It is undeniable that most business organizations rely on the Internet to conduct their highly competitive businesses nowadays. Cyber security is one of the important elements for companies to guarantee the normal operation of their business activities. However, there is no panacea in cyber security protection. Common security practices used are to deploy hardware and software security protection tools to combat the known security threats which may become more and more powerful later. In fact, the attackers and security practitioners are at war from time to time. As a result, such a tools-based security protection strategy cannot be sustained. On the other hand, the related awareness training for employees is ignored in a number of companies, which has made biased the decisions made by staff when facing cyber security breaches. In this study, in order to find ways to sustain such protection, we conduct a quantitative analysis to explore the key elements contributing to the SETA implementation of the companies and organizations. We evaluate the performances of eight supervised learning models in a dataset collected from cyber security breach surveys on UK businesses to perform a fundamental analysis. The detailed analysis is performed via the feature importance of features generated in the model with better performance in the task of detecting the companies and organizations with SETA implementation. The experiment result shows that the awareness related factors play the most significant role in the SETA implementation decision-making for the businesses, and most of the businesses are lacking the awareness to prevent the potential cyber security risks in the stuff using externally-hosted web services and products as well as services depending on online services.

BACKGROUND Antimicrobial resistance (AMR) has emerged as one of the most serious threats to global public health. The proliferation of social media platforms as digital channels for health information has facilitated the dissemination of personal experiences with antibiotic use, but some of this information may be misleading. Such misleading information may encourage the inappropriate use of antibiotics, thereby accelerating AMR. Despite growing concerns, there remains a significant gap in the literature regarding both the quality of information in these posts and the specific features that drive their popularity. OBJECTIVE Taking sore throat as a focal case, this study aims to achieve two primary objectives: (1) to systematically evaluate the informational quality of social media posts that share personal experiences with antibiotic treatment and (2) to quantitatively identify specific features associated with higher levels of audience engagement. METHODS The data were collected from a leading social media platform, Xiaohongshu, to conduct this study. Two clinical pharmacists assessed the appropriateness of the medication-related practices described in the posts and coded the relevant features, which are identified according to the Elaboration Likelihood Model (ELM) and the Health Belief Model (HBM). Negative binomial regression was employed to examine the effects of these features on user engagement, as measured by the number of likes and collections. RESULTS Among the 1061 posts analyzed, the overall quality of the antibiotic-related posts was worrisome. As many as 31.39% of the posts involved antibiotic use without clinical indication, and many displayed issues related to suboptimal drug selection and inappropriate dosing regimens. In examining features associated with post popularity, this study found that the presence of treatment information in post title (likes: incidence rate ratio (IRR)=1.65, P<.001; collections: IRR=1.92, P<.001), the use of a well-designed cover image (likes: IRR=1.43, P=.002; collections: IRR=1.67, P<.001), and the inclusion of recovery outcome (likes: IRR=2.02, P<.001; collections: IRR=2.28, P<.001), detailed description (likes: IRR=2.08, P<.001; collections: IRR=2.14, P<.001), and advice (likes: IRR=1.78, P<.001; collections: IRR=2.14, P<.001) in the content were all significantly associated with increased user engagement. In addition, the imperative language and linguistic intensifier in post titles had different effects during the COVID-19 pandemic's peak and stable phases. CONCLUSIONS The findings of this study indicate that the quality of personal antibiotic treatment experiences shared on social media is concerning. Furthermore, the study identified several post features that are associated with increased popularity in this context. The findings provide valuable insights for health authorities in developing more effective health promotion strategies.