Page 349 - “O‘ZBEKISTON – 2030 STRATEGIYASI: AMALGA OSHIRILAYOTGAN ISLOHOTLAR TAHLILI, MUAMMOLAR VA YECHIMLAR”

and acceptance of AI-enabled tools. Deep learning models, in particular, often
function as “black boxes,” complicating clinicians’ understanding of underlying
decision processes and limiting trust in AI recommendations.
Another critical issue concerns generalizability across heterogeneous
populations and data environments. Algorithms trained on unbalanced datasets
may produce inconsistent results and reinforce healthcare disparities. Obermeyer et
al. [11] demonstrated the presence of racial bias in clinical AI systems, underscoring
the necessity for equitable data collection, transparent validation, and continuous
algorithm auditing to ensure fairness and inclusivity in AI-supported CDSS.
Legal and ethical compliance also poses major constraints. Adhering to
frameworks such as the Health Insurance Portability and Accountability Act (HIPAA)
and the General Data Protection Regulation (GDPR) is essential for maintaining data
privacy and patient trust. Moreover, the complexity of medical device certification,
software validation, and CDSS authorization processes hinders large-scale
implementation. These regulatory and ethical considerations represent central
challenges for developers and healthcare organizations seeking to deploy AI
solutions safely and responsibly [12].

Figure 1. AI applications across healthcare domains (Author’s elaboration).

AI-powered Clinical Decision Support Systems (CDSS) increasingly enable risk
prediction and preventive intervention, helping clinicians anticipate and manage
potential complications through machine learning (ML) analysis of real-time patient
data.
Choi et al. [13] showed that AI models using electronic health records (EHRs)
accurately identify diabetic patients at risk of cardiovascular disease, allowing timely
preventive action. Likewise, Ryu et al. [14] developed a convolutional neural network
(CNN) that predicts diabetic retinopathy from OCTA images with 91–98% accuracy,
demonstrating AI’s potential to detect early disease markers, prevent irreversible
damage, and improve outcomes.

RESULTS
The study confirms that integrating Artificial Intelligence (AI) into Clinical
Decision Support Systems (CDSS) markedly improves outcomes in diagnosis,
prognosis, and therapy. AI techniques—machine learning (ML), natural language 346

V SHO‘BA:

Raqamli transformatsiya va ilg‘or boshqaruv tizimlarini joriy etish samaradorligi

https://www.asr-conference.com