Healthcare Generative Artificial Intelligence Tools in Medical Diagnosis, Treatment, and Prognosis.

• Author: Horak, Jakub

1. Introduction

   Generative artificial intelligence algorithms (Barnes, 2022; Nica et al, 2023; Peters et al, 2023; Valaskova et al, 2022) can configure prevention strategies, personalized healthcare planning and medical decisions, and treatment options. The purpose of our systematic review is to examine the recently published literature on healthcare generative artificial intelligence tools and integrate the insights it configures on medical diagnosis, treatment, and prognosis. By analyzing the most recent (2023) and significant (Web of Science, Scopus, and ProQuest) sources, our paper has attempted to prove that ChatGPT can assist in medical knowledge, diagnosis, treatment, and prognosis, improve case report writing and patient outcomes, analyze large medical datasets, and forecast side effects. The actuality and novelty of this study are articulated by addressing how ChatGPT can be deployed in clinical decision support and practice, producing precise differential diagnosis lists and providing reliable disease and medical query-related information with regard to real-time monitoring and personalized treatment, that is an emerging topic involving much interest. Our research problem is whether ChatGPT-generated detailed and streamlined data analysis (Andronie et al, 2021; Krizanova et al, 2019; Popescu et al, 2017) can harmonize healthcare practice and patient experience with optimized medical reporting, diagnostics, and treatment plans by precise and thorough medical and treatment information.

   In this review, prior findings have been cumulated indicating that ChatGPT can act as an interactive medical education tool in terms of biomedical, clinical, epidemiological knowledge, and reasoning, shaping patient care and supporting related problem-solving and reflective practice (Barber, 2022; Nica, 2017; Popescu, 2018; Vatamanescu et al, 2020) in an interactive learning environment. The identified gaps advance how generative artificial intelligence algorithms (Andronie et al, 2023; Lazaroiu, 2018; Popescu et al., 2017) can shape clinical and translational medicine and practice and standardized medical procedures. Our main objective is to indicate that generative artificial intelligence tools (Barbu et al, 2021; Nica, 2018; Rowland, 2022; Vinerean et al, 2022) enable medical decision-making and analyze patient data, resulting in better patient outcomes and in enhanced access to care.

2. Theoretical Overview of the Main Concepts

   ChatGPT can expedite disease prevention measures and therapeutic design and simulate virtual interactions between healthcare professionals and patients. Generative artificial intelligence algorithms can configure multimodal medical image fusion and post-treatment care, resulting in precision medicine and personalized healthcare implementation by specific healthcare recom-mendations. Generative artificial intelligence tools optimize patient education, enhance medical diagnosis and treatment, and provide personalized care and advice. Medical information intricacy and individual patient case specific character may lead to misinterpretation generated by ChatGPT, and thus to misguided self-diagnosis, erroneous decision-making, existing condition aggravation, delayed treatment, and adverse health outcomes. The manuscript is organized as following: theoretical overview (section 2), methodology (section 3), ChatGPT can provide efficient and accurate health education (section 4), ChatGPT can assist in medical knowledge, diagnosis, treatment, and prognosis (section 5), ChatGPT can provide precise and thorough medical and treatment information (section 6), discussion (section 7), synopsis of the main research outcomes (section 8), conclusions (section 9), limitations, implications, and further directions of research (section 10).

3. Methodology

   Throughout March 2023, we performed a quantitative literature review of the Web of Science, Scopus, and ProQuest databases, with search terms including "healthcare generative artificial intelligence tools" + "medical diagnosis," "medical treatment," and "medical prognosis." As we inspected research published in 2023, only 174 articles satisfied the eligibility criteria. By eliminating controversial findings, outcomes unsubstantiated by replication, too imprecise material, or having similar titles, we decided upon 38, generally empirical, sources (Tables 1 and 2). Data visualization tools: Dimensions (bibliometric mapping) and VOSviewer (layout algorithms). Reporting quality assessment tool: PRISMA. Methodological quality assessment tools include: AMSTAR, Distiller SR, MMAT, and ROBIS (Figures 1-6).

4. ChatGPT Can Provide Efficient and Accurate Health Education

   ChatGPT can be deployed in clinical decision support and practice, producing precise differential diagnosis lists and providing reliable disease and medical query-related information with regard to real-time monitoring and personalized treatment (Gilson et al., 2023; Liu et al., 2023; Thurzo et al., 2023; Vaishya et al., 2023), being decisive in remote healthcare through predictive analytics. ChatGPT can act as an interactive medical education tool in terms of biomedical, clinical, epidemiological knowledge, and reasoning, shaping patient care and supporting related problem-solving and reflective practice in an interactive learning environment.

   Responses produced by ChatGPT should be checked and enhanced, and personalized study plans, detailed recommendations, and learning materials be integrated through guidance and critical thinking, expert-reviewed content, and specific learning needs (Abd-alrazaq et al., 2023; Cadamuro et al., 2023; Morath et al., 2023; Sallam et al., 2023) to avoid inaccuracies and misinterpretations in medical education. Generative artificial intelligence tools optimize patient education, enhance medical diagnosis and treatment, and provide personalized care and advice.

   ChatGPT can provide efficient and accurate health education and enable home self-care, furthering virtual healthcare professional--patient interactions (Corsello and Santangelo, 2023; Das et al., 2023; Giannos and Delardas, 2023; Liu et al., 2023; Rahsepar et al., 2023) through remote assistance and guidance due to intelligent extraction, summarization, data exchange, analysis, and interoperability, in addition to unstructured clinical note information extraction. ChatGPT can expedite disease prevention measures and therapeutic design and simulate virtual interactions between healthcare professionals and patients. (Table 3)

5. ChatGPT Can Assist in Medical Knowledge, Diagnosis, Treatment, and Prognosis

   ChatGPT's advice is intelligible for users having no specialized medical knowledge, but patients may erroneously construe the recommendations and thus self-diagnose or self-administer treatment (Altamimi et al., 2023; Ge and Lai, 2023; Meo et al., 2023; Shahsavar and Choudhury, 2023; Temsah et al., 2023), resulting in postponement in identifying professional assistance or in receiving deficient care. Medical information intricacy and individual patient case specific character may lead to misinterpretation generated by ChatGPT, and thus to misguided self-diagnosis, erroneous decision-making, existing condition aggravation, delayed treatment, and adverse health outcomes.

   Patients can use annotated dataset-based fine-tuned ChatGPT to ask healthcare practitioners meaningful questions with regard to medical report results and complex medical conditions (Elkassem and Smith, 2023; Johnson et al., 2023; Mondal et al., 2023; Sanmarchi et al., 2023), optimizing understanding and engagement with respects to follow-up recommendations. ChatGPT can assist in medical knowledge, diagnosis, treatment, and prognosis, improve case report writing and patient outcomes, analyze large medical datasets, and forecast side effects.

   Generative artificial intelligence algorithms can shape clinical and translational medicine and practice and standardized medical procedures (Baumgartner, 2023; De Angelis et al., 2023; Juhi et al., 2023; Putra et al., 2023) through medical decision support, machine learning-based pattern recognition, medical image assessment for diagnosis and patient...