university. This experiment was part of a larger project conducted in that setting. Inclusion criteria required participants to be at least 18 years old and have a cell phone with SMS capabilities. After being presented the benefits of vitamin C from a trusted online source (British Broadcasting Corporation Web site), all participants were required to take one vitamin C pill per day for preventive reasons for 1 month. During the experiment, participants received SMS reminders on their cell phones (with fresh content, including brief jokes and interesting related information on how to stay healthy) about taking the vitamins. Reminders were provided by an automated wireless application at an average rate of one per day, at random times and in an informal language, as if sent by a virtual friend. Data were collected through surveys at baseline (demographics and prior experience with cell phones and SMS) and endpoint of the experiment (perceptions after using the mobile technology effectively). Survey questions used 7-point Likerttype scales and were adapted from previously validated questionnaires. Perceived risk survey questions were adapted from questionnaires in consumer behavior research (Stone and Grønhaug 1993; Stone and Mason 1995) and information systems research (Featherman and Pavlou 2003) while extrinsic and intrinsic motivation questions were adapted from relevant IS studies (Venkatesh and Davis 2000; Venkatesh et al. 2002). Appendix A shows survey questions pertaining to the theoretical model, and their sources. Results Fifty one participants completed the 1-month trial. One case was discarded after a visual inspection detected uncommon patterns in the responses. The remaining 50 cases were used in the subsequent data analysis. Demographic analyses indicated that participants in the experiment were generally young people (average age 24 years), balanced on gender (56% females) and each with significant experience with the technology under scrutiny (average of 4.1 years cell phone experience and average of 2.6 years SMS experience). The theoretical model was analyzed using the Partial Least Squares (PLS) modeling technique. This is suitable for small sample studies when the purpose of the research is exploratory (Chin 1998), also when including formative indicators (Thomas et al. 2005). The 50-case valid sample was larger than the minimum sample of 40 participants required by PLS for the theoretical model used in this research (Jarvenpaa et al. 2004). The PLS analysis comprised two steps: evaluation of the measurement model, followed by the evaluation of the structural model (Bontis 1998; Jarvenpaa et al. 2004). Operationalization of per-

Adoption of mobile ICT for health promotion 245 ceived overall risk was done through the repeated indicators approach (the hierarchical component model) by measuring the second order factor through the indicators of the first order perceived risk facets (Chin 1997; Lohmoller 1989). Measurement model A first analysis assessed construct reliability by calculating Cronbach s alpha in SPSS 17.0. Then PLS-Graph 3.00 with bootstrap was run, following closely the guidelines of Gefen and Straub (2005). All the measures were in the expected range except one item of perceived financial risk (PFR2). This item, displaying very low factor loading and item-total correlation, was dropped from subsequent analyses and the PLS program was re-run. Statistics of the resulting measurement model are presented in Table 1. Cronbach s alpha and internal consistency values were above 0.7 and average variance extracted (AVE) were above 0.5 for all but one of the constructs, indicating adequate reliability and convergent validity (Fornell and Larcker 1981). This conclusion was substantiated by high item loading on the factors (above 0.7 for all items), itemtotal correlations above 0.3 and by the high t-values for the item loadings. Cronbach s alpha for PFR with three items was 0.710 but after dropping PFR2, as suggested by the PLS analysis, it became 0.679. However, the AVE is 0.737 and the internal consistency 0.817. Therefore, the two-item scale for PFR was considered acceptable from the reliability point of view. Consequently, all items in Table 1 were retained in the measurement model (Bontis 1998; Nunnally 1978). A test for discriminant validity is to compare the measurement model construct correlations to the square root of the average variance extracted. Results are shown in Table 2: the diagonal in the table shows the square roots of AVE, and off-diagonals are the correlations obtained, as described by Gefen and Straub (2005). As the diagonal elements in Table 2 are substantially larger than the corresponding off-diagonal elements, there is some confidence in the model having appropriate discriminant validity (Compeau et al. 1999). In conclusion, after the above analyses, we considered reliability, and convergent and discriminant validity conditions were met, so we could proceed to the structural analysis of the model. Structural model Path coefficients and t-values were obtained by running PLS- Graph 3.00 with bootstrap. Results are indicated in Fig. 2. A visual inspection of Fig. 2 shows that most of the hypotheses we made are supported: & & & psychological risk and privacy risk have a strong significant influence on the overall risk (H1-2 and H1-3); intrinsic motivation positively influences extrinsic motivation (H3); and, intrinsic motivation affects positively (H4-1) andper- ceived overall risk negatively (H4-3) the intention to use mobile technology for a preventive health intervention. Table 1 Measurement model statistics Item Mean Standard deviation Factor loading Error Item-total correlations t-value Internal consistency (Cronbach s alpha; AVE) PFR1 3.59 1.49 0.952 0.175 0.517 5.43 0.817 (0.679; 0.737) PFR3 4.21 1.64 0.753 0.254 0.517 2.96 PPR1 2.38 1.36 0.911 0.025 0.768 36.31 0.984 (0.867; 0.796) PPR2 2.87 1.68 0.883 0.054 0.753 16.31 PPR3 2.94 1.62 0.880 0.036 0.739 24.02 PYR1 1.98 0.94 0.888 0.036 0.730 24.19 0.988 (0.891; 0.823) PYR2 1.85 1.03 0.924 0.027 0.836 33.95 PYR3 1.66 0.86 0.909 0.049 0.808 18.54 IM1 4.53 1.41 0.950 0.024 0.885 39.13 0.991 (0.936; 0.888) IM2 4.64 1.50 0.939 0.023 0.857 39.44 IM3 4.49 1.43 0.937 0.031 0.865 30.09 EM1 4.77 1.72 0.827 0.101 0.721 8.13 0.983 (0.923; 0.821) EM2 5.27 1.42 0.927 0.039 0.849 23.73 EM3 5.02 1.62 0.973 0.007 0.936 128.20 EM4 5.15 1.54 0.888 0.057 0.807 15.44 BI1 5.61 1.31 0.986 0.007 0.944 140.25 0.997 (0.971; 0.972) BI2 5.90 1.40 0.985 0.007 0.944 143.69 PFR perceived financial risk, PPR perceived privacy risk, PYR perceived psychological risk, IM intrinsic motivation, EM extrinsic motivation, BI behavioural intention to use, 1...4 scale items

246 M. Cocosila, N. Archer Table 2 Correlations and square root of Average Variance Extracted for first order constructs Hypotheses that were not supported include: H1-1 (there is no significant effect from perceived financial risk to the overall risk), H2 (the influence of perceived overall risk on extrinsic motivation is not significant) and H4-2 (it cannot be supported that extrinsic motivation has a positive effect on behavioural intention). In addition, the theoretical model proposed by this research displayed a reasonably high explanatory power. The variance explained by the endogenous constructs was at levels usually met in information systems studies (Moon and Kim 2001): R 2 =0.335 for extrinsic motivation and R 2 = 0.487 for behavioural intention. Effect size PFR PPR PYR IM EM BI PFR 0.850 PPR 0.175 0.891 PYR 0.249 0.606** 0.907 IM 0.283* 0.034 0.189 0.942 EM 0.378** 0.021 0.143 0.554** 0.906 BI 0.598** 0.092 0.200 0.421** 0.602** 0.986 PFR perceived financial risk, PPR perceived privacy risk, PYR perceived psychological risk, IM intrinsic motivation, EM extrinsic motivation, BI behavioural intention to use; Significance levels: *=0.05; **=0.01 A refinement of the R 2 analysis is to look at the individual effect of the independent variables on the dependent variables, thus giving a better picture of the predictive power of the theoretical model. Predictive power (or impact) of the exogenous variables on the endogenous variables can be assessed by a formula introduced by Chin (1998): f 2 =(R 2 included R 2 excluded)/(1 R 2 included) where R 2 included is the R 2 value when the independent construct is included in the model and R 2 excluded when this construct is not included in the model. Step values for R 2 recommended by Chin (1998) based on Cohen s (1988) work are: 0.02 for small, 0.15 for medium, and 0.35 for large effects of the predictor at a structural level. This analysis is useful in determining if a single independent variable has a significant influence on the predictive power of the dependent variable. Predictive power was calculated only for Behavioural Intention (this being the only endogenous construct with at least two incoming significant paths) by removing one significant link in turn and running the model in order to record the new R 2 values. The results obtained are presented in Table 3. The calculated values for f 2 show that intrinsic motivation (IM) is more important than perceived risk (POR) for the behavioral intention (BI) to adopt mobile health for preventive interventions. This is confirmed by the path coefficients and significance levels for the two paths (POR- BI and IM-BI) in the model analysis results represented in Fig. 2. Control variable analyses Demographic variables were tested as possible influencers of the model. Age, gender, and SMS experience were added to the model and PLS was re-run. Virtually no difference between the R 2 values for the endogenous constructs (EM and BI) with the control variables and without these was detected. Since technology could do little to support an activity that people do not find important, an additional factor was tested as a possible influencer. This was called attitude toward adherence (i.e., the health activity targeted by the mobile service in this case). Attitude toward adherence questions described in Appendix A were adapted from the Beliefs about Medicines Questionnaire developed by Horne Table 3 Size effects on the behavioural intention Note: PFR - perceived financial risk, PYR - perceived psychological risk, PPR - perceived privacy risk, POR - perceived overall risk, IM - intrinsic motivation, EM - extrinsic motivation, BI - behavioural intention to use; Significance levels: * = 0.05; ** = 0.01; *** = 0.001 Fig. 2 Structural evaluation of the theoretical model. PFR perceived financial risk, PYR perceived psychological risk, PPR perceived privacy risk, POR perceived overall risk, IM intrinsic motivation, EM extrinsic motivation, BI behavioural intention to use; Significance levels: *=0.05; **=0.01; ***=0.001 Endogenous construct Exogenous constructs 2 BI R included 0.487 POR IM R excluded 2 0.376 0.357 f 2 0.10 0.24 Effect small medium

Adoption of mobile ICT for health promotion 247 et al. (1999) and improved and revalidated by Horne et al. (2004). The Cronbach s alpha for this construct was 0.797 and AVE 0.518, thus demonstrating satisfactory reliability. When re-running the model with this construct we noticed slight increases of R 2 for both EM and BI. As loadings and AVE values for the constructs of the initial model did not change when attitude toward adherence was added, we concluded that the changes noticed were structural. An analysis of the path coefficients from the attitude toward adherence to the two endogenous variables showed a positive and significant influence (Table 4). Discussion and conclusions The purpose of this research has been to investigate user acceptance of mobile ICT for a health promotion intervention. We proposed a theoretical model blending both factors favouring adoption (i.e., motivations) and factors disfavouring it (i.e., perceived risk). The model was tested by surveying participants in an empirical investigation where they were exposed to the technology for 1 month. The first research question we asked was: What is the influence of a multi-dimensional perceived risk on the multi-dimensional motivation to use mobile ICT for a health promotion intervention? In response to this question we found a negative influence of the second-order perceived risk construct on the extrinsic motivation, but this was not significant at the 0.05 level (it was marginally significant at the 0.1 level). A justification of this finding could be provided when examining the response to the second research question we asked: What is the combined influence of multi-dimensional risk perception and motivation on the intention to use mobile ICT for a health promotion intervention? We found that perceived overall risk was a significant deterrent to intention to use the technology. This is similar to consumer behaviour research where a risk perception negatively affects consumer intent to complete a purchase (Laroche et al. 2004). We also found intrinsic motivation to influence positively both extrinsic motivation and user intent to use the technology, similar to previous IS studies (Igbaria et al. 1995). However, extrinsic motivation was not shown to be a significant antecedent of intention to use the technology. Table 4 Path coefficients from attitude toward adherence to the endogenous constructs in the controlled model Control variable EM BI Attitude toward adherence Path coefficient 0.246 0.222 t-value 1.922 2.097 Significance 0.06 0.04 An explanation of the non-significant role played by extrinsic motivation is that participants were young and normally healthy people who did not perceive the usefulness of a mobile application to support a preventive intervention. Previous research in healthcare has demonstrated that adherence to preventive treatments is the most difficult to ensure for people who do not perceive their necessity (Anna et al. 2004). Nonetheless, even without the contribution of extrinsic motivation, intrinsic motivation was found to be a sufficient reason for adoption. This is similar to some IS studies that found enjoyment more important than usefulness perception (van der Heijden 2004). This suggests that, when developing mobile ICT applications for health interventions in practice, making these applications attractive may be a sufficient reason for adoption when future goals are not necessarily apparent. We also found that attitude toward adherence played an important role that deserves more attention in future research. That is, our analysis showed that if people tend to have a positive attitude toward the health activity supported by the technology, they also tend to have a favourable perception of the technology itself. The last research question we asked was: How appropriate is the theoretical model we propose for explaining the intention to use mobile ICT for a health promotion intervention? We found that the majority of the hypothesized model paths were significant, with reasonably high coefficients. We also found moderately high values for the variance explained for the endogenous constructs. According to the literature, these indicate a good model: it has significant relationships between the constructs, and high R 2 values (Bontis et al. 2000). The second-order perceived overall risk construct used three first-order risk facets: financial, psychological, and privacy. As hypothesized, perceived psychological risk and perceived privacy risk had strong and significant influences on the overall risk construct (paths of 0.556 and 0.525, respectively, both significant at the 0.001 level). Consistent with similar research (Turel et al. 2007), since two of the three components of the overall risk are significant in the overall risk variable, this second-order construct could be termed as appropriate in capturing the effects of first-order risk facets. While psychological and privacy risk showed a significant and strong influence on the overall risk, financial risk did not show such an effect. A possible explanation to this could be found in consumer behaviour studies which showed that perceived financial risk is so strong in some cases that it plays a role separate from the other risk facets (Stone and Grønhaug 1993). An alternate explanation is that the possibility of spending additional money for a mobile service was not an issue for people already accustomed to cell phone use. Despite these issues,

248 M. Cocosila, N. Archer the model was considered appropriate due to its parsimony and to the possibility of generalization: depending on the context, various risk facets could be added or removed but the general layout would be the same. This study also had some limitations, as do virtually all empirical studies in information systems. The relatively small sample size imposed by feasibility was the most important issue. However, the size was 20% more than the minimal number required by PLS techniques. As in many such empirical studies participants self-selected and were generally young people. However, although they were recruited through a university website, the resulting sample contained a combination of students and non-students. Further, there is no reason to believe this sample would differ from the general youthful population in terms of the need for preventive health behaviours. Further, it is always important to look at the future when developing a new information and communication technology; accordingly, this research investigated the perceptions of tomorrow s potential users of mobile healthcare interventions. So, from this point of view the sample could be considered as representative for the health prevention needs of future adult generations that are currently users of SMS on cell phones. Overall, this research did not involve more limitations than those usually encountered in IS studies. This research has implications for both theoretical and practical knowledge. From a theoretical point of view it integrates a multi-dimensional perceived risk latent variable into the motivational model thus proposing and validating a pro-con ICT adoption model. While motivation is a factor pro adoption, perceived risk is a deterrent ( con ). By capturing these two opposing broad views on technology adoption, the model has a high degree of generalizability. Thus the combination of extrinsic motivation-intrinsic motivation may vary in relative weight but it will always have a favourable role in the intention to use the technology in any context. Also, the relative importance of the perceived risk facets may differ from a technology use situation to other situations, but their overall effect would express aggregated user resistance to adoption for various reasons. From a practical point of view, this research shows that intrinsic motivation might be a key ingredient in the success of some ICT applications in a healthcare context. Like covering some pills in an easy-to-swallow coat, ICT applications in general should be appealing to individuals when the usefulness is less apparent (e.g., for a health promotion activity). Concordant with previous research showing the key role of psychological risk (Stone and Grønhaug 1993; Cocosila et al. 2009), we too found this risk to be an important obstacle to adoption. Therefore, developers and marketers of ICT applications for health promotion (or for well individuals, in general) should try to mitigate the risk by targeting prospective user doubts about the necessity of such services. Our study is also important because by proving how and why a simple but solid application may work might also encourage the use of mobile information technology by other stakeholders in healthcare (e.g., home care nurses, primary care physicians, etc.). Future research should look more carefully at the integration of the perceived risk into the motivational model. The role of financial risk should be investigated in detail. Another possibility is to consider overall risk as a first-order construct having as antecedents the risk facets as suggested by certain consumer behaviour studies (Stone and Grønhaug 1993). Tests for various riskfacetsincludedinthemodelaswellasforseveral technology contexts should be performed. Obviously, future tests with larger samples and for different user groups would help to assess the validity of the research model proposed here. This study was an exploratory investigation of factors favouring or disfavouring the adoption of mobile ICT for health promotion interventions. It is hoped that future investigations will bring more contributions to this type of research, by considering both perceived obstacles and factors favouring the adoption of new information and communication technologies. Appendix A: Theoretical Model Questionnaire All items were measured with 7-point Likert type scales having as anchors Strongly Agree and Strongly Disagree. TMT in the below questions means Text Messaging Telehealth and denotes the prototype application using SMS on cell phones for health promotion that was tested in the study. Perceived Financial Risk (adapted from Stone and Grønhaug (1993) and Stone and Mason (1995)) Signing up for TMT would be a poor way to spend my money. I would be concerned about how much I would pay if I subscribed to TMT. If I subscribed to TMT, I would be concerned that I would not get my money s worth. Perceived Privacy Risk (adapted from Featherman and Pavlou (2003)) My use of TMT would cause me to lose control over the privacy of my information. Signing up for and using TMT would lead to a loss of privacy for me because my personal information could be used without my knowledge. Internet hackers (criminals) might take control of my information if I used TMT.

Adoption of mobile ICT for health promotion 249 Perceived Psychological Risk (adapted from Stone and Grønhaug (1993) and Stone and Mason (1995)) The thought of signing up for TMT makes me feel uncomfortable. The thought of signing up for TMT gives me an unwanted feeling of anxiety. The thought of signing up for TMT causes me to experience unnecessary tension. Extrinsic Motivation (adapted from Venkatesh et al. (2002)) Using TMT helped me to take the daily vitamin C pill at proper time. Using TMT helped me to not forget about the daily vitamin C. Using TMT helped me to take the vitamin C every day. I found TMT to be useful in reminding me to take my vitamin C daily. Intrinsic Motivation (adapted from Venkatesh et al. (2002)) I found TMT to be enjoyable. The actual process of using TMT was pleasant. I had fun using TMT. Behavioural Intention (adapted from Venkatesh et al. (2002)) Assuming I had access to TMT, I intend to use it. Given that I had access to TMT, I predict that I would use it. Attitude Toward Adherence (adapted from Horne et al. (2004)) Without vitamin C doctors would be less able to cure people for colds and flu. Taking vitamin C helps many people to be healthy. Taking vitamin C helps many people to prevent or recover faster from colds and flu. The benefits of taking vitamin C outweigh the risks. References Anna, V., Jose-Maria, B., Maria-Teresa, D., Caterina, G., Consolacion, D., Dulcis, S., et al. (2004). The role of mobile phones in improving vaccination rates in travelers. Preventive Medicine, 38, 503 509. Bauer, S., Percevic, R., Okon, E., Meermann, R., & Kordy, H. (2003). Use of text messaging in the aftercare of patients with bulimia nervosa. European Eating Disorders Review, 11, 279 290. Bontis, N. (1998). Intellectual capital: an exploratory study that develops measures and models. Management Decision, 36(2), 63 76. Bontis, N., Keow, W. C. C., & Richardson, S. (2000). Intellectual capital and business performance in Malaysian industries. Journal of Intellectual Capital, 1(1), 85 100. Childers, T. L., Carr, C. L., Peck, J., & Carson, S. (2001). Hedonic and utilitarian motivations for online retail shopping behavior. Journal of Retailing, 77, 511 535. Chin, W. W. (1997). Overview of the PLS Method, University of Houston. http://disc-nt.cba.uh.edu/chin/plsintro.htm, accessed: February 10, 2009. Chin, W. W. (1998). The Partial Least Squares approach for structural equation modelling. In A. Marcoulides (Ed.), Modern methods for business research (pp. 295 336). Mahwa: Lawrence Erlbaum Associates. Cocosila, M., Turel, O., Archer, N., & Yuan, Y. (2007). Perceived health risk of 3G cell phones: do users care? Communications of the ACM, 50(6), 89 92. Cocosila, M., Archer, N., & Yuan, Y. (2009). Early investigation of new information technology acceptance: a perceived risk motivation model. Communications of the Association for Information Systems, 25(30). http://aisel.aisnet.org/cais/vol25/ iss21/30, accessed: February 19, 2010. Cohen, J. (1988). Statistical power analysis for the behavioral science (2nd ed.). Hillsdale: Lawrence Erlbaum Associates. Compeau, D., Higgins, C., & Huff, S. (1999). Social cognitive theory and individual reactions to computing technology: a longitudinal study. MIS Quarterly, 23(2), 145 158. Conchar, M. P., Zinkhan, G. M., Peters, C., & Olavarrieta, S. (2004). An integrated framework for the conceptualization of consumers perceived-risk processing. Journal of the Academy of Marketing Science, 32(4), 418 436. Davis, F., Bagozzi, R., & Warshaw, P. (1992). Extrinsic and intrinsic motivation to use computers in the workplace. Journal of Applied Social Psychology, 22(14), 1111 1132. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior. New York: Plenum Press. Doolin, B., Dillon, S., Thompson, F., & Corner, J. L. (2005). Perceived risk, the Internet shopping experience and online purchasing behavior: a New Zealand perspective. Journal of Global Information Management, 13(2), 66 88. Downer, S. R., Meara, J. G., & Da Costa, A. C. (2005). Use of SMS text messaging to improve outpatient attendance. The Medical Journal of Australia, 183(7), 366 368. Featherman, M., & Pavlou, P. (2003). Predicting e-services adoption: a perceived risk facets perspective. International Journal of Human Computer Studies, 59(1), 451 474. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39 50. Gefen, D., & Straub, D. (2005). A practical guide to factorial validity using PLS-Graph: tutorial and annotated example. Communications of the Association for Information Systems, 16, 91 109. Horne, R., Graupnera, L., Frosta, S., Weinmanb, J., Wright, S. M., & Hankins, M. (2004). Medicine in a multi-cultural society: the effect of cultural background on beliefs about medications. Social Science & Medicine, 59(6), 1307 1313. Horne, R., Weinman, J., & Hankins, M. (1999). The Beliefs about Medicines Questionnaire: the development and evaluation of a new method for assessing the cognitive representation of medication. Psychology & Health, 14(1), 1 24. Igbaria, M. (1993). User acceptance of microcomputer technology: an empirical test. OMEGA: International Journal of Management Science, 21(1), 73 90. Igbaria, M., Iivari, J., & Maragahh, H. (1995). Why do individuals use computer technology? A Finnish case study. Information Management, 29(5), 227 238. Jacoby, J., & Kaplan, L. B. (1972). The components of perceived risk. Proceedings of the 3rd Annual Conference of the Association for Consumer Research. Jarvenpaa, S. L., & Todd, P. A. (1996). Consumer reactions to electronic shopping on the World Wide Web. International Journal of Electronic Commerce, 1(2), 59 88. Jarvenpaa, S. L., Shaw, T. R., & Staples, D. S. (2004). Toward contextualized theories of trust: the role of trust in global virtual teams. Information Systems Research, 15(3), 250 267.

250 M. Cocosila, N. Archer Keen, P. (1991). Shaping the future: Business design through information technology. Boston: Harvard Business School Press. Laroche, M., McDougall, G. H. G., Bergeron, J., & Yang, Z. (2004). Exploring how intangibility affects perceived risk. Journal of Service Research, 6(4), 373 389. Lim, N. (2003). Consumers perceived risk: sources versus consequences. Electronic Commerce Research and Applications, 2, 216 228. Lohmoller, J.-B. (1989). Latent variable path modelling with Partial Least Squares. Heidelberg: Physica-Verlag. Mantzana, V., Themistocleous, M., Irani, Z., & Vincenzo, M. (2007). Identifying healthcare actors involved in the adoption of information systems. European Journal of Information Systems, 16(1), 91 102. Moon, J.-W., & Kim, Y.-G. (2001). Extending the TAM for a World- Wide-Web context. Information Management, 38(4), 217 230. Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill. Pavlou, P. A. (2003). Consumer acceptance of electronic commerce Integrating trust and risk with the Technology Acceptance Model. International Journal of Electronic Commerce, 7(2), 101 134. Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54 67. Schaper, L., & Pervan, G. (2004). A model of information and communication technology acceptance and utilisation by occupational therapists. Proceedings of The IFIP TC8/WG8.3 International Conference, Prato, Italy. Shareef, M. A., Kumar, U., & Kumar, V. (2008). Role of different electronic-commerce (EC) quality factors on purchase decision: a developing country perspective. Journal of Electronic Commerce Research, 9(2), 92 112. Shimp, T. A., & Bearden, W. O. (1982). Warranty and other extrinsic cue effects on consumers risk perceptions. Journal of Consumer Research, 9(1), 38 47. Stone, R. N., & Grønhaug, K. (1993). Perceived risk: further considerations for the marketing discipline. European Journal of Marketing, 27(3), 39 50. Stone, R. N., & Mason, B. J. (1995). Attitude and risk: exploring the relationship. Psychology & Marketing, 12(2), 135 153. Thomas, R. D., Lu, I. R. R., & Cedzynski, M. (2005). Partial Least Squares: A critical review and a potential alternative. Proceedings of Administrative Sciences Association of Canada (ASAC), Toronto, Ontario, Canada. Turel, O., Serenko, A., & Bontis, N. (2007). User acceptance of wireless short messaging services: deconstructing perceived value. Information Management, 44(1), 63 73. van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly, 28(4), 695 704. van der Heijden, H., Ogertschnig, M., & van der Gaast, L. (2005). Effects of context relevance and perceived risk on user acceptance of mobile information services. Proceedings of the Thirteenth European Conference on Information Systems, Regensburg, Germany. Venkatesh, V. (1999). Creation of favorable user perceptions: exploring the role of intrinsic motivation. MIS Quarterly, 23(2), 239 260. Venkatesh, V., & Davis, F. (2000). A theoretical extension of the technology acceptance model: four longitudinal field studies. Management Science, 46(2), 186 204. Venkatesh, V., Speier, C., & Morris, M. G. (2002). User acceptance enablers in individual decision making about technology: toward an integrated model. Decision Sciences, 33(2), 297 316. Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: toward a unified view. MIS Quarterly, 27(3), 425 478.