The most recent technology, the Internet of Things (IoT), is essential to the digitalisation of the healthcare sector (Boyes et al. 2018). Furthermore, the adoption of connected healthcare is accelerating globally over the past two decades as healthcare IoT cases gain momentum (Shah & Chircu 2018). The impact of leveraging digitalisation across various levels of processes, products and services demonstrates that the healthcare industry is one of the significantly revolutionised sectors (Kumar, Verma & Bharti 2020; Moro Visconti & Morea 2020). Furthermore, the electronic medical record (EMR) represents a change from the antiquated practice of keeping paper records to computerised electronic data management. It is now acknowledged as a contemporary productivity, efficiency and effectiveness in medical care facilitation (Adeleke et al. 2014; Sockolow et al. 2014). In addition, it is supported by systems for Internet networks, providing flexibility in the ability to transfer information, bring about change in the field of healthcare and enhancing patient health outcomes. Nene (2015) claims that a patient at South Africa’s King Edward VIII Hospital had to wait more than eight hours before receiving treatment for diarrhoea because staff members needed to find his medical file in the file room. Medical records can occasionally be difficult to locate in the filing department as a result of, among other things, records being lost or misfiled (Nene 2015). These records are occasionally simply misplaced by healthcare providers and later discovered at their offices. In addition, the duplicate record would exist because the original record would have been changed by the time the replacement record was found.

In another study by Marutha (2011), doctors were prevented from operating on a patient at Nkhesani Hospital in Limpopo province, South Africa, because the patient’s medical records were missing. The patient’s left leg was partially paralysed as a result of a car accident in 2005. According to the report, the doctors needed the missing file, which contained information about the patient’s accident, health problems and illness, in order to trace the history of the injuries before the patient could have surgery. Unfortunately, the patient’s diabetes records were the only part of the medical history that could be accessed. According to the Limpopo Department of Health and Social Development, which intervened and provided clarification in this case, the hospital had a minor file mix-up. It is evident from the aforementioned cases that when patient health records in paper format are missing, patients face significant challenges and risk their lives, which may result in death. The medical institution could also be harmed by this. Duplicate data is regarded as poor and untrustworthy patient data and is a concern for many organisations and systems, according to Bardach, Real and Bardach (2017). Another disadvantage of paper-based medical records is that data cannot be shared by many users because the physical document is often only read in one place (Dowding, Turley & Garrido 2015).

Clinical decision-making is thought to be improved by the integration of various information tools into EMR, such as test ordering, electronic prescriptions, decision support systems, digital imaging and telemedicine. Furthermore, integrating such information into standard clinical practice might encourage the use of a secure and effective healthcare system (Sockolow, Bowles & Adelsberger 2014). Access to patients’ vital health information from various providers, which greatly enhances care coordination and healthcare practice efficiency, is one of the main advantages of EMR that is stated (Jawhari, Keenan & Zakus 2016). Several of these advantages have been supported by earlier studies (Kruse, Mileski & Alaytsev 2015). In the end, the push for electronic health records (EHR) may make it possible for patients to access and possibly transfer their own medical data, giving them more control over their own care and management of their medical data (Chao et al. 2013; King et al. 2014). Furthermore, because of the complexity of healthcare delivery, conventional decision-making methods for implementing new technologies based on trustworthy and predictable systems are inappropriate (Kuziemsky 2016). This study’s aim was to investigate the critical success factors that influence the adoption of evidence-based healthcare practice (EBHP) in a South African public hospital. A theoretical framework based on a model based on the updated Information Systems (IS) Success Model (D&M) model, the Technology-Organisation-Environment (TOE) framework and Technological Acceptance Model (TAM) was developed to assist in identifying the critical success factors. Structural equation modelling (SEM) was used to test and validate the conceptual framework using survey data gathered at Dr George Mukhari Academic hospital.

The next section will discuss in greater detail on the theories that were used to develop the conceptual framework, the constructs and the formulation of the hypothesis based on the framework.

This study suggests a hybrid model for the adoption of EBHP. The hybrid model is based on updated versions of Davis, Bagozzi and Warshaw (1989), the TAM and the TOE, as well as DeLone and McLean’s IS Success Model (DeLone & McLean 2003; Tornatzky & Fleischer 1990). Technology Acceptance Model was chosen for this study because it concentrates on how individual users perceive technology and assesses how those perceptions might affect their behavioural intention. The goal of this study is to develop a framework for EBHP at a South African public hospital. For this purpose, the factors that contribute to the adoption of EBHP will be analysed using an integration of two theories namely the TAM and TOE into D&M IS Success Model, which is the main underpinning theory in this study.

The D&M IS Success Model assisted in finding the individual level acceptance of the new technology like EHR system, whereas the TOE model was a logical extension of D&M IS Success Model from an individual perspective to organisational factors. While D&M IS Success Model and TOE frameworks helped to understand the decision method for technology adoption and some aspects of implementation, TAM theory provided the usability factor from the individual user perspective. For this study, the EBHP construct was used in place of the Net-benefits construct from the D&M IS Success Model. This construct is a dependent variable in the proposed research paradigm (see Figure 1), while the independent variables are medical error reduction (MER), information quality (IQ), knowledge quality (KQ), perceived ease of use and organisational context.

The influence of EHR on medical mistake reduction, IQ, KQ, improved patient coordination, disease diagnosis and treatment, perceived usability, technological context and organisational context were all examined in this study. The findings highlight the importance of improving patient responsiveness and orientation to expat demands in South Africa’s healthcare system.

In the section that follows, we will discuss in greater detail how the constructs from TAM and TOE that were incorporated into the D&M IS Success Model as well as how they relate to each other. These relationships are shown in Figure 1: hypothesis research model.

The following section describes the process undertaken to analyse the data collected from the questionnaire. The first step was to undertake basic descriptive tests for 300 participants using SPSS Version 23, followed by exploratory factor analysis (EFA) to determine whether items loaded on their respective constructs as anticipated (Gaskin et al. 2014; Hair et al. 2010). Cronbach’s alpha was performed to check the internal consistency of the items for reliability (Bagozzi & Yi 1998), followed by confirmatory factor analysis (CFA) to determine the convergent and discriminant validity of the measurement model (Gaskin et al. 2014; Holmes-Smith 2010). Lastly, SEM in the form of path analysis was used to test the hypotheses (Holmes-Smith 2010). The results of the data analysis are discussed in Section 7.

To develop an integrated model for EBHP, this study used the updated (D&M) IS Success Model, TOE and TAM. Structural equation was then used to analyse data. In addition, to evaluate how well the model fit the data from the EFA, CFA was also used. The results of study confirmed that between EHRs have a positive significant influence on IQ (β = 0.154; p < 0.05). For this reason, H1 was supported. The findings of this study corroborate those of Meeks et al. (2014), who found that the adoption of EHR may have a considerable effect on patient quality and safety.

As predicated in the research model, knowledge quality has a positive significant effect on better patient coordination of patient care (β = 0.985; p < 0.05). However, this H2 was supported. This finding is in line with previous research on the adoption of technological innovation (Shaltoni 2017). Findings of the study reaffirm a positive association between higher knowledge quality and better patient coordination, resolving the issue of contradictory findings in the literature. The results of H3 confirmed that EHRs have a positive significant influence on knowledge quality (β = 0.558; p < 0.05). This hypothesis was also supported. These findings are supported by Zayyad and Toycan’s (2018) who claim that EHR facilitates timely communication among all healthcare providers as well as that access to patient medical records and related information. The study did not confirm that EHRs have a positive significant effect on perceived ease-of-use (β = 0.021; p > 0.05). For this reason, H4 was not supported. This finding is in contradiction with that of Hwang, Dutta and Chang (2019), who discovered a significant correlation between attitude and perceived ease of use in the context of physician acceptance of EHR.

The results of the study, however, did confirm that top management’s support has a positive significant influence towards the use of EHRs (β = 0.087; p > 0.05). Therefore, H5 was supported in this study. However, the result is consistent with prior research, which demonstrated that leadership roles influence effective implementations in a variety of situations (Aarons, Ehrhart & Farahnak 2014). In addition, the study confirmed that knowledge quality and better patient coordination of patient care are positively and significantly related (β = 0.985; p < 0.05). For this reason, H6 was supported. This result is consistent with Berghout, Fabbricotti and Buljac (2017) assertion that doctors are highly knowledgeable professionals and that, as the largest group employed by hospitals, their decisions are typically based on their knowledge and experience. Furthermore, this study’s results did not support BCP has a negative insignificant impact in the reduction of medical errors (β = 0.048; p > 0.05). According to Mitchell (2019), the adoption of an EHR will increase patient outcomes by enhancing the ability to diagnose illnesses and reducing or even preventing medical errors. These findings contradict with those of Mitchell. For this reason, H7 was not supported.

The results of this study confirmed that EBHP has positive significant influence on DTD (β = 0.154; p < 0.05). H8 was supported as a result. These results are consistent with Pyrene’s (2015) claim that access to medical records is important as it is practically difficult for medical professionals to offer the best diagnosis or treatment without it. In addition, the study also confirmed that better patient coordination of patient care is positively influenced by the use of EHRs (β = 0.64; p < 0.05). This is also predicated in the research model and hypothesis H9 was supported. These results concur with those of Meeks et al. (2014), who claimed that the use of EHRs improves decision-making, record-keeping and patient care monitoring, allowing medical practitioners to offer better care. Furthermore, the results of H10 confirmed that EBHP has a direct positive influence on better patient coordination of patient care (β = 0.75; p < 0.05). Therefore, H10 was supported in this study. The outcomes of these findings support the claims made by Krist et al. (2014) that implementing an EHR system can facilitate health information exchange by catalysing the sharing of clinical information and patient care coordination among healthcare providers, potentially leading to higher-quality patient care. Since the model was tested and matches the data, the updated (D&M) IS Success Model, TOE and TAM frameworks can be used in the study to investigate the crucial successful factors for the adoption of EBHP in hospitals. This theoretical framework was developed to assist in determining the critical factors required for the adoption of EBHPs at a public hospital in South Africa.

This study contained a cross-sectional survey to understand which critical factors influence EBHP adoption at a public hospital in South Africa. With 300 valid responses, the results showed that the proposed model was acceptable with high significant of statistics for model fit indices. Nurses made up the majority of the participants (225 in total). The updated D&M IS Success Model, TAM and TOE formed the theoretical foundation. Seven dependent variables, EBHP and five independent variables: IQ, EHR, KQ, MER, BCP and DTD, were adopted from the updated and modified D&M IS Success Model. Perceived ease of the variable adopted for TAM while from TOE, organizational context was adopted and these variables were independent variables. The findings of this study reveal that the adoption of EBHP is strongly influenced by EHR, which is a crucial construct. In addition, the findings of this study also showed a strong positive significant relationship between the adoption of EBHP and the constructs of EHRs, MER, IQ and KQ. One of the key findings of this study was that the model used here for the adoption of EBHPs by the medical healthcare professionals at a public hospital in South Africa fit the data well and could be used to steer future research.

The current study extended and justified the existing D&M IS Success Model by adding KQ and reduction of medical errors as new constructs. BCP was also added to replace the moderating construct ‘intention to use’ and the construct user satisfaction was modified to read DTD. Net-benefits construct, the dependent variable in D&M IS Success Model was changed to read EBHP. With the adoption of the developed EBHP framework in public hospitals, healthcare professionals will be able to shift from a culture of providing care based on tradition, intuition, and authority to a situation where decisions will be guided and supported by the best available medical evidence.

This study contributes to advancing knowledge about the adoption of EBHP in the context of developing countries like South Africa, which may provide a fruitful avenue for future research. A sample population from one academic hospital in Gauteng province was used for the current investigation. Therefore, it is possible that the results do not accurately reflect the critical success factors that have a bearing on the adoption of EBHP in South African public hospitals. Future study is therefore required with a larger sample size, more varied sampling locations and a variety of sampling strategies.

In addition, the insignificant relationship between perceived ease of use and EHR H4 needs to further be investigated in a different hospital setting in other provinces. Furthermore, research should be conducted to determine the critical success factors that influence the adoption of EBHP in South African public hospitals, with additional data collection strategies such as interviews and focus groups being used to determine which is the most viable. Furthermore, the model of the current research can be modified and be applied to the adoption of other types of e-health services, such as telemedicine and mHealth (mobile Health), in South Africa or other developing countries.