Background: Clinical decision support systems based on artificial intelligence (AI-CDSS) are being widely adopted in medical specialties to enhance diagnostic precision, streamline workflows, and improve patient outcomes. Though potentially valuable, usability, trust, and integration issues are major hurdles to large-scale adoption. Objective: The objective was to conduct a synthesis of the latest evidence regarding the uses, effects, and impediments of AI-CDSS in multidisciplinary clinical practice, alongside comparing rates of adoption and clinician attitudes between care environments. Methods: A scoping review design was used, and literature searches were undertaken in PubMed, Scopus, and Web of Science between January 2020 and July 2025. Searches included studies reporting on the implementation or evaluation of AI-CDSS in clinical practice in primary care, oncology, emergency medicine, and community health. Data were extracted to a matrix and synthesized thematically. Results: Sixty-two studies were included, most published since 2021, demonstrating increasing interest. AI-CDSS for primary care enhanced diagnostic performance and consultation speed, whereas oncology applications demonstrated improvements in cancer diagnosis and treatment planning. Emergency medicine reported results highlighted the need for expeditious triage and workflow optimization, while community health applications highlighted the optimization of resource allocation. Inhibitors were limited transparency, clinician skepticism, alert fatigue, and integration complexities. Comparative results showed that institutional context, usability of the system, and perceived trust were highly influential for willingness to implement AI-CDSS. Conclusion: AI-CDSS show evident benefits across specialties but need explainable design, clinician buy-in, and compliance with regulation for sustainable incorporation into healthcare systems.
The convergence of artificial intelligence (AI) and clinical decision support systems (CDSS) has emerged as a revolutionary force in numerous fields of medicine. These systems are intended to enhance diagnostic precision, optimize the selection of treatment, and improve patient outcomes by giving clinicians evidence-based information. In primary care settings, AI-enabled CDSS has been seen to demonstrate potential for initial rollout, rationalizing clinical workflow, and enhancing decision-making capacity in low-resource settings [1,7].
Successful adoption, however, involves consideration of both the technological aspects and the human factors that affect clinician interaction with these systems, such as usability, trustworthiness, and interpretability [2]. In oncology, AI technologies such as breast cancer diagnosis show the prowess of machine learning in subspecialty diagnosis with more emphasis on health technology assessment in safety, cost-effectiveness, and ethics [3]. In addition, the broadening of the use of AI in emergency medicine puts the power of large language models in context-dependent, real-time decision support in the acute environment [6].
At international levels, heterogeneity in CDSS adoption suggests the necessity of context-adjusted strategies. For example, Chinese implementation strategies have varying applications and issues, including infrastructure preparedness and integration into existing care pathways [5]. Alongside functionality, the ethical dimension of AI in health care has also received more attention, particularly in facilitating empathic care and ensuring technological advancements are consonant with patient-centered values [4].
While there is evidence that AI-based CDSS have the potential to influence physician performance and patient outcomes in a positive way, more evaluation is needed to address implementation barriers and enhance their optimization in multidisciplinary teams [8]. This emerging literature highlights the potential and challenges of utilizing AI to augment decision-making, the need for a more augmented approach that spans technology, human factors, and systems-level integration.
Study Design
This research employed the scoping review approach to chart the landscape of artificial intelligence-based clinical decision support systems (AI-CDSS) research in multidisciplinary medicine. A scoping review design was employed to map the variety of implementations, settings, and adoption issues of AI, as well as the technical and human aspects influencing adoption.
Data Sources and Search Strategy
A systematic review of the major scientific databases, including PubMed, Scopus, and Web of Science, was undertaken to find peer-reviewed articles from January 2020 to July 2025. The search terms included the combination of keywords artificial intelligence, clinical decision support systems, multidisciplinary medicine, primary care, oncology, emergency medicine, and health technology assessment. The search strategy was regularly updated to enhance sensitivity and specificity, and the references of the relevant articles were also scanned for finding other relevant sources.
Eligibility Criteria
Articles were deemed eligible if they were documenting implementation, evaluation, or conceptual analysis of AI-supported CDSS in clinical use. Primary care, specialty-specific use (e.g., oncology, emergency medicine), or more general multidisciplinary environments were included. Qualitative and quantitative study designs were entertained in an effort to understand system operation, human–technology interface, ethical concerns, and organizational effects. Non-English language publications, commentaries lacking empirical basis, and publications solely dealing with technical model development and not clinically applicable were not included.
Study Selection
All identified records were entered into a reference management system and duplicates removed prior to screening. Titles and abstracts were independently screened against eligibility criteria by two researchers, with full-text evaluation of studies deemed potentially relevant. Disagreements were resolved by consensus and discussion, and a third reviewer consulted where necessary.
Data Extraction and Charting
An evidence framework for the structured extraction of data was developed to identify key information, including study characteristics, clinical context, AI-CDSS characteristics, reported outcomes, and determinants of identified barriers or facilitators to implementation. Extracted data were mapped onto an evidence matrix to enable thematic mapping across multidisciplinary domains.
Data Synthesis
Thematic synthesis was used to merge results across studies included. The procedure involved categorization of evidence into themes of technological capabilities, clinical performance, user interface, ethical issues, and system-level integration. Narrative synthesis was used to reveal emerging patterns, variation by specialty, and contextual factors influencing implementation effects. Synthesis was directed for comparability but permitted heterogeneity in study design and clinical settings.
62 studies that were found to be eligible were included in this review. These studies crossed a variety of clinical settings, ranging from primary care, oncology, emergency care, and more general community-based healthcare. Most of the studies were from Europe and North America, with a growing representation from Asia in recent years, especially from China and India. Trends in publication show a consistent increase in interest in AI-CDSS, with a significant upsurge following 2021.
Table 1. Distribution of Included Studies by Clinical Domain and Geographic Region
Clinical Domain |
Number of Studies |
Major Regions of Origin |
Example Applications |
Primary Care |
18 |
Europe, North America |
Risk stratification, chronic disease support |
Oncology |
14 |
Europe, Asia |
Cancer detection, treatment planning |
Emergency Medicine |
11 |
North America, Europe |
Triage support, rapid decision guidance |
Community Health Care |
9 |
Africa, Asia |
Resource allocation, preventive screening |
Multidisciplinary Use |
10 |
Global |
Integration across multiple specialties |
Across disciplines, AI-CDSS exhibited diverse capabilities. In primary care, diagnostic decision support and management of chronic diseases were the most common functionalities supported by most systems. Oncology software was primarily designed to address early detection and imaging-based diagnosis, with breast cancer detection featured as the most common use case. Acute triage and intervention support was the focus of real-time data stream-based AI integration in emergency medicine. Community-based health care studies brought out the ability of AI in combating inequalities by rationalizing scarce resources.
Figure 1. Trends in Publications on AI-Based CDSS Across Clinical Domains (2020–2025)
Figure 1 showing an upward trend, with oncology and emergency medicine experiencing the sharpest rise after 2022.
Reported results were not uniform across sites. In primary care, AI-CDSS enhanced diagnostic precision and shortened consultation times, whereas oncology systems showed dramatic improvements in sensitivity for the detection of early cancers. In emergency medicine, AI-CDSS facilitated quicker triage decisions and improved workflow efficiency. While these were positive effects, various studies highlighted issues such as transparency of algorithms, trust of clinicians, and integration into current electronic health records.
Table 2. Reported Outcomes of AI-CDSS Across Clinical Domains
Domain |
Positive Outcomes |
Reported Challenges |
Primary Care |
Improved diagnostic accuracy, reduced consultation time |
Usability issues, clinician trust in AI outputs |
Oncology |
Higher sensitivity in early cancer detection, improved treatment planning |
Integration with imaging workflows, regulatory barriers |
Emergency Medicine |
Reduced triage time, improved patient flow |
Real-time data integration, alert fatigue |
Community Health |
Improved preventive screening, better resource allocation |
Limited infrastructure, equity concerns |
Multidisciplinary |
Enhanced collaboration, holistic decision support |
Data interoperability, ethical and legal issues |
Human factors became a persistent driver of adoption success. Systems that prioritized transparency, interpretability, and clinician feedback loops tended to become positively integrated. Systems that did not possess explainability tended to be resisted by clinical practitioners. The tension between technological sophistication and clinician usability was a persistent theme throughout the studies.
Figure 2. Key Factors Influencing Adoption of AI-CDSS
Figure 2 highlighting percentages of studies reporting clinician trust, system usability, transparency, workflow integration, and ethical considerations as critical factors.
The results of this review note the growing application of AI-based clinical decision support systems (AI-CDSS) in various areas of medicine, with significant gains achieved in oncology and emergency medicine as well as nascent integration into primary and community health settings. In general, AI-CDSS exhibited beneficial effects on diagnostic performance, workflow, and patient outcomes, with non-uniform adoption attributed to ongoing issues with usability, trust, and system transparency.
Such findings are consistent with current clinician adoption behaviors evidence. It has been shown that physicians' willingness to adopt AI-CDSS is influenced by institutional setting and available resources, with hospital-level factors significantly influencing adoption readiness (Yu et al., 2025) [9]. Likewise, predictors like performance expectancy, effort expectancy, and perceived trustworthiness were found to significantly determine practitioners' intentions to adopt AI in clinical processes (Dingel et al., 2024) [10]. Our synthesis is consistent with these findings, as studies that stressed interpretability and workflow alignment all had higher clinician acceptance.
Barriers identified in this review, such as integration challenges, alert fatigue, and lack of transparency, are echoed in broader literature. For example, qualitative analyses highlight persistent skepticism among clinicians, particularly when systems operate as “black boxes” without providing clear reasoning for outputs (Giebel et al., 2025; Scipion et al., 2025) [11, 12]. In spite of these obstacles, the literature indicates that clinicians are not resistant to AI per se but that their uptake is context dependent and contingent on systems being able to simplify and not complicate decision-making (Vijayakumar et al., 2023) [14].
Clinical findings published in oncology and diabetes management also attest to the efficacy of AI-CDSS in achieving more accurate diagnosis and therapeutic support. In cancer diagnosis and diabetes management, AI-CDSS have shown enhanced accuracy and consistency in decision-making, highlighting their promise in disease-specific interventions (Huang et al., 2023; Ouanes & Farhah, 2024) [13,15]. These advantages need to be well integrated into current health systems to guarantee scalability and sustainability.
Together, this review concludes that AI-CDSS have vast potential for multidisciplinary care but need balanced development strategies with both technological performance and human–system interaction. Future deployments should emphasize clinician involvement, explainability, and regulatory alignment to promote adoption at scale and optimize patient benefit.
Artificial intelligence-supported clinical decision support systems are changing multidisciplinary medicine at a fast pace by improving diagnostic accuracy, streamlining workflow effectiveness, and aiding patient-centered care in various settings. Although evidence has shown tangible advantages in areas like oncology, emergency care, and primary care, their general acceptance continues to be influenced by technological transparency, trust among clinicians, and system integration. Mitigating these obstacles using explainable design, human-centered development, and policy-based frameworks will be critical to successful implementation. Ultimately, the success of AI-CDSS integration rests on balancing technological advancement with human considerations, ensuring the tools supplement clinical acumen instead of substituting it and have a valuable role to play in the future of healthcare provision.