evaluation models

Given that artificial intelligence (AI) is a newly emerging field transforming the practice of medicine, the end-user perspective is of paramount importance for its success and acceptance.

This systematic review aimed to capture an overview of qualitative and quantitative surveys related to patients' opinions on the use of AI within healthcare settings.

In this systematic review, a query was conducted on PubMed for original studies that surveyed patients' opinions using qualitative and quantitative methods. The key inclusion criteria were papers assessing patient viewpoints assessed by interviews and those exploring opinions via questionnaires. The data extraction process involved collaborative analysis to ensure reliability.

This systematic review analyzed 26 studies on patient perspectives of AI in healthcare. The majority of articles used quantitative surveys (65.4%) or qualitative interviews (19.2%), with convenience and purposive sampling being the most common. Cancer patients were the most frequent group studied (26.9%), with research on AI applications in cancer care. The key factors examined in quantitative surveys were prior AI exposure, perceptions of the advantages/drawbacks of AI, and privacy/trust concerns. Qualitative studies focused on AI knowledge, usage barriers, benefits, and facilitators.

This literature review examined how demographic factors, trust, and knowledge impact patient perspectives on integrating AI in healthcare. The obtained results highlighted the need for educational initiatives to address knowledge gaps and facilitate the smooth integration of AI-powered solutions, leveraging their potential to enhance patient care and service delivery.

End user opinions are crucial for the success of health applications, particularly in the emerging field of artificial intelligence (AI) in medicine. Understanding end users' perspectives is essential for the acceptance and effectiveness of AI.AimThis systematic review aims to comprehensively analyze existing literature on end users' perspectives and acceptance models for AI applications. By synthesizing and critically evaluating research, this review seeks to identify key themes, methodologies, and knowledge gaps.

A systematic review was conducted in PubMed in 2023 to identify relevant peer-reviewed articles written in English. Inclusion criteria focused on original studies that validated assessment AI models from users' perspectives. Information extracted included publication details, countries of research, participant characteristics, data gathering and analysis methods, and attributes of the proposed models.

Out of 3714 records, 19 papers were included in the study that were published between 2019 and 2022. Participants belonged to six categories: physicians, medical students, nurses, patients, and general public. The most important assessed factors in identified papers were “ethical issues, trust, and anxiety”, “usability”, “self-efficacy and knowledge”, “social”, “benefits”, “quality of the AI products and service support”, “AI acceptance, resistance of AI, attitude, and satisfaction” were explored. In addition, the commonly examined several moderating variables, including perceived ease of use, perceived usefulness, and perceived risks.

The findings contribute to understanding current trends and practices in end users' perspective research. Future studies should continue exploring end users' perspectives to enhance the development and implementation of effective AI systems in healthcare.

Over the last few decades, several theoretical frameworks have been proposed to evaluate electronic health records (EHRs). These frameworks provide a theoretical basis for assessing the impact and outcomes of technology adoption in healthcare settings. This can help identify areas for improvement and ensure that EHRs effectively support healthcare delivery and patient care.

The purpose of this study is to present a comprehensive review of the use of validation methods in Electronic health records.

Out of a total of 62 EHR evaluation frameworks in our previous literature review, at the final stage, 34 relevant articles were included for analysis. Variables such as participants and study setting, analysis software, data gathering methods, missing data, and outlier handling, theoretical basis models to develop the EHR evaluation model, the relationship between variables of the EHR evaluation models with evaluation items, sampling technique and sample size reliability assessment methods and values, and statistical validation methods and criteria values were extracted.

Among the 34 papers that disclosed the validation methods utilized, the most widely used technique was Structural Equation Modeling (SEM), employed in 26.5% of the studies. Other methods utilized were Confirmatory Factor Analysis (CFA), and Exploratory Factor Analysis (EFA). A reliability assessment was performed in 82% of the articles. Cronbach's alpha and composite reliability (CR) were popular reliability (internal validation) methods on identified papers.

It is our belief that the results of this study can assist researchers in examining and modifying EHR evaluation frameworks to suit their specific needs. Additionally, we believe that our findings serve as a solid foundation for the creation of new EHR evaluation frameworks. Furthermore, we recommend that researchers utilize the findings presented in this article to enhance the implementation and utilization of SEM, CFA, and EFA methods in EHR evaluation models.

This study is focused on common theories that influenced the history of program evaluation and introduce the educational program evaluation proposal format based on the updated theory. Literature searches were carried out in March-December 2010 with a combination of key words, MeSH terms and other free text terms as suitable for the purpose. A comprehensive search strategy was developed to search Medline by the PubMed interface, ERIC (Education Resources Information Center) and the main journal of medical education regarding current evaluation models and theories. We included all study designs in our study. We found 810 articles related to our topic, and finally 63 with the full text article included. We compared documents and used expert consensus for selection the best model. We found that the complexity theory using logic model suggests compatible evaluation proposal formats, especially with new medical education programs. Common components of a logic model are: situation, inputs, outputs, and outcomes that our proposal format is based on. Its contents are: title page, cover letter, situation and background, introduction and rationale, project description, evaluation design, evaluation methodology, reporting, program evaluation management, timeline, evaluation budget based on the best evidences, and supporting documents. We found that the logic model is used for evaluation program planning in many places, but more research is needed to see if it is suitable for our context.