Research Report A model for the acceptance and use of mHealth in South Africa: A UTAUT and TTF perspective University of Witwatersrand: MCom Information Systems Student Name: Livhuwani Grace Mongwe Student Number: 1256754 Supervisor: Mitchell Hughes and Professor Ray Kekwaletswe 1 10 November 2023 Declaration of Originality I declare that this is my own unaided work and is, to the best of my knowledge and belief, original, except as acknowledged in the text. I have read and understood the Senate policy on plagiarism, and I am aware that plagiarism is the intentional or unintentional “failure to acknowledge the ideas or writing of another” or “presentation of the ideas or writing of another as one’s own”. In this context “other” means any other person including a student, academic, professional, published author or other resource such as the Internet. Failing to acknowledge the use of ideas of others constitutes an important breach of the values and conventions of the academic enterprise. I am aware that plagiarism offences will be dealt with in terms of the Senate policy and may be subject to disciplinary action. Signed: LG Mongwe Date: 29 March 2023 2 10 November 2023 Acknowledgements I wish to thank: • My husband, Arthur, for his unwavering support and encouragement when I would encounter challenges during my study period. • My mother, Dr Nkhensani Masekoa, whose love for education encouraged me to persevere. • My supervisor, Mitchell Hughes for his support, guidance, and insight during my studies. • My supervisor, Professor Ray Kekwaletswe, who encouraged me to get started even when I had lost my motivation. • God, for his guidance and enabling me to have access to the resources needed to complete my studies. 3 10 November 2023 Table of Contents Declaration of Originality ............................................................................................................................. 1 Acknowledgements ...................................................................................................................................... 2 List of Figures ................................................................................................................................................ 6 List of Tables ................................................................................................................................................. 7 Abstract ........................................................................................................................................................ 9 CHAPTER 1: INTRODUCTION TO THE STUDY ............................................................................................. 10 1.1 Background ....................................................................................................................................... 10 1.2 Problem Statement ......................................................................................................................... 11 1.3 Research Questions and Objectives ................................................................................................ 12 1.3.1 Research Questions ................................................................................................................... 12 1.3.2 Research Objectives .................................................................................................................. 12 1.4 Delimitations .................................................................................................................................... 13 CHAPTER 2: LITERATURE REVIEW ............................................................................................................. 14 2.1 mHealth Technology ........................................................................................................................ 14 2.2 mHealth Technology in South Africa ............................................................................................... 16 2.3 mHealth Acceptance and Use .......................................................................................................... 18 2.4 Theoretical Frameworks .................................................................................................................. 20 2.4.1 The Unified Theory of Technology Acceptance and Use of Technology (UTAUT) .................. 20 2.4.2 Task-Technology Fit (TTF).......................................................................................................... 23 2.5 Model of the study ........................................................................................................................... 25 2.5.1 Task Characteristics ................................................................................................................... 25 2.5.2 Technology Characteristics ....................................................................................................... 25 2.5.3 Task Technology Fit ................................................................................................................... 25 2.5.4 Performance expectancy .......................................................................................................... 26 2.5.5 Effort Expectancy ...................................................................................................................... 26 2.5.6 Facilitating Conditions ............................................................................................................... 26 2.5.7 Construct relationship among UTAUT and TTF ........................................................................ 26 CHAPTER 3: RESEARCH METHODOLOGY ................................................................................................... 30 3.2 Research Paradigm and Approach ................................................................................................... 30 3.2 Research Design ............................................................................................................................... 31 3.2.1 Types of research designs ......................................................................................................... 31 3.2.2 Survey Research ........................................................................................................................ 32 4 10 November 2023 3.3 Data Collection Methods ................................................................................................................. 33 3.3.1 Operationalization/ Measurement and Instrument Construction .......................................... 33 3.3.2 Sampling and Respondents ....................................................................................................... 34 3.3.3 Pilot testing................................................................................................................................ 35 3.3.4 Questionnaire administration .................................................................................................. 35 3.4 Data Analysis Strategy ..................................................................................................................... 35 3.5 Ethical Considerations ...................................................................................................................... 36 CHAPTER 4: PRESENTATION OF RESULTS .................................................................................................. 37 4.1 Data Screening .................................................................................................................................. 37 4.1.1 Missing Data .............................................................................................................................. 37 4.2 Demographic Analysis ...................................................................................................................... 37 4.3 Validity and Reliability ..................................................................................................................... 40 4.4 Normality Tests ................................................................................................................................ 45 4.5 Correlation Analysis ......................................................................................................................... 46 4.6 Multiple Regression ......................................................................................................................... 52 4.7 Regression Analysis .......................................................................................................................... 53 4.7.1 The Regression analysis for the TTF Model .............................................................................. 54 4.7.2 The Regression analysis for the UTAUT Model ........................................................................ 56 4.7.3 The Regression analysis for the Combined Model ................................................................... 59 4.7.4 Summary of Findings ................................................................................................................. 61 CHAPTER 5: DISCUSSION ............................................................................................................................ 62 5.1 Effects of TTF factors on mHealth adoption .................................................................................... 62 5.1.1 Task Characteristics ................................................................................................................... 62 5.1.2 Technology Characteristics ....................................................................................................... 63 5.1.3 Task-Technology Fit ................................................................................................................... 63 5.2 Effects of UTAUT factors on mHealth adoption .............................................................................. 64 5.2.1 Performance Expectancy .......................................................................................................... 64 5.2.2 Effort Expectancy ...................................................................................................................... 64 5.2.2 Facilitating Conditions ............................................................................................................... 65 CHAPTER 6: CONCLUSION ...................................................................................................................... 66 6.1 Summary of Findings ........................................................................................................................ 66 6.2 Implications for Academia ............................................................................................................... 67 6.3 Implications for Practice .................................................................................................................. 67 5 10 November 2023 6.4 Limitations ........................................................................................................................................ 68 6.5 Future Work...................................................................................................................................... 68 6.6 Conclusion ........................................................................................................................................ 69 REFERENCES ................................................................................................................................................ 70 Appendix A: Final Research Instrument .................................................................................................... 83 Appendix B: Ethics Approval Certificate .................................................................................................... 89 Appendix C: Participation Letter ................................................................................................................ 90 Appendix D: Turnitin Report ...................................................................................................................... 91 6 10 November 2023 List of Figures Figure 1: UTAUT Model (Venkatesh et al., 2003, p. 447)............................................................................ 21 Figure 2: TTF Model (Goodhue & Thompson, 1995) .................................................................................. 24 Figure 3: Proposed conceptual framework................................................................................................. 27 file:///C:/Users/grace/OneDrive/Documents/2023/School/2%20Nov%202023/LG%20Mongwe_Research%20Report%20(Chapter%201%20to%206)_02%20Nov%202023.docx%23_Toc149860797 file:///C:/Users/grace/OneDrive/Documents/2023/School/2%20Nov%202023/LG%20Mongwe_Research%20Report%20(Chapter%201%20to%206)_02%20Nov%202023.docx%23_Toc149860798 7 10 November 2023 List of Tables Table 1: Summary of Hypotheses ............................................................................................................... 28 Table 2: Key Attributes of Quantitative Research (Basias & Pollalis, 2018, p. 93) ..................................... 31 Table 3: Measurement Constructs .............................................................................................................. 33 Table 4: Gender Statistics ........................................................................................................................... 37 Table 5: Occupation .................................................................................................................................... 38 Table 6: Age Group ..................................................................................................................................... 38 Table 7: Level of Education ......................................................................................................................... 38 Table 8: Cross-tabulation of Smart phone ownership and Use of mHealth Application ............................ 39 Table 9: Cross-tabulation of Use of mHealth Application and Employment .............................................. 39 Table 10: Cross-tabulation of Gender and Use of mHealth Application ..................................................... 39 Table 11: Level of Education and Use of Mhealth Application ................................................................... 40 Table 12: Internal Consistency Values ........................................................................................................ 41 Table 13: Item-Total Correlation statistics .................................................................................................. 41 Table 14: KMO and Bartlett's Test .............................................................................................................. 42 Table 15: Principal Axis Factor Analysis ...................................................................................................... 42 Table 16: Rotated Factor Matrix ................................................................................................................. 44 Table 17: Skewness and Kurtosis ................................................................................................................ 45 Table 18: Tests of Normality ....................................................................................................................... 45 Table 19: Correlation of Task Characteristics and Task Technology Fit ...................................................... 46 Table 20: Correlation of Technology Fit and Task Technology Fit .............................................................. 46 Table 21: Correlation of Task Technology Fit and User Adoption of mHealth. .......................................... 47 Table 22: Correlation between Performance Expectancy and User Adoption of mHealth ........................ 47 Table 23: Correlation between Effort Expectancy and User Adoption of mHealth .................................... 48 Table 24: Correlation between Effort Expectancy and User Performance Expectancy ............................. 48 Table 25: Correlation of Facilitating Conditions and User Adoption of mHeath. ....................................... 49 Table 26: Correlation between Task Technology Fit and Performance Expectancy................................... 49 Table 27: Correlation between Technology Characteristics and Effort Expectancy ................................... 50 Table 28: Summary of Hypothesis Testing Results ..................................................................................... 50 Table 29: Multiple Regression Analysis Model Summary ........................................................................... 52 Table 30: ANOVA Results for Multiple Regression ..................................................................................... 52 Table 31: Multiple Regression Coefficients Results .................................................................................... 52 Table 32: Impact of Task Characteristics on the user adoption of mHealth Technology (Model Summary) .................................................................................................................................................................... 54 Table 33: Impact of Task Characteristics on the user adoption of mHealth Technology (ANOVA Results)54 Table 34: Impact of Task Characteristics on the user adoption of mHealth Technology (Coefficients Results) .................................................................................................................................................................... 55 Table 35: Impact Technology Characteristics on the user adoption of mHealth Technology (Model Summary) .................................................................................................................................................... 55 Table 36: Impact of Technology Characteristics on the user adoption of mHealth Technology (ANOVA) 55 Table 37: Impact of Technology Characteristics on the user adoption of mHealth Technology (Coefficients) .................................................................................................................................................................... 56 8 10 November 2023 Table 38: Impact of Performance Expectancy on the adoption of mHealth Technology (Model Summary) .................................................................................................................................................................... 56 Table 39: Impact of Performance Expectancy on the adoption of mHealth Technology (ANOVA) .......... 56 Table 40: Impact of Performance Expectancy on the adoption of mHealth Technology (Coefficients) .... 57 Table 41: The effect of Effort Expectancy on the adoption of mHealth Technology (Model Summary) ... 57 Table 42: The effect of Effort Expectancy on the adoption of mHealth Technology (ANOVA) .................. 57 Table 43: The effect of Effort Expectancy on the adoption of mHealth Technology (Coefficients ............ 58 Table 44: The effect of Facilitating Conditions on the adoption of mHealth Technology (Model Summary) .................................................................................................................................................................... 58 Table 45: The effect of Facilitating Conditions on the adoption of mHealth Technology (ANOVA) .......... 59 Table 46: The effect of Facilitating Conditions on the adoption of mHealth Technology (Coefficients) ... 59 Table 47: The effect of Task Technology Fit on the adoption of mHealth Technology (Model Summary) 59 Table 48: The effect of Task Technology Fit on the adoption of mHealth Technology (ANOVA) ............... 60 Table 49: The effect of Task Technology Fit on the adoption of mHealth Technology (Coefficients) ........ 60 Table 50: Results of Hypothesis testing (Summary) ................................................................................... 61 9 10 November 2023 Abstract Scaling up high impact community based mHealth interventions is one of the agenda items mentioned in the National Digital Health Strategy of South Africa for the period 2019 -2024. Although many mHealth interventions have been explored, many of them end up in the pilot phase and do not reach full implementation. A common theme which was found as a possible driver of scalability is designing an mHealth application that considers usability and acceptability by users. The purpose of this study was to synthesize a model for the acceptance and use of mHealth in the South African health sector. A positivist research approach was used to test the adoption factors using the Task-Technology Fit (TTF) and the Unified Theory of Acceptance and Use of Technology (UTAUT). Seven factors that could potentially impact the adoption of mHealth technology in South Africa were tested. The data for the study was collected through an online survey questionnaire which was shared through social media platforms. Results of this study were used to answer questions related to factors that have an impact on the adoption of mHealth applications in the health sector in South Africa. The study found that when adoption factors were combined into the UTAUT and TTF model, the only factor that was significant was facilitating conditions. The study findings in this regard were not consistent with other studies and it is therefore recommend that other scholars explore the reasons for these differences. The other factors were found to be significant when bivariate regression was used to compare the factors to the dependant variable of user acceptance and use of mHealth technology. The study further found that the combined model of Task Technology Fit has a positive impact on the adoption of mHealth technology in South Africa. The implication of the finding is that mHealth designers should build the functionalities of the innovation with the idea of making the task that the innovation supports easier to perform. Keywords: IT Adoption, mHealth adoption, mHealth, User Acceptance, TTF and UTAUT 10 10 November 2023 CHAPTER 1: INTRODUCTION TO THE STUDY 1.1 Background In 2018, 10% percent of the world’s population owned a smart phone. More than 41 000 health related applications were available on Google Play store in 2017 (Healthcare Apps Available Google Play 2021, n.d.). There were over 53 000 health related Apps on Google Play store and Apple store as at the first quarter of 2021 (Healthcare Apps Available Google Play 2021, n.d.). The difference between quarter 1 of 2017 and quarter 1 of 2021 shows a significant increase. The computational power of mobile phones has improved such that they can be leveraged to improve the health care sector (Balapour et al., 2019). These are just some of the indicators of the rapid growth of this technology. mHealth is defined as the use of technology to support health care services through the use of voice calls, SMS messaging systems, wireless transmission of data and mobile phone applications (Osei & Mashamba- Thompson, 2021). mHealth is a subset of telehealth and can also simply be defined as the use of portable technology to perform health related services and interventions (Lu et al., 2021). mHeath interventions include treatment support, diagnosing support, health monitoring, data accuracy and many other inventions that can support the health sector (Osei & Mashamba-Thompson, 2021). These interventions go beyond patient care. They also provide support to health care professionals and the health sector by facilitating the collection and sharing of data (Qudah & Luetsch, 2019). They also offer some level of independence and involvement on the part of the patient as they empower patients to be actively involved in their care and to receive information that provides actionable actions that they can take with regards to their health care (Qudah & Luetsch, 2019). Some examples of mHealth applications include healthy living applications, symptom checking applications, medication reminder applications etc. A study conducted in 2019 stated that even though the adoption of mHealth technology is widely discussed, there is a gap in terms of predicting the intention to accept and use this mHealth technology (Balapour et al., 2019). Many mHealth applications have been built in South Africa even though many of them remain as part of pilot studies (Ojo, 2018). The purpose of the study was to synthesise a model which can be used to explain the acceptance and use of mHealth in South Africa. The model was based on the Unified Theory of Technology Acceptance and Use of Technology (UTAUT) together with the Task-Technology Fit (TTF) model. This was with the aim of 11 10 November 2023 trying to mitigate one of the issues that hinder the usage of mHealth applications at a larger scale, building an application with usability in mind. The report will be broken down into six Chapters. Chapter 1 will provide insight into the problem that the research aimed to address and will introduce the research topic and chapter 2 will be a literature review to better understand topics that have been researched related to the research topic. Chapter 3 will be a discussion on the research methodology used to address the problem. Then chapter 4 will be a presentation of the results and Chapter 5 will be an in-depth discussion of the results and how to interpret them. Chapter 6 will be the conclusion which will summarize the findings, implications of study for academia and for the industry. Research limitations and future research will also be discussed in Chapter 6. 1.2 Problem Statement Even though many mHealth applications have been developed in South Africa most of them remain as part of pilot studies and are not fully accepted and used (Ojo, 2018; Latif et al., 2017). A study which focused on mHealth interventions in developing countries found that most mHealth interventions do not move beyond the pilot phase and further suggested that national health policies should include scalability and sustainability as important success measures for an mHealth intervention (Latif et al., 2017). Another study which aimed to summarize the challenges that must be overcome for an mHealth intervention to move from pilot phase to full-scale adoption found that one of the challenges that must be overcome is that mHealth applications must be built with usability in mind (Wallis et al., 2017). Scaling up high impact community based mHealth interventions is one of the agenda items mentioned in the National Digital Health Strategy of South Africa for the period 2019 -2024, indicating a further gap in the area of scaling up (NDoH,2019). Although there are many mHealth interventions which proved to have the potential to improve the health system and services provided, many of them end up at the pilot phase (Ojo, 2018). Many studies have mentioned that to enjoy the benefits that come with mHealth, policies that are aimed at resolving innovation challenges at scale should be explored. Most current studies related to mHealth technology are related to pilot projects (Said, 2022), thus motivating for further studies on the adoption of mHealth technology. The study aimed to synthesise a model for the acceptance and use of mHealth by individual users in South Africa which can be used to better understand the factors that can increase the chances of an mHealth innovation moving beyond the pilot phase. 12 10 November 2023 The factors impacting the adoption of mHealth in South Africa are still under-investigated and not fully understood, and this limits the understanding of factors that cause mHealth innovations to not move beyond the pilot phase. The problem statement provides a motivation for the development of the model and to further understand the factors that impact mHealth adoption. 1.3 Research Questions and Objectives The purpose of the study was to synthesise a model that can be used to explain the acceptance and use of mHealth at the individual user level. The model was based on the UTAUT and TTF theories of technology acceptance and use. 1.3.1 Research Questions The questions that the study aimed to answer were as follows: Primary Research Question RQ1: What are the factors that affect the use and acceptance of mHealth applications by individual adult users in the South African health sector? Secondary Research Questions RQ2: What are the technology characteristics that affect the use and acceptance of mHealth in the South African health sector by adults? RQ3: What are the task characteristics that affect the use and acceptance of mHealth in the South African health sector by adults? RQ4: Which behavioural constructs influence an individual user’s decision to use and accept a mHealth in the South African health sector by adults? 1.3.2 Research Objectives This section contains the goals and objectives of the study. They are listed below: RO1: To review existing literature to identify UTAUT and TTF factors that are dominant in the adoption mHealth technology. RO2: To identify the factors that affect the acceptance and use of mHealth in the South African health sector. RO3: To determine the significant factors in the UTAUT and TTF models that affect the adoption of mHealth in the South African health sector. RO4: To potentially conceptualize a model for the adoption of mHealth in the South African context. 13 10 November 2023 1.4 Delimitations 1. The study participants were limited to the province of Gauteng in South Africa. 2. The study was limited to adult participants who are over the age 18. 3. The study was not focused on a specific type of mobile health application but rather on any mobile health application in the health sector. 14 10 November 2023 CHAPTER 2: LITERATURE REVIEW 2.1 mHealth Technology mHeath technology can be defined as software that is installed in mobile devices such as a phone, smart phone, tablet etc. (Kumar et al., 2018). It is a subset of e-health and includes the use of mobile phones, tablets, smart watches for health related services, health information and data collection (Pai & Alathur, 2021). It enhances access to health related information for both patients and health care providers, and assists with remote patient care, health related data collection which leads to timely health care decisions and recommendations (Hood et al., 2016). mHealth can interact with the health system itself, health care providers and patients in improving the quality of health care (Osei & Mashamba-Thompson, 2021). mHealth technology has been recognized as playing a pivotal role in the surveillance and monitoring of infectious diseases such as Ebola, HIV, SARS and most recently COVID-19 (Osei & Mashamba-Thompson, 2021). The role that mHealth played in supporting the surveillance on the mentioned infectious diseases is an indicator of the increasing growth and impact of this technology. mHealth has also been studied as a possible health solution in managing chronic illnesses such as diabetes, epilepsy, hypertension indicating that mHealth could possibly play an important role in the management of diseases. In the use of diabetes self-management, one study found that there were too many mHealth applications which were overwhelming for users (Hood et al., 2016). This is different to other studies which found that there were too few applications to assess. The study by Hood et al., further found that most diabetes applications target patients broadly and not by the type of diabetes they have or other demographic factors. Targeting the applications by more specific details could be an important consideration for application designers to consider. The number of patients that were using an application for a study in the management of epilepsy had decreased significantly at the end of the study (Choi et al., 2021). This could be an indicator of application use issues and a further indicator that mHealth use should be studied further. Most studies in the management of chronic illnesses found that mHealth could have a positive impact on the management of these diseases. A literature review aimed at assessing the possible impact of mHealth in encouraging women to have their cervical cancer screening found that SMS interventions are useful for short term interventions but factors such as the SMS structure, whether the SMSes will be combined with 15 10 November 2023 phone calls or other communication types may be more beneficial for long term interventions (Bhochhibhoya et al., 2021). This was further supported by another study which was aimed at assessing the impact of mobile in the management of cervical cancer, the study found that short, clear and concise SMSes are more impactful in encouraging women to go for their cervical cancer screenings and could be a potential adoption hinderance if not taken into consideration (Moodley et al., 2019). This brings the spotlight into taking technology characteristics under consideration when designing an mHealth application. A literature review conducted to analyse commonly used mHealth applications found that SMS interventions for health promotion and rising health awareness were the most commonly used types of mHealth applications in low income countries (Abaza & Marschollek, 2017). In South Africa, most studies related to mHealth have focused on HIV care, information on physical activity, child health and gender based violence and other social issues (Mogoba et al., 2019). The topic trends in mHealth studies in South Africa are related to some of the social issues that burden the country. An investment in understanding how mHealth can be used to manage or educate on some of these social issues would be a worthy and positive investment. Mobile phones have been identified as having the capabilities to improve access to health and delivery in poor remote South Africa (Anstey Watkins et al., 2018). This support for mobile phones impact could further support the possibilities of tapping into understanding the impact that mHealth applications could have in the South African health sector. Most studies related to mHealth innovations indicate that scalability is a major challenge. A study which aimed to summarize the challenges related to moving an mHealth pilot study to full-scale adoption found that one of the challenges that must be overcome is that mHealth must be built with usability in mind (Wallis et al., 2017). The usability traits that should be considered include the user’s age, technological capabilities and role in relation to the application (Wallis et al., 2017). A study on the prospects and challenges of mHealth in developing countries, which South Africa is part of, found that in order to enjoy the benefits that come with mHealth interventions countries must design and develop strategies that overcome the challenges that mHealth interventions face at full implementation (Nsor-Anabiah et al., 2019). One of the two challenges identified as a barrier into the scalability of mHealth interventions in the developing world is developing applications which consider socio economic factors such as acceptance by users, technology awareness and other human factors (Istepanian et al., 2020). A study that had the purpose of analysing mHealth interventions in developing countries found that most mHealth interventions do not move beyond the pilot phase and further suggested that national health policies 16 10 November 2023 should include scalability and sustainability as important success measures for an mHealth intervention (Latif et al., 2017). The common theme is that interventions do not reach full scale and that usability should be an important consideration when building mHealth applications. This offers an opportunity to further explore factors that could impact the adoption and use of these mHealth applications. 2.2 mHealth Technology in South Africa The World health Organization (WHO) states that health related indicators in the Africa region continue to show slow progress despite efforts to decrease mortality and to slow down the impact of infectious diseases and that socioeconomic and epidemiological issues related to migration and climate change contribute to the disease burden in Africa (WHO, 2011). Furthermore, the challenges highlighted by WHO require new technological innovations emanating from the Africa region itself and even though this is a great opportunity, innovators still face challenges with scaling up their innovations (WHO, 2011). Scaling up high impact community based mHealth interventions is one of the agenda items mentioned in the National Digital health strategy for South Africa for the period 2019 -2024 (National Department of Health [NDoH], 2019). This is a good indicator that scaling up mHealth application is part of the government’s priorities. MomConnect, a nationwide mHealth system designed to equip South African pregnant women with pregnancy health related information with the aim of minimizing infant and maternal mortality has become the blue print for demonstrating that large scale digital health systems are feasible not just in South Africa but world wide (Mehl et al., 2018). A study which aimed to better understand the different approaches that can be taken to scale up a technological innovation noted that South Africa has an advantage in that there is a supportive policy which supports the absorption of innovations and also skilled partners, a high penetration of mobile phones and acceptable levels of technology literacy (Swartz et al., 2021). This shows that South Africa can scale up mHealth interventions successfully. Another review of mHealth interventions in developing countries suggested that there is a necessity for further research to focus on how countries can be assisted to reach their mHealth goals and increase the acceptance of mHealth interventions as a way to improve healthcare (Kruse et al., 2019). The success of MomConnect, supportive policy, high phone literacy and an increase in phone use are just some of the reasons that indicate that a model to assess the use and acceptance of mHealth technology in South Africa would be beneficial and is worth exploring further. 17 10 November 2023 A study aimed at exploring the impact of mHealth technology to report adverse events found that in order to understand the impact of mHealth interventions, factors that motivate further use should be researched (Chaiyachati et al., 2013). A study which aimed to scope the mHealth landscape in South Africa found that most of the mHealth initiatives are donor funded and no business model existed for implementation beyond the funding period, another scaling challenge (Botha & Booi, 2016). The study further found that most mHealth initiatives in South Africa are targeted towards the general public (individual users) and focus on drug adherence, HIV drugs distribution and pregnancy related information (Botha & Booi, 2016). A review on mHealth use in Sub Saharan Africa found a similar theme, which is that beneficiaries of mHealth interventions are HIV/AIDs patients, pregnant women, breast feeding mothers, children and general adults (Bervell & Al-Samarraie, 2019). A pilot study for an application called Masivukeni, an application meant to assist with Anti-retroviral therapy (ART) readiness counselling raised questions about the feasibility of scaling (McCreesh-Toselli et al., 2021). These studies and their attention on scalability or moving beyond the pilot phase indicate that scalability is still a serious challenge in the adoption and use of mHealth applications. There have been several studies conducted in South Africa to assess the possibility of mHealth being used support the health care services. A few are summarized in the next paragraphs. A study conducted in Zimbabwe, South Africa and Malawi found that SMS reminders improved the numbers of self-supported HIV testing (Govender et al., 2019). Another study related to SMS reminders found medication adherence support for hypertension delivered on a patient’s phone to be helpful and acceptable (Leon et al., 2015). A study which focused on piloting MHealth to support HIV self-testing services, found that the application was suitable for uploading the HIV results emanating from self-testing (Gous et al., 2020). In 2019, there was a study conducted to determine the tasks that could be automated using an mHealth solution to deliver home based HIV testing and counselling with the aim of improving linkage to care and adherence to ART medication (Comulada et al., 2019). A study which was aimed at studying out the effectiveness of a mHealth for diabetes found that patients preferred mHealth as a method of seeking medical treatment (Petersen et al., 2020). A study aimed at investigating the use of WhatsApp as an mHealth tool to support fracture management by doctors who are not orthopaedic doctors revealed that the use of WhatsApp decreased the delay to care and also reduced unnecessary referrals to already overburdened facilities and also empowered doctors who are not orthopaedic doctors to manage fractures (Kauta et al., 2020). These 18 10 November 2023 few mentioned studies just demonstrate the power of mHealth and its impact on how people can receive health care services in an improved way. Most studies on mHealth inventions in South Africa were on the management of HIV/AIDS and most of the studies reviewed indicate a positive possibility of how mHealth applications can impact the South African health sector. 2.3 mHealth Acceptance and Use A study aimed at understanding the determinants of mHealth adoption in developing countries, found that Performance Expectancy (PE) and Effort Expectancy (EE) are drivers in a user’s intention to use and adopt an mHealth application (Alaiad et al., 2019). Performance Expectancy relates to the user’s idea that a technological innovation will improve how they manage their health and as a result improve their lives. Effort Expectancy relates to how easy they think the technology is to use and whether it is convincement for completing the tasks that it is meant to automate. The study further indicated that when mHealth designers are building applications with PE in mind, they should consider applications that give users greater freedom in terms of managing their health and consider the amount of time it would take to use the mHealth application (Alaiad et al., 2019). Another technology characteristic the study suggested is building the mHealth application user interface in such a way that users can complete their tasks correctly and in a manner that does not take an unnecessary amount of time (Alaiad et al., 2019). The study also suggested that technology builders should consider the skills and demographics of the users that they are building the mHealth technology for if they want to consider the Effort Expectancy behavioural construct. The study findings are similar to a study which aimed to determine the constructs that can improve the acceptance and use of mHealth applications in that performance and effort expectancy have a significant impact on the acceptance and use of mHealth applications (Nunes et al., 2019). It is worth noting that Effort Expectancy only had a significant impact on users who were male. A study aimed at studying the factors that influence the behavioural intention to use physical activity mHealth applications in South Africa also found that Performance Expectancy, Effort Expectancy and Social Influence (SI) were dominant factors (Ndayizigamiye et al., 2020). The findings in the above studies just indicate that there is a link between the behavioural constructs that influence the user acceptance and use of mHealth and the technology characteristics. This brings into the spotlight the idea that studying a technological innovation’s intended users and mapping their behaviour to the characteristics of the technology has the potential to influence the user acceptance and use of the technology. 19 10 November 2023 A study aimed at understanding physicians’ views of how mHealth can impact the quality of health found that ease of use and timeliness are important in influencing physicians to view mHealth positively (O’Connor et al., 2020a). A study conducted with the purpose of understanding the acceptance of an mHealth system which was piloted in one of the districts in Ghana with the purpose of supporting children’s healthcare found that all the four constructs of UTAUT had an influence on acceptance and use (Brinkel et al., 2017). The level which individuals believe that the system will help them gain competitive advance, ease of use, organizational and technical infrastructure and social factors such as education had an impact o the intention to use the mHealth system (Brinkel et al., 2017). One study which deviates from most studies analysed found that trust was a motivator in the adoption of mHealth and cited that users struggle with adopting mHealth applications because they do not believe that mHealth applications can protect their personal information (Barutçu, 2018). Trust could be a critical decider for the intention to use and accept mHealth applications since some applications will be collecting personal health information. A systematic review aimed at analysing studies related to the use and acceptance of mHealth found that perceived usefulness and perceived ease of had the most impact on the behavioural intention to use an mHealth application (Kaium et al., 2019). Another review related to analysing the factors that could increase or decrease mHealth adoption also found perceived ease as a major contributor to the adoption (Agbenyo, 2019). A study into the adoption of mHealth by the elderly found that Performance Expectancy and Effort Expectancy did not have an impact in the adoption of mHealth, this finding was different from most studies (Palas et al., 2022). It found that Social Influence had an impact on the adoption and tied it to the cultural influence of family on health decisions. The common theme in most of the studies identified is that /Performance Expectancy and Effort Expectancy have a significant impact on the intention to use and accept mHealth. Users want to use technology that is easy to use and technology which will help perform the task that the technology aims to automate. This also then directly impacts the technology characteristics of the technology; technology must be easy to use and should add value to the lives of its users. 20 10 November 2023 2.4 Theoretical Frameworks 2.4.1 The Unified Theory of Technology Acceptance and Use of Technology (UTAUT) The UTAUT emanated from the notion that researchers were confronted with multiple theories and would often pick and choose the constructs from their favourite models, so a model that was directed towards a unified view of acceptance needed to be developed (Venkatesh et al., 2003). It was developed using common concepts and similar characteristics from eight different theories related to the use and acceptance of technology (Venkatesh et al., 2003). The UTAUT uses four variables that are predictors for adoption, Performance Expectancy, Effort Expectancy and social influence and another variable which has a direct impact on actual usage, Facilitating Conditions (FC) (Arfi et al., 2021). Age, gender, experience and voluntariness are moderating factors of the relationships in the model (Arfi et al., 2021). Performance expectancy refers to the belief that individuals who use the system will perform higher (Ayaz & Yanartaş, 2020). It is the degree which an individual believes that using the system will help them achieve a higher job performance (Alam et al., 2021). Consumers of a certain technology tend to be motivated to use and accept that technology if they believe that it will add value and make their daily lives easier (Alam et al., 2021). Effort Expectancy refers to the level at which it is convenient to use the system (Ayaz & Yanartaş, 2020). Social influence refers to the level which an individual whom the user holds in high regards deems the system as important (Ayaz & Yanartaş, 2020). It is related to the influence that the thoughts and activities of peers has on the behaviour of the individual (Alam et al., 2021). Facilitating condition is the degree which a consumer believes there is support and infrastructure to support the use the technology. 21 10 November 2023 The framework has been used to analyse the use and acceptance of various technological innovations at the individual level and at the organizational levels through studying out the patterns of individuals in the specific organization. Some use the extended version of the UTAUT, which is the UTAUT2. UTAUT2 was introduced by the same researchers who initially developed UTAUT. They extended the model by adding additional constructs and relationships to tailor it to a consumer use context (Venkatesh et al., 2012). UTAUT was used in the study. The UTAUT theory has been used to study the acceptance and use of mHealth from a patient’s perspective in Bangladesh (Alam et al., 2021). The study found that the model is a good model to use for the prediction of a young generation’s intention to use mHealth technology and was also aligned with other gender roles studies which found females to have higher technology anxiety and lower self-expectancy when dealing with new technological innovations (Alam et al., 2021). It has also been used to study the reason that people use location based applications for emergency situations which found that Effort Expectancy and facilitating conditions do not have a significant impact on the behavioural intention to use the technology, which means that users did not take the complexity of the technology and the level of support available for the technology into consideration when using the technology (Ayuning Budi et al., 2021). There have also been two studies where the model was used to Figure 1: UTAUT Model (Venkatesh et al., 2003, p. 447) 22 10 November 2023 understand the use and acceptance of internet of things technology. The one study found that Performance Expectancy played a pivotal role in the user’s usage of the technology and also that Effort Expectancy also had a positive impact on usage behaviour (Ronaghi & Forouharfar, 2020). Another study which used UTAUT to better understand the use and acceptance of internet of things technology similarly found that Performance Expectancy and Effort Expectancy have a positive impact on usage of the technology, if the technology improves a user’s daily life and is easy to use then it is likely to be adopted and used by users (Isaac et al., 2019). It is also worth noting that UTAUT was used in conjunction with Task-technology fit in the second study mentioned. Since age is also mentioned as a moderating factor in the relationships that exists within the model, other studies have focused on the acceptance and use by users from a particular age group. For instance, it was used to better understand the use and acceptance of information technology by older adults (Macedo, 2017). The study found that intention to use is a high determining factor in the actual usage of the technology and that older adult’s usage behaviour is directly proportional to the value that they think the technology adds to their lives (Macedo, 2017, p. 945). The study also found that effort efficacy and facilitating conditions play an important role in the acceptance of technology by older adults, and that technology should not be complex and that a supportive environment for the use of the technology should exist (Macedo, 2017). A study on the acceptance of an electronic document management system also re-enforced the fact that Performance Expectancy was an important determinant in the behavioural intention to use (Ayaz & Yanartaş, 2020). UTAUT has also been used to study the acceptance of telehealth services (Cimperman et al., 2016) and also patients’ intention to use an online portal to post an emergency department’s waiting times (Jewer, 2018). Other health related studies include a study which investigated nurses’ intention to use an Electronic Patient Record system (Maillet et al., 2015) and also to study the factors that influence the acceptance and use of an Electronic Record System in a context where there are limited resources (Shiferaw & Mehari, 2019). The study from the limited resource setting found that those developing health systems in Sub-Saharan Africa should focus on improving the attitudes, efficacy and awareness for health care professionals for the successful acceptance and use of those systems (Shiferaw & Mehari, 2019). A study which aimed to investigate the factors that influence a physician’s adoption of an Electronic Health Record (EHR) system in Bangladesh found that facilitating conditions play a pivotal role in the adoption of an electronic record system, so the availability of infrastructural support is important to physicians (Hossain et al., 2018). This study however found an insignificant relationship between Effort Expectancy, Performance Expectancy and the physician’s behavioural intention to use the EHR system (Hossain et al., 2018). Another study which aimed to understand the adoption of a mobile online hospital and their use by patients in West 23 10 November 2023 China found that Performance Expectancy, Effort Expectancy, and facilitating conditions were the main factors that were found to motivate patients to use the online hospital (Addotey-Delove et al., 2023). A similar study related to the adoption of mobile technology into the health system in Bangladesh also found Effort Expectancy and facilitating conditions to be significant in the adoption of the technology (M. Alam et al., 2019). Another common denominator in these studies related to health is facilitating conditions, users will use a technology if there is an enabling environment and infrastructure. UTAUT has been used in several studies related to the adoption of technology in the health sector, thus making it a suitable model for the study. 2.4.2 Task-Technology Fit (TTF) The TTF model emanated from two areas of research that complement each other, from user attitude as a predictor of system usage and task-technology fit as a predictor of performance (Goodhue & Thompson, 1995). It extends beyond the original proposed model by (DeLone & McLean, 1992) which states that both system usage and attitude related to the technology lead to individual user performance impact (Goodhue & Thompson, 1995). The TTF model focuses on the role that task-technology fit plays on performance impact, this is noteworthy since task-technology fit was a missing construct in many various models or was just hinted at and not explained explicitly (Goodhue & Thompson, 1995). The task characteristics in the model refer to the individual user’s actions and the technology characteristics refer to the technology that the individuals use to perform their tasks (Yang et al., 2013). Another definition for TTF is a way in which the technological innovation’s characteristics meet the task requirements of a user in such a way that performance is increased (Shahbaz et al., 2021a). The model essentially suggests that users will adopt a technological innovation based on how it will improve the efficiency of their daily tasks (Oliveira et al., 2014). It further suggests that the user’s performance while using the technological innovation will be determined by how well the technology fits the user’s task requirement (Park et al., 2015). 24 10 November 2023 Figure 2 is a pictorial view of the model. The figure demonstrates that task and technology characteristics affect the Task Technology Fit which lead to user utilization of the technological innovation and also individual performance of the user (Zhou et al., 2010a). There have been several studies which used TTF in the health sector. A study on the adoption factors of mobile technology for frontline patient care using TTF and human drivers found that understanding fit to be an important consideration when building a health information system (Junglas et al., 2009). One study which was intended to study the adoption of a personal health record system by patients who have chronic illnesses used TTF together with the Intervention Theory and the Protection Motivation Theory. The study found that task and technology characteristics play an important role in the Task Technology Fit and can influence adoption (Laugesen & Hassanein, 2017). Another study which was meant to better understand the adoption pattern of an electronic health record (EHR) by health care providers in the United States of America (USA) found that designers of EHR systems should take the characteristics of the organization into consideration when designing EHR system so as to ensure fit and in turn proving that technology and task characteristics are significant in the adoption of a EHR system (Gan, 2015). TTF in partnership with the UTAUT model has been used to study the acceptance of wearable devices in the health care sector. The task and technology characteristics were found to have an impact on TTF which further impact the decision to use and accept the wearable devices, the authors noted that this was different to other results on the adoption of wearable devices in the health sector (Wang et al., 2020a). TTF is a model that is just above 25 years old and has been used since its inception and is still being used in recent studies. The results in most of the studies are consistent in that they found task and technology Figure 2: TTF Model (Goodhue & Thompson, 1995) 25 10 November 2023 characteristics as an important determinant of requirements gathering when building a new technology. The model is also a good partner when matched with other theories as indicated by the studies listed above. TTF has a solid theoretical background. The reasons mentioned make TTF a suitable model to also use in the current study. The proposed model uses a combination of the UTAUT and the TTF models to derive a model for the factors that affect the use and acceptance of mHealth technology in the health sector in South Africa. Task characteristics and technology characteristics as determinants of Task Technology Fit which will also affect user adoption. 2.5 Model of the study 2.5.1 Task Characteristics The task characteristics of a technology can refer to actions that individual users put in as input for outputs that meet their information needs (Tam & Oliveira, 2016). A task can vary in how routine it is, the amount of time it takes to execute and whether is dependent on other tasks or not. The theory is that the more complex a technology is or the less functionalities an information technology offers, user’s positive view of the technology will decrease (Tam & Oliveira, 2016). This leads to the first hypothesis below: H1: Task characteristics of mHealth positively affect the Task Technology Fit. 2.5.2 Technology Characteristics Technology characteristics refers to the technological tool used by individual users to carry out their tasks (Tam & Oliveira, 2016). Attributes such as design and ease of use can affect how users view the technological innovation and also usage of the technology (Tam & Oliveira, 2016). Technology design can also affects how and how much a technology will be used, so a poor technology design or design choices not fitting the task will be overlooked for other technology (Tam & Oliveira, 2016). If a technology is designed with user’s requirements in mind, it creates a perception or idea that the technology will improve the user’s work activities or daily activities (Shahbaz et al., 2021b). Technology is likely to be adopted if it sufficiently supports the tasks it must perform (O’Connor et al., 2020b). Therefore: H2: Technology fit of mHealth positively affects the Task Technology Fit. 2.5.3 Task Technology Fit 26 10 November 2023 TTF refers to a fit between task and technology characteristics which will influence the use and task performance of individual users technology (Tam & Oliveira, 2016). If the tasks that a technology supports increase and become too complex for a technology to support, then the TTF decreases. If the technology functions increase to meet the task requirements, then TTF increases. The task characteristics refers to the use of mHealth applications to improve health (Yu et al., 2019). Therefore: H3: Task Technology Fit has a positive influence on the use of mHealth. 2.5.4 Performance expectancy Performance Expectancy refers to the positive impact that using a technology will have on users performing certain tasks (Wang et al., 2020a). It is a primary determinant in the use of technology (Wang et al., 2020). Therefore: H4: Performance Expectancy positively affects user adoption of mHealth. 2.5.5 Effort Expectancy Effort Expectancy refers to the degree which the user perceives the technology to be easy to use. If users believe that a technological innovation is easy to use then there is a higher change of them accepting the technological innovation (Wang et al., 2020a). Therefore: H5a: Effort Expectancy positively affects user adoption of mHealth. H5b: Effort Expectancy positively affects Performance Expectancy. 2.5.6 Facilitating Conditions Facilitating conditions refers to the degree which an induvial user believes organizational and technology infrastructure are available for the use of the technological innovation (Wang et al., 2020). Infrastructure such as internet bandwidth or ability for use the application offline can encourage technology use (Wan et al., 2020). These conditions facilitate the skills, knowledge, and support to use the technology (Wang et al., 2020). Therefore: H6: Facilitating Conditions positively affect user adoption of mHealth technology. 2.5.7 Construct relationship among UTAUT and TTF Previous studies have found that Task-Technology Fit has a positive impact on Performance Expectancy, which means that users will only adopt mHealth if the functionalities of the mHealth application meet 27 10 November 2023 their health management tasks (H. Wang et al., 2020). Many studies have also reported that the attributes of mHealth may reduce the effort to monitor health related issues. Therefore: H7: Task Technology Fit of mHealth has a positive relationship with Performance Expectancy. H8: Technology characteristics of mHealth have a positive relationship with Effort Expectancy. Figure 3: Proposed conceptual framework. 28 10 November 2023 Figure 3 is the proposed conceptual framework for the acceptance and usage of mHealth technology in the health sector in South Africa at the individual level. The figure demonstrates the possible constructs that can have a significant impact on the user adoption of mHealth through the different hypotheses. The independent variables from the TTF theory are task characteristics and technology characteristics and from the UTAUT theory, the independent variables are Performance Expectancy, Effort Expectancy, Social Influence and Facilitating Conditions. The independent variable from the combined model is the Task Technology Fit. The dependant variable is the user adoption of mHealth. Social Influence was not included as a construct as its not of interest for this study. Table 1 contains a summary of the hypotheses which were derived from the TTF and UTAUT theoretical framework. Table 1: Summary of Hypotheses Construct Hypothesis Number & Description References Task Characteristics H1: Task characteristics positively affects the Task Technology Fit (Tam & Oliveira, 2016, p. 237) Technology Characteristics H2: Technology fit of mHealth positively affects the Task Technology Fit (Tam & Oliveira, 2016, p. 237) (Shahbaz et al., 2021b, p. 4) Task Technology Fit H3: Task Technology Fit has a positive influence on the use of mHealth (Tam & Oliveira, 2016, p. 237). Performance Expectancy H4: Performance Expectancy positively affects user adoption of mHealth (Wang et al., 2020a, p. 4) Effort Expectancy H5a: Effort Expectancy positively affects user adoption of mHealth H5b: Effort Expectancy positively affects Performance Expectancy (Wang et al., 2020a, p. 4) 29 10 November 2023 Facilitating Conditions H6: Facilitating Conditions positively affects user adoption of mHealth technology (H. Wang et al., 2020, p. 2) TTF and UTAUT H7: Task Technology Fit of mHealth has a positive relationship with Performance Expectancy. H8: Technology characteristics of mHealth have a positive relationship with Effort Expectancy (H. Wang et al., 2020, p. 3) 30 10 November 2023 CHAPTER 3: RESEARCH METHODOLOGY 3.2 Research Paradigm and Approach The research paradigm used in the study is the positivist paradigm. The positivist paradigm can be described as a method for combining deductive logic with empirical observations of individual behaviour to try formulate rules that are based on a probability in order to come up with some pattern or generalization for human behaviour (Antwi & Hamza, 2015, p. 219). Researchers who work under the positivist paradigm work in quantitative terms to explain how items interact, how certain events are shaped and the causes of certain outcomes (Antwi & Hamza, 2015). Positivist researchers believe that reality stays the same and is not subjective to circumstances, and that it can be studied objectively (Rahi, 2017). Oates (Oates, 2005) states that positivism research has two fundamental assumptions or theoretical roots. The first assumption is that the world is ordered and regular, not just a haphazard occurrence (Oates, 2005). The second assumption is that the world can be analysed objectively without any emotions or feelings attached (Oates, 2005). Positivism focuses on finding out phenomena in the world without our personal preferences attached but in an objective manner (Oates, 2005). Researchers that use the positivism approach believe that we can understand the world through experiments and observation, it is also called the scientific or empirical method (Rahi, 2017). Positivism removes the researcher from what is being researched, numeric measurement is used with the aim of generating acceptable knowledge (Wahyuni, 2012). Numeric measurement was used for in the study, that is why the positivist research paradigm was appropriate for the study. The Positivist research is rooted in numeric and quantifiable data, so the research approach used in the study is the quantitative research approach. This is the scientific approach and focuses on collecting data from a sample that represents the population and the analysing the data but ignoring an individual’s feelings and emotions (Rahi, 2017). Quantitative researchers consider it very important to state their hypotheses and to test those hypotheses with empirical evidence to see if the hypotheses are supported or not (Antwi & Hamza, 2015). Positivism was selected as suitable for the study because the TTF and UTAUT frameworks were used to test the relationship between the dependant and independent variables and testing the hypotheses to establish the impact of the variables on the user adoption of mHealth technology. Furthermore, each hypothesis was tested using various statistical methods to test whether they are true or not, which also makes the paradigm suitable since. 31 10 November 2023 Table 2 shows the key characteristics of the quantitative research approach and demonstrates even further why it was the selected research approach for the proposed study. Table 2: Key Attributes of Quantitative Research (Basias & Pollalis, 2018, p. 93) Key Attribute Description Examines Interpretation The quantitative research method uses statistics and data analyses to interpret large amounts of data Usually Selected when - When it is necessary to analyse a large amount of data to verify or prove hypotheses - The research can be conducted using simple questionnaires with short answers that can be easily compared. - Data attribute Numeric data that can be collected using questionnaires Advantages - The result is quantifiable and therefore considered objective. - Quantitative data makes it easier to highlight changes and differences. - It makes it easier to compare numeric data. - Facilitates the development of valuation indicators 3.2 Research Design 3.2.1 Types of research designs The research design refers to the way that the researcher chooses to organize the different components of the research paper in a logical and coherent way such that the research problem is addressed adequately (Kumar, 2014). The research design allows us to translate our collected research into valuable and usable information (Nardi, 2018). It is a comprehensive plan for data collection in an empirical study and has the purpose of answering research questions or testing hypotheses (Bhattacherjee, 2012). It contains an overview of how data will be collected, measured and then analysed (Mishra & Alok, 2017). This is where the researcher communicates the plan that they will undertake to complete their study and describes the measures that they will take to collect valid, objective, and accurate data. There are many different types of research designs which a researcher can choose from depending on the research question they wish to address (Walliman, 2017). The choice depends on the type of problem that the researcher wishes to resolve (Walliman, 2017). The research design that was used in the study is the 32 10 November 2023 explanatory design type. It has the purpose of explaining observed phenomena and behaviour (Bhattacherjee, 2012). Explanatory research looks at phenomena through casual relationships (Rahi, 2017). It explains why things happen the way they do (Kumar, 2014). The main aspect is to make it clear as to why and how there is a relationship between variables (Kumar, 2014). The advantages of explanatory research can be summarized as follows: - They play a role in identifying the norms behind a wide range of issues (cscscholarship, 2018) - It is easy to replicate an explanatory study (cscscholarship, 2018) - They are associated with a high level of internal validity since there is an organized way of selecting the study subjects (cscscholarship, 2018) It was the most appropriate design for the study because it looks at cause and effect relationships, our research was aimed at summarizing the factors that affect the use and acceptance of mHealth applications by individuals in the South African health sector. The purpose of further understanding these factors was so that barriers to mHealth adoption could be mitigated and aiming to understand why sometimes there is adoption of mHealth and sometimes there is not. The reasons mentioned resonated with the core goal of explanatory research. 3.2.2 Survey Research The survey method is a systematic way of collecting data about people, thoughts, behaviour and their preferences in a standardized and systematic way using questionnaires or interviews (Bhattacherjee, 2012). Its main purpose is to gain a better understanding about a specific group or a sample representing a larger population (Jonker & Pennink, 2010). Surveys are mostly used to study individuals but can be used to study other groups such as organizations (Bhattacherjee, 2012). Though in such a case, a key informant for that group is used. In the case of the current study, the key informants are individual users of mHealth in the South African health sector. Surveys have many advantages when compared to other research methods, the first one is that they are a reliable way of measuring data that cannot be observed (Bhattacherjee, 2012). The second advantage is that they are an easy method for selecting data remotely (Bhattacherjee, 2012). Thirdly, they are seen as a convenient method of data collection by respondents since they can respond in their own time and space (Bhattacherjee, 2012). The fourth advantage is that the survey method is economical in terms of the researcher’s resources such as time, effort and money (Bhattacherjee, 2012). The study used the survey method because it is more suitable when trying to 33 10 November 2023 understand a group of individuals. The study also used a sample to represent a larger population, Gauteng province was used to generalize mHealth adoption factors related to South Africa. Thirdly the questionnaire was disseminated remotely another core positive of the survey method that was considered when making the decision to use the survey method. Lastly the survey method is also excellent for measuring people’s perceptions or beliefs about a specific topic. 3.3 Data Collection Methods Data collection is defined as the process of gathering information about a specific research matter in order to answer research questions, test hypotheses or evaluate outcomes (Kabir, 2016) . The reason that data was collected in the study was to test the hypotheses stated in Section 2.5 of this report. 3.3.1 Operationalization/ Measurement and Instrument Construction Operationalization is the process of designing precise measurements for abstract items (Bhattacherjee, 2012). The first step is to define an operational definition for the constructs that the researcher is interested in (Bhattacherjee, 2012). The constructs were developed using various adoption literature. A 7-point numerical scale survey was adopted with 7 indicating Strongly Agree and 1 indicating Strongly Disagree. This is consistent with other similar studies (Zhou et al., 2010b). The constructs are defined as per Table 3. Table 3: Measurement Constructs Constructs and Items Source Task Characteristics (TAC) - I need to access my health-related information anytime anywhere. - I need to acquire my health-related information anywhere. - I need to know my health-related information anywhere. (Zhou et al., 2010b) Technology Characteristics (TEC) - MHealth provide ubiquitous service. - MHealth provide real time services. - MHealth provide services where my privacy is protected. (Zhou et al., 2010b) Task technology fit (TTF) - The functions in mHealth are enough for the management of my health. - The functions in mHealth are appropriate for the management of my health. - Generally, the functions of the mHealth technology I use meet my needs. (Wang et al., 2020b) Performance expectancy (PEE) (Zhou et al., 2010b) 34 10 November 2023 - I feel mHealth is useful in collecting health related information. - Using mHealth enables me to obtain health related information quickly. - Using mHealth provides me with the information that I need. Effort Expectancy (EPE) - I can quickly and easily use a new technological tool. - I find that using mHealth that I use easy. - I think learning to operate new technology is easy for me. - My interaction with the mHealth that I use is clear and understandable. (Wang et al., 2020b) (Zhou et al., 2010b) Facilitating Conditions (FAC) - I have the necessary resources to use mHealth that I use. - I have the necessary knowledge to use the mHealth that I use. - Support is available if I struggle with a function of the mHealth that I use. (Wang et al., 2020b) User Adoption of mHealth (UAU) - I predict that I would use a mHealth to manage my health-related information. - I can develop a habit of using a health-related mobile application. - I foresee myself using a mHealth to manage health related issues (Wang et al., 2020b) The survey was created using Google forms, see Appendix A for the final research instrument. 3.3.2 Sampling and Respondents Sampling is a statistical process which involves selecting a subset from the population of interest such that observations and statistical inferences can be made about the population (Bhattacherjee, 2012). The first step is defining the target population (Bhattacherjee, 2012) . The population for the study was users who use mHealth in South Africa. The population categorisation for the study were adults (anyone older than 18) who own a smart phone and those who were users and are potential users of mHealth and are based in the province of Gauteng, South Africa, similar to other studies (Wang et al., 2020b). The second step in the sampling process is to select the sampling frame. This can be in the form of a contact list that is accessible (Bhattacherjee, 2012). Once the sampling frame was determined, the sample size was determined using Cochran’s Sample Size Formula, this formula as used because it is ideal for determining the sample size for large populations (Nanjundeswaraswamy & Divakara, 2021). The last step in the sampling process is choosing the sample using a selected sampling method (Bhattacherjee, 2012). The convenience sampling method was used for the research. Convenience 35 10 November 2023 sampling collects data from whoever is willing to take part in the study and is conveniently accessible (Scholtz, 2021). It was chosen to avoid high research costs and because it makes data collection easy. 3.3.3 Pilot testing Pilot testing is an important part of the research process, it can assist with the detection of research design issues or issues with the research instrument (Bhattacherjee, 2012). It ensures that the measurement instrument is a reliable and valid source of measure (Bhattacherjee, 2012). The instrument was piloted to 10 respondents before being disseminated to the sample for data collection. The 10 respondents were not part of the final study results. No major updates were made after the pilot except for suggestions on the presentation of the questionnaire as some respondents in the pilot phase struggled to navigate the questionnaire. 3.3.4 Questionnaire administration Once the pilot test was done, the data collection instrument link was administered online via e-mail and link sharing on social media. This was done through WhatsApp contacts and LinkedIn in contacts. The survey was shared between December 2021 and April 2022. The questionnaire was then turned off to stop accepting responses. 3.4 Data Analysis Strategy Internal consistency was tested. This is a measure of consistency amongst different items of the same construct (Bhattacherjee, 2012). This is a measure of how likely respondents rate items of a multi item construct (Bhattacherjee, 2012). Internal consistency was measured using Cronbach’s alpha to estimate this consistency. The second measure for consideration is construct validity. It measures whether the proposed construct is a good measure of the topic of interest (Bhattacherjee, 2012). Correlational and factor analysis was used for this measure. Face validity refers to whether a measurement seems to be a good measure of a construct. The analysis also analysed for content validity, this is an indication of whether scale items match the content domain of a construct. There are two types of statistics for analysing data, descriptive and inferential statistics. Descriptive statistics aggregate data that are grouped to examine typical values and the characteristics of the data whereas inferential statistics make conclusions about a group of data and can draw conclusions on a population based on a sample (Guetterman, 2019). The difference between the two is that inferential statistics goes beyond just describing the data (Guetterman, 2019). A combination of inferential statistics and descriptive statistics was used to analyse the data collected during the study. A comprehensive summary of the different findings and tests conducted on the data is available in Chapter 4 of this report. 36 10 November 2023 3.5 Ethical Considerations Ethics refers to the conforming of rules or regulations set by a specific group or a professional body (Bhattacherjee, 2012). Participants should be made aware that their participation is voluntary and that they can withdraw at any time (Bhattacherjee, 2012). This was stipulated in the e-mail and social media link sharing when the questionnaire was administered. Another consideration considered was to ensure confidentiality of participants (Bhattacherjee, 2012). This was done though ensuring that the questionnaire does not collect self-identifying data. Disclosure of who is doing the study, what the study is about, why they are doing the study, and the purpose of the study was shared to assist the participants with deciding on whether they want to part take in the study or not. It is also important to ensure that all results are reported, no matter what the outcome is (Bhattacherjee, 2012). This was adhered to for this study. The last consideration was any recommendations made by the Wits School of Business Sciences’ ethics committee under the approval for protocol CBUSE1963, see Appendix B for more details. 37 10 November 2023 CHAPTER 4: PRESENTATION OF RESULTS 4.1 Data Screening The purpose of this Chapter is to present the findings of the study. Both descriptive and inferential statistical results are presented in the chapter. A total of 107 participants voluntarily took part in the study. All participants gave consent confirming the below: 1. My participation in the survey is solely voluntary. 2. It is my right to withdraw from the survey at any given stage. 3. Only people who live in Gauteng, and are 18 years and above, can respond to the questionnaire. Only 70 of the participants were used as part of the analysis. Those excluded were missing responses in some questions due to the skip logic. Those excluded had answered ‘No’ to having used an mHealth application. The implication for the study was that only responses for those who responded to have used mHealth were used. The data set was imported into IBM SPSS for analysis. 4.1.1 Missing Data All the included responses in the survey were correctly completed in accordance with the skip logic in the survey. This is due to the questionnaire being designed in such a way that participants only respond to questions that were relevant to them. Another factor which led to there no being missing data is that all the questions were marked as required in the electronic survey shared and participants could not proceed to the next question before providing a response for the previous question. 4.2 Demographic Analysis The sample comprised of 41 females and 28 males, which translates to 58.6% and 40% respectively. 1.4% of the respondents did not provide their gender. Table 4: Gender Statistics Gender Frequency Percent Valid Percent Cumulative Percent Unspecified 1 1.4 1.4 1.4 Female 41 58.6 58.6 60.0 Male 28 40.0 40.0 100.0 Total 70 100.0 100.0 38 10 November 2023 Table 5 indicates the occupation that the respondents have. 84.3% said that they were employed, 8.6% said that they were self-employed, 5.7% said that they were students and 1.4% were unemployed. Table 5: Occupation Occupation Frequency Percent Valid Percent Cumulative Percent Employed 59 84.3 84.3 84.3 Self-Employed 6 8.6 8.6 92.9 Student 4 5.7 5.7 98.6 Unemployed 1 1.4 1.4 100.0 Total 70 100.0 100.0 were under the age of 30 and 44.3% indicated that they were between the ages of 31 -40. Whereas 24.3% indicated that they were between the ages of 41-50 and 10% were between the ages of 51-60. Only 4.3% of the respondents were older than 60Table 6 indicates the respondent break down by age group. 17.1% of the respondents indicated that they. Table 6: Age Group Age range Frequency Percent Valid Percent Cumulative Percent 18-30 12 17.1 17.1 17.1 31-40 31 44.3 44.3 61.4 41-50 17 24.3 24.3 85.7 51-60 7 10.0 10.0 95.7 Older than 60 3 4.3 4.3 100.0 Total 70 100.0 100.0 Table 7 indicates the respondent break down by level of education. 57.1% of the respondents indicated that their highest level of education was a postgraduate qualification and 25.7% indicated a bachelor’s degree as their highest qualification. Only 5.7% of the respondents indicated that their highest level of education was high school. Table 7: Level of Education Frequency Percent Valid Percent Cumulative Percent Bachelor’s degree 18 25.7 25.7 25.7 Diploma 8 11.4 11.4 37.1 High School 4 5.7 5.7 42.9 39 10 November 2023 Postgraduate 40 57.1 57.1 100.0 Total 70 100.0 100.0 Table 8: Cross-tabulation of Smart phone ownership and Use of mHealth Application Have you used a mobile health application? Total Yes Do you own a smartphone? Yes 70 70 Total 70 70 All the respondents who indicated to have owned a smartphone also indicated that they have used a mobile health application before as shown in Table 8. Table 9: Cross-tabulation of Use of mHealth Application and Employment Have you used a mobile health application? Total Yes Occupation Employed 59 59 Self-Employed 6 6 Student 4 4 Unemployed 1 1 Total 70 70 Table 9 represents the statistics on mHealth application usage and employment. Out of the 70 respondents who reported to have used a mHealth application, only one indicated to have been unemployed. Table 10: Cross-tabulation of Gender and Use of mHealth Application Have you used a mobile health application? Total Yes What is your gender? Unspecified 1 1 Female 41 41 Male 28 28 Total 70 70 40 10 November 2023 Table 10 represents the statistics on mHealth application usage and gender. Females represented 59% of the respondents who indicated to have used a mHealth application, whereas 40% of the respondents who indicated to have used a mHealth application were male. Table 11: Level of Education and Use of Mhealth Application Have you used a mobile health application? Total Yes Level of education Bachelor’s degree 18 18 Diploma 8 8 High School 4 4 Postgraduate 40 40 Total 70 70 Table 11 represents the statistics on mHealth application usage and level of education. 40 out of the 70 respondents who indicated to have used a mHealth application indicated that their level of education was a postgraduate qualification. 4.3 Validity and Reliability Reliability is the degree to which the measure of a construct maintains consistency. It refers to the notation that we should get the same results if we measure a construct multiple times (Bhattacherjee, 2012). Validity refers to the degree which a measure sufficiently represents the construct that it is supposed to measure (Bhattacherjee, 2012). The reliability and validity of the scales that were used to determine the factors that affect the use and acceptance of mHealth were determined using the Cronbach’s Alpha test. This had to be determined before the model could be tested. The internal consistency (reliability) of the measurement scales were assessed through the Cronbach’s alpha (α) coefficient. Cronbach’s alpha measures if all items in a in a test evaluate the same construct or concept (Tavakol & Dennick, 2011). Data is deemed reliable and acceptable if the evaluated Cronbach’s alpha is 0.70 or higher. The reliability statistics value was produced using SPSS for all the 22 items and had a value of 0.933. This indicates a reliable level of internal consistency for the scale. Table 12 represents these statistics. 41 10 November 2023 Table 12: Internal Consistency Values Cronbach's Alpha N of Items 0.933 22 Table 13 shows the item-by-item statistics indicating what the Cronbach Alpha would be if an item is deleted. The Cronbach Alpha increases to 0.936 when mHealth provides ubiquitous service (TEC1) item is deleted. It also has the lowest Item-Total correlation indicating that it is not a good measure for the construct and that it may be dropped. Table 13: Item-Total Correlation statistics Item Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Cronbach's Alpha if Item Deleted TAC1 120.56 293.410 0.609 0.930 TAC2 120.70 291.401 0.646 0.930 TAC3 120.61 294.936 0.542 0.932 TEC1 121.23 310.875 0.243 0.936 TEC2 120.93 297.372 0.601 0.930 TEC3 121.31 287.030 0.676 0.929 TTF1 121.81 285.052 0.708 0.928 TTF2 121.60 286.272 0.752 0.928 TTF3 121.53 288.050 0.762 0.928 PEE1 120.99 296.304 0.619 0.930 PEE2 121.03 299.680 0.519 0.932 PEE3 121.30 287.981 0.747 0.928 EPE1 120.91 301.790 0.460 0.933 EPE2 120.99 300.420 0.465 0.933 EPE3 120.89 298.856 0.558 0.931 EPE4 120.93 295.343 0.660 0.929 FAC1 120.74 295.121 0.733 0.929 FAC2 120.93 293.111 0.687 0.929 FAC2 121.33 289.151 0.656 0.929 UAU1 120.70 300.416 0.668 0.930 UAU2 120.63 304.469 0.499 0.932 UAU3 120.56 306.163 0.493 0.932 Factor analysis is a data reduction method used to statistically group large number of items into a smaller set of variables based on their correlation patterns (Bhattacherjee, 2012). Factor analysis was used for variable reduction and to investigate the structure of the data. Before factor analysis was performed on 42 10 November 2023 the data, the Kaiser-Meyer-Oklin (KMO) test was done to test that the sample size was sufficient for factor analysis. The KMO test measures whether the sampling was adequate (Shrestha, 2021a) for each construct in the model. The calculated KMO value was 0.836 as shown in Table 14, therefore the data could be used for factor analysis as KMO values between 0.8 and 1.0 indicate that sampling is adequate (Shrestha, 2021b). Bartlett’s test was also run on the data to test the assumption that variances are homogenous across groups. A level of correlation between the variables must exist to run correlation analysis test. Bartlett’s test aims to identify the correlation. By using SPSS, the value for Bartlett’s was determined and for this study it was significant i.e., p < 0.001, therefore, showing that the correlation matrix has significant correlations amongst some of the variables indicating that using factor analysis was appropriate (Shrestha, 2021b). Table 14: KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .836 Bartlett's Test of Sphericity Approx. Chi-Square 1219.110 df 231 Sig. .000 Principal Component Analysis (PCA) was conducted with the seven factors that impact the adoption of mHealth adopted from the TTF and UTATUT models. The goal of principal component analysis is to extract important statistical information from the data (Nguyen, 2017). It was found that seven factors accounted for 82.6% of the total variance in the model as shown in Table 15. Coefficients of less than 0.40 were suppressed from the results for visual clarity. The Varimax method of rotation was used as it is found to minimize the number of variables which have high loadings on a factor to only have variables which load strongly on the factors. Table 15: Principal Axis Factor Analysis Total Variance Explained Component Initial Eigenvalues Rotation Sums of Squared Loadings Total % Of Variance Cumulative % Total % Of Variance Cumulative % 1 9.535 43.341 43.341 3.816 17.347 17.347 2 2.463 11.197 54.538 3.305 15.025 32.372 3 1.956 8.892 63.431 2.998 13.629 46.001 4 1.350 6.135 69.566 2.759 12.541 58.543 5 1.190 5.410 74.976 2.243 10.195 68.738 6 .960 4.362 79.338 1.572 7.145 75.882 43 10 November 2023 7 .716 3.254 82.592 1.476 6.710 82.592 8 .605 2.751 85.343 9 .509 2.315 87.658 10 .447 2.033 89.692 11 .392 1.781 91.472 12 .345 1.567 93.040 13 .254 1.155 94.195 14 .235 1.069 95.264 15 .212 .962 96.225 16 .185 .842 97.068 17 .153 .693 97.761 18 .147 .666 98.427 19 .140 .634 99.061 20 .094 .426 99.487 21 .068 .307 99.794 22 .045 .206 100.000 Extraction Method: Principal Component Analysis. The Rotated Component Matrix in Table 16 shows the loadings for the seven components. The third measure of Facilitating Conditions was observed to load into a different component (Effort Expectancy) from the other two measures of Facilitating Conditions. This measure was included as a measure of Effort Expectancy. Also, the second measure of Technology Characteristics was observed to also load into another component (Task Technology Fit) from the other two measures. Most of the items had a high loading value and thus indicated a high level of correlation with the constructs they were trying to measure. 44 10 November 2023 Table 16: Rotated Factor Matrix Rotated Component Matrixa Component 1 2 3 4 5 6 7 TTF1 .800 TTF2 .748 EPE4 .698 .505 TTF3 .682 .439 TEC2 .572 .439 FAC3 .543 .519 TAC1 .914 TAC2 .905 TAC3 .886 PEE2 .907 PEE1 .739 PEE3 .702 EPE2 .541 .454 UAU2 .840 UAU3 .839 UAU1 .787 EPE1 .442 .767 EPE3 .736 FAC2 .701 FAC1 .677 TEC3 .419 .434 .463 .410 TEC1 .893 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 9 iterations. 45 10 November 2023 4.4 Normality Tests Composite scores were calculated for each of the factors by averaging scores of the items measuring the variable. The composite scores were tested for normality to determine if they were a representative of a normally distributed population. The Kolmogorov-Smirnov and Shapiro-Wilk is used to determine how well a random sample is evenly distributed (Şahintürk & Özcan, 2017). The two tests were used to compare the distributions of the data against reference normal distributions. These tests yielded non-significance scores (p<0.05), which indicated that there was a significant deviation from the normal distribution. The conclusion on the distributions of the factors was to not assume normal distribution of the variables. This meant that non-parametric tests were used since the assumptions of normality were not met. Table 17: Skewness and Kurtosis Table 18: Tests of Normality Factor Skewness Kurtosis UAU -1.056 1.029 TTF -0.864 1.378 TAC -1.918 3.69 PEE -0.753 0.177 EPE -0.588 -0.525 FAC -1.518 3.356 TEC -0.676 0.224 Tests of Normality Kolmogorov-Smirnova Shapiro-Wilk Statistic df Sig. Statistic df Sig. UAU .197 70 .000 .856 70 .000 TTF .125 70 .009 .941 70 .003 TAC .269 70 .000 .701 70 .000 PEE .116 70 .021 .918 70 .000 EPE .147 70 .001 .927 70 .001 FAC .209 70 .000 .844 70 .000 TEC .202 70 .000 .926 70 .001 a. Lilliefors Significance Correction 46 10 November 2023 4.5 Correlation Analysis Composite values of the independent and dependent constructs were computed in SPSS and correlation analysis was conducted to determine the strengths of the relationships between them. The Spearman correlation was used since the normality test showed that the distribution of factors does not follow a normal distribution (Mukaka, 2012). Hypothesis 1 H1: Task characteristics of mHealth positively affect the Task Technology Fit. Table 19: Correlation of Task Characteristics and Task Technology Fit Correlations TAC TTF Spearman's rho TAC Correlation Coefficient 1.000 .354** Sig. (2-tailed) . .003 N 70 70 TTF Correlation Coefficient .354** 1.000 Sig. (2-tailed) .003 . N 70 70 **. Correlation is significant at the 0.01 level (2-tailed). The relationship between Task Characteristic and Task Technology Fit was found to be statistically significant (0.354, p < 0.001) at a weak level. This finding provides support for H1, that the higher the Task Characteristic, the higher the Task Technology Fit Hypothesis 2 H2: Technology fit of mHealth positively affects the Task Technology Fit. Table 20: Correlation of Technology Fit and Task Technology Fit Correlations TTF TEC Spearman's rho TTF Correlation Coefficient 1.000 .488** Sig. (2-tailed) . .000 N 70 70 TEC Correlation Coefficient .488** 1.000 Sig. (2-tailed) .000 . N 70 70 **. Correlation is significant at the 0.01 level (2-tailed). 47 10 November 2023 The relationship between Technology Fit and Task Technology Fit was found to be statistically significant (0.488, p < 0.01) at a moderate level. This finding provides support for H2, that the higher the Technology Fit, the higher the Task Technology Fit. Hypothesis 3 H3: Task Technology Fit has a positive influence on the use of mHealth. Table 21: Correlation of Task Technology Fit and User Adoption of mHealth. Correlations TTF UAU Spearman's rho TTF Correlation Coefficient 1.000 .491** Sig. (2-tailed) . .000 N 70 70 UAU Correlation Coefficient .491** 1.000 Sig. (2-tailed) .000 . N 70 70 **. Correlation is significant at the 0.01 level (2-tailed). The relationship between Task Technology Fit and the behavioral intention to adopt mHealth to be statistically significant (0.491, p < 0.01) at a moderate level. This finding provides s