BACKGROUND Previous studies have identified that the effective management of cyber security in large healthcare environments is likely have a significant dependence on human and social factors, as much as technical controls. However, there have been limited attempts to confirm this with measured and integrated studies focussed on users in order to understand their motivations and behaviours. OBJECTIVE This study aimed to gather data from a variety of users working within large Australian healthcare provider (LAHP) environments, to record and verify their motivations in understanding and applying essential cyber security messaging and operational controls. METHODS An explanatory sequential mixed-methods approach was undertaken, gathering quantitative data via a user survey (n=103), with variables mapped to the extended Technology Acceptance Model (TAM2) and results scrutinized via an exploratory factor analysis. A parallel qualitative review was also undertaken via in-depth interviews with staff (n=9), which were also encoded to the TAM2 model and examined for thematic frequency. The coded comments were analysed to create an emic-to-etic understanding, which was integrated with the quantitative results to produce verified conclusions. RESULTS Via both the quantitative and qualitative investigations, users prioritised the TAM2 Perceived usefulness of cyber security measures, and the operational subjective norms of their professional peers as being the main influences on their knowledge and behaviours. CONCLUSIONS To pragmatically manage and improve cyber security governance in large healthcare environments, efforts should be focussed on demonstrating how controls can benefit staff and patients without getting in the way of complex treatments. Further consideration needs to be given to how clinicians need to share data and collaborate on patient care, with tools and processes provided to support and manage this securely.

Background Previous studies have identified that the effective management of cyber security in large health care environments is likely to be significantly impacted by human and social factors, as well as by technical controls. However, there have been limited attempts to confirm this by using measured and integrated studies to identify specific user motivations and behaviors that can be managed to achieve improved outcomes. Objective This study aims to document and analyze survey and interview data from a diverse range of health care staff members, to determine the primary motivations and behaviors that influence their acceptance and application of cyber security messaging and controls. By identifying these issues, recommendations can be made to positively influence future cyber security governance in health care. Methods An explanatory sequential mixed methods approach was undertaken to analyze quantitative data from a web-based staff survey (N=103), with a concurrent qualitative investigation applied to data gathered via in-depth staff interviews (N=9). Data from both stages of this methodology were mapped to descriptive variables based on a modified version of the Technology Acceptance Model (TAM; TAM2). After normalization, the quantitative data were verified and analyzed using descriptive statistics, distribution and linearity measures, and a bivariate correlation of the TAM variables to identify the Pearson coefficient (r) and significance (P) values. Finally, after confirming Cronbach α, the determinant score for multicollinearity, and the Kaiser-Meyer-Olkin measure, and applying the Bartlett test of sphericity (χ2), an exploratory factor analysis (EFA) was conducted to identify the primary factors with an eigenvalue (λ) >1.0. Comments captured during the qualitative interviews were coded using NVivo software (QSR International) to create an emic-to-etic understanding, which was subsequently integrated with the quantitative results to produce verified conclusions. Results Using the explanatory sequential methodology, this study showed that the perceived usefulness of security controls emerged as the most significant factor influencing staff beliefs and behaviors. This variable represented 24% of all the variances measured in the EFA and was also the most common category identified across all coded interviews (281/692, 40.6%). The word frequency analysis showed that systems, patients, and people represented the top 3 recurring themes reported by the interviewees. Conclusions To improve cyber security governance in large health care environments, efforts should be focused on demonstrating how confidentiality, integrity, availability, policies, and cloud or vendor-based controls (the main contributors of usefulness measured by the EFA) can directly improve outcomes for systems, staff, and patients. Further consideration also needs to be given to how clinicians should share data and collaborate on patient care, with tools and processes provided to support and manage data sharing securely and to achieve a consistent baseline of secure and normalized behaviors.

Purpose This paper proposes a novel cyber security risk governance framework and ontology for large Australian healthcare providers, using the structure and simplicity of the Unified Modelling Language (UML). This framework is intended to mitigate impacts from the risk areas of: (1) cyber-attacks, (2) incidents, (3) data breaches, and (4) data disclosures. Methods Using a mixed-methods approach comprised of empirical evidence discovery and phenomenological review, existing literature is sourced to confirm baseline ontological definitions. These are supplemented with Australian government reports, professional standards publications and legislation covering cyber security, data breach reporting and healthcare governance. Historical examples of healthcare cyber security incidents are reviewed, and a cyber risk governance UML presented to manage the defined problem areas via a single, simplified ontological diagram. Results A clear definition of ‘cyber security’ is generated, along with the ‘CYBER-AIDD’ risk model. Specific examples of cyber security incidents impacting Australian healthcare are confirmed as N = 929 over 5 years, with human factors the largest contributor. The CYBER-AIDD UML model presents a workflow across four defined classes, providing a clear approach to implementing the controls required to mitigate risks against verified threats. Conclusions The governance of cyber security in healthcare is complex, in part due to a lack of clarity around key terms and risks, and this is contributing to consistently poor operational outcomes. A focus on the most essential avenues of risk, using a simple UML model, is beneficial in describing these risks and designing governance controls around them.