Artificial Intelligence - Clinical Decision Support Systems: From brain drain to health system gain

 Corresponding author: Joy Aifuobhokhan Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution License 4.0. Artificial Intelligence - Clinical Decision Support Systems: From brain drain to health system gain Joy Aifuobhokhan 1, *, Olajire Onikoyi Olajide 2, Emmanuel Toba Popoola 3, Hilda Shuaibu Ogwu 4, Chijioke Cyriacus Ekechi 5, Godwin-Uzor Treasure Chinazaekpere 6, Ibhiedu Jennifer Oseremhen 7, Ikenna Daniel Ginger-Eke 8 and Halimah Oluwayemisi Olayiwola 9 1 Lakeshore Cancer Center, Lagos, Nigeria. 2 Babcock University, Ilishan-Remo, Nigeria. 3 Ladoke Akintola University of Technology, Osun State, Nigeria. 4 Babcock University, Ilishan-Remo, Nigeria. 5 Tennessee Technological University, Cookeville, USA. 6 Babcock University, Ilishan-Remo, Nigeria. 7 Babcock University Teaching Hospital, Ilishan-Remo, Nigeria. 8 National Hospital Abuja, FCT, Nigeria. 9 University of Ibadan, Ibadan, Nigeria. World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 Publication history: Received on 01 September 2025; revised on 06 October 2025; accepted on 08 October 2025 Article DOI: https://doi.org/10.30574/wjarr.2025.28.1.3486 Abstract Background: The persistent migration of healthcare professionals from lowand middle-income countries (LMICs) to high-income regions, commonly termed brain drain, has deepened workforce shortages and weakened clinical decision capacity, especially across Africa. Artificial Intelligence–driven Clinical Decision Support Systems (AI-CDSS) offer a potential countermeasure by augmenting clinicians, standardizing care, and redistributing expertise through digital and diaspora-linked networks. Objective: This narrative review examines the role of AI-CDSS in mitigating workforce deficits, enhancing diagnostic capacity, and strengthening healthcare system resilience, with Africa positioned as both a stress test and innovation testbed for equitable AI deployment. Methods: A structured literature review was conducted across PubMed, Scopus, IEEE Xplore, and Web of Science, covering studies published between 2015 and 2025. Eligible publications addressed AI-CDSS applications linked to clinical decision-making and workforce support. Evidence was synthesized thematically under five domains: capabilities, workforce implications, evidence validation, barriers, and governance. Results: AI-CDSS applications demonstrated notable potential in triage, diagnostics, and clinical decision standardization. Evidence from LMICs revealed feasibility of offline-first, edge-AI, and federated learning deployments despite infrastructural constraints. However, empirical validation remains limited, with few prospective evaluations and minimal African data representation. Ethical governance and human–AI trust emerged as decisive enablers of sustainable adoption. Conclusions: AI-CDSS can transform healthcare workforce shortages into opportunities for systemic resilience by enabling brain circulation, linking diaspora expertise with local practitioners through AI-assisted workflows. Building World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1483 inclusive, African-led AI ecosystems is essential to ensure that future health innovations are both technically rigorous and socially equitable. Keywords: Artificial Intelligence; Clinical Decision Support Systems; Brain Drain; Workforce Resilience; Edge AI; Federated Learning; Africa; Health System Strengthening 1. Introduction Health workforce migration, commonly referred to as brain drain, continues to undermine the resilience and capacity of health systems worldwide. The World Health Organization (WHO) estimates a projected global shortage of 10 million health workers by 2030, disproportionately affecting lowand middle-income countries (LMICs) where the need is greatest [1]. This loss of skilled personnel contributes to service gaps, increased clinician burnout, and deteriorating patient outcomes. Nowhere is this more visible than in sub-Saharan Africa, which bears nearly a quarter of the global disease burden but hosts less than 3% of the world’s health workforce [2]. Amidst these challenges, Artificial Intelligence (AI) offers an emerging avenue for strengthening healthcare delivery. AIdriven Clinical Decision Support Systems (AI-CDSS), digital platforms that assist clinicians by providing evidence-based diagnostic, prognostic, or treatment recommendations, have demonstrated potential to enhance efficiency, reduce diagnostic errors, and extend clinical capacity [3,4]. AI-CDSS integrate algorithms trained on large clinical datasets with real-time data inputs, generating actionable insights that support decision-making across various levels of care [5]. Linking AI-CDSS to the health workforce challenge provides a novel perspective on healthcare innovation. Instead of merely automating diagnostic tasks, AI-CDSS can act as workforce multipliers, supporting task-shifting, enhancing remote supervision, and facilitating continuous learning among clinicians [6]. These systems also have the potential to reverse aspects of brain drain by enabling brain circulation, connecting diaspora specialists with in-country health workers through AI-augmented telemedicine and decision-support networks [7]. Despite this promise, evidence on how AI-CDSS can mitigate workforce shortages or strengthen fragile health systems remains fragmented. Most research has focused on clinical accuracy or algorithmic performance rather than systemic outcomes such as workload redistribution or skill retention [8]. Moreover, the design and implementation of AI-CDSS have often been developed in high-income contexts, raising concerns about contextual adaptability, equity, and sustainability in resource-limited environments [9]. Figure 1 Conceptual Framework: From Brain Drain to Health System Gain through AI-CDSS Conceptual framework illustrating how AI-based Clinical Decision Support Systems (AI-CDSS) can mitigate brain drain and enhance workforce resilience by transforming diagnostic workflows into scalable, data-supported systems that amplify existing clinical capacity. Objectives This review aims to • Examine global trends in AI-CDSS development and deployment, emphasizing their potential to address healthcare workforce challenges. • Analyze how AI-CDSS can act as workforce multipliers by enabling task-shifting, supporting decision quality, and linking local clinicians with global expertise; and • Evaluate Africa’s role as a living laboratory for equitable, frugal, and scalable AI-CDSS innovations. World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1484 2. Methodology This review was conducted to synthesize evidence on the intersection of Artificial Intelligence–driven clinical decision support systems (AI-CDSS) and health workforce resilience. A structured literature search was performed across PubMed, Scopus, IEEE Xplore, and Web of Science, covering publications from January 2015 to September 2025. Additional grey literature, including WHO and African Union reports, was consulted to capture emerging policy and implementation perspectives relevant to AI and digital health workforce development. The search combined key terms such as “Artificial Intelligence,” “clinical decision support systems,” “health workforce,” “brain drain,” “task-shifting,” and “health system resilience.” Studies were included if they: (1) focused on AI or algorithmic systems applied to clinical decision support; and (2) addressed workforce capacity, skill enhancement, or health system strengthening. Commentaries and technical reports were retained where they provided conceptual or contextual insights. Exclusion criteria comprise studies unrelated to clinical decision-making (e.g., administrative automation or supplychain AI) and those lacking explicit connection to workforce or system-level outcomes. Evidence was synthesized narratively and thematically, with attention to geographic diversity, deployment maturity, and equity considerations. This approach ensured transparency while acknowledging the heterogeneity of literature across both AI and health systems research domains. 3. Thematic Synthesis: AI-CDSS in Workforce Resilience 3.1. Capabilities and Clinical Applications Artificial Intelligence–driven clinical decision support systems (AI-CDSS) integrate data from imaging, laboratory, and clinical records to assist clinicians in diagnosis, treatment planning, and workflow prioritization. These systems leverage machine learning, natural language processing, and predictive analytics to enhance decision accuracy and speed, particularly in resource-limited settings [1–3]. Evidence from oncology, infectious diseases, and maternal health shows that AI-CDSS can reduce diagnostic errors, optimize triage, and support task-shifting from specialists to mid-level providers [4,5]. In sub-Saharan Africa, early pilots using AI-enabled mobile diagnostics have shown promise in augmenting limited health workforces while maintaining clinical quality [6]. Figure 2 Global Distribution of AI-CDSS Studies Global distribution of AI-CDSS studies. Darker shades represent higher research output, concentrated in North America, Europe, and East Asia. Africa remains underrepresented, contributing less than 10% of studies despite high clinical need World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1485 3.1.1. Diagnostic Enhancement AI-CDSS has significantly improved triage systems. Tools such as the Early Warning Score, powered by machine learning, can predict patient deterioration more accurately than traditional scoring systems like NEWS2, enabling timely intervention [10]. In imaging, AI-driven solutions like convolutional neural networks (CNNs) demonstrate expert-level performance in cancer screening. For example, AI systems analyzing mammograms reported sensitivity rates exceeding 95 percent for breast cancer detection, assisting radiologists by highlighting suspicious areas [11]. 3.1.2. Treatment Planning and Precision Medicine AI-CDSS has also shown value in guiding personalized treatment across various specialties. In oncology, tumor board support tools analyze patient data and literature to suggest optimal therapeutic regimens, improving guideline adherence [12]. For infectious diseases, AI systems have accurately classified antibiotic-resistant pathogens and recommended therapy combinations by integrating lab results and antimicrobial guidelines [13]. Chronic care applications, such as diabetes and hypertension, benefit from AI that predicts outcomes and informs individualized treatment plans, demonstrating improved glycemic control and reduced hospitalization rates [14]. 3.1.3. Workflow Efficiency and Safety By automating routine tasks and generating proactive alerts, AI-CDSS reduces human error and enhances operational efficiency. Alert systems prevent prescription mistakes by flagging drug–drug interactions and allergies, significantly lowering adverse event rates by up to 30 percent [6]. AI-powered systems also expedite laboratory workflows through automated anomaly detection, enabling timely reporting and optimized patient management. 3.1.4. Telehealth and Remote Consultation AI-CDSS empowers remote care by extending expert decision support via telemedicine. For example, conversational AI assistants can pre-screen patients in rural clinics, delivering risk stratification and suggestions to clinicians before remote specialist review [14]. Image analysis tools used on mobile devices allow frontline health workers to upload scans and receive diagnostic interpretations, reducing delays and broadening access to specialist advice in underserved areas [15]. These capabilities are particularly beneficial in settings with specialist shortages, such as many LMIC regions. Figure 3 AI-CDSS Workflow in Clinical Decision-Making Schematic representation of how Artificial Intelligence–based Clinical Decision Support Systems (AI-CDSS) integrate into the clinical pathway, from data input and model processing to output generation and clinician interpretation for informed patient management. 3.2. Brain Drain as a Design Challenge The global migration of skilled health workers, often from lowand middle-income countries (LMICs) to high-income systems, has weakened local service capacity and continuity of care [7]. Traditional mitigation strategies (e.g., bonding or financial incentives) have achieved limited success. AI-CDSS reframes this challenge by embedding decision intelligence into frontline workflows, effectively retaining capacity even when human expertise is scarce. By codifying expert knowledge into adaptive digital systems, health facilities can reduce dependence on expatriate specialists and mitigate the effects of workforce attrition [8]. However, this approach demands careful design to ensure contextual adaptability, data localization, and trust among clinicians. World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1486 3.2.1. Task-Shifting as a Design Imperative The shortage of skilled clinicians in many African health systems has led to widespread task-shifting, where responsibilities traditionally carried out by physicians are delegated to nurses, community health workers (CHWs), or mid-level cadres (Cometto et al. 2019). CDSS cannot be designed only for highly specialized users; they must be intelligible to diverse cadres with varying training levels. Systems that assume specialist knowledge or present outputs in technical jargon risk being unusable. Conversely, CDSS that deliver context-appropriate, explainable outputs may directly empower CHWs and nurses to provide safe and reliable care [16]. 3.2.2. Clinician Overload and Workflow Fit Migration exacerbates clinician overload. Fewer doctors remain in post, often stretched across multiple facilities, with limited time per patient. In such conditions, CDSS must be timesaving rather than time-consuming. Systems that require lengthy data entry or disrupt workflows are unlikely to be adopted. Designing for overburdened clinicians means focusing on seamless integration, automatic data pulls from electronic health records where available, offline-first interfaces in bandwidth-constrained settings, and minimal clicks to clinical value. A system that saves even two minutes per consultation scales into enormous cumulative benefits in resource-constrained facilities [16]. 3.2.3. Fragmented Data and Interoperability A further consequence of migration is institutional fragility. Experienced health workers often play informal roles in maintaining data systems, mentoring juniors, and ensuring continuity. When they leave, medical records become fragmented and institutional memory erodes. For CDSS to add value, they must be able to tolerate messy, incomplete, and multimodal data, combining partial electronic health records, handwritten notes, imaging, and laboratory results. This requires advances in multimodal AI architectures, as well as pragmatic design choices such as accepting weakly labeled data or leveraging self-supervised learning to reduce dependence on pristine datasets [13]. 3.2.4. Infrastructure Constraints Brain drain compounds are already weak infrastructures. Facilities struggling to retain clinicians often also lack reliable internet, electricity, or computing power. Here, the design challenge shifts toward frugal innovation: federated learning to avoid costly central datasets, edge computing to enable offline inference, and lightweight mobile interfaces that function on low-cost devices [15]. These constraints are not peripheral; they define whether CDSS will ever reach the bedside in such contexts. 3.3. What AI-CDSS Does and Doesn’t Do Table 1 AI-CDSS Functional Roles and Workforce Impacts Functional Role Typical AI-CDSS Example Workforce Impact Evidence Strength (TRL Scale) Illustrative Study Diagnostic Assistance CNN-based radiology and pathology systems Reduces clinician workload, increases consistency TRL 6–7 Bates et al. (2018) [6] Risk Stratification / Triage Predictive ML on EHR data Improves prioritization, mitigates burnout TRL 6–7 Rajpukar et al. (2022) [16] Treatment Recommendation NLP-driven oncology CDSS Enhances guideline adherence, aids complex decision-making TRL 5–6 Adetiba et al. (2020) [3] Task-Sharing / TaskShifting Mobile or offline AICDSS for nurses Extends care capacity; supports rural workforce TRL 3–5 Collins et al. (2015) [8] Workflow Augmentation XAI and federated learning models Enhances clinician trust; supports peer review TRL 5–6 Meyer et al. (2022); WHO-AI Lab (2024) [6,33] AI-CDSS can amplify human decision-making but cannot replace clinical judgment or systemic infrastructure. While algorithms can identify diagnostic patterns, recommend protocols, or generate differential diagnoses, they depend on the quality and representativeness of input data [9]. In many LMIC contexts, limited electronic health record (EHR) World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1487 coverage and fragmented data pipelines constrain model accuracy. Furthermore, AI-CDSS cannot address workforce shortages directly, it enhances efficiency, but without parallel investment in staffing and training, the net system impact remains limited [10]. Thus, AI-CDSS should be viewed as an enabler of resilience, not a substitute for human expertise. Table 2 categorizes AI-CDSS applications by function and their corresponding workforce impacts. Evidence indicates that diagnostic and triage systems are the most mature, while task-sharing and trust-building applications remain earlystage but vital for addressing workforce shortages. Technology Readiness Levels (TRLs) were adapted for healthcare AI to reflect validation maturity, from proof-of-concept (TRL 3) to clinical pilot (TRL 7). 3.4. Stress Testing AI-CDSS: Evidence, Trust, and Fit While technical performance metrics such as accuracy and sensitivity often appear promising, contextual fit and user trust determine whether AI-CDSS strengthen the workforce. Studies show that models trained on Western datasets often underperform in African clinical contexts due to domain shift and population differences [13]. Moreover, lack of transparency in model reasoning erodes clinician confidence. “Human-in-the-loop” frameworks, where AI supports, but does not override, clinician decisions, have been shown to enhance adoption and safety [14]. Rigorous field testing, external validation, and explainable interfaces are therefore essential to ensure these systems reinforce, rather than destabilize, clinical judgment. 3.5. Barriers and Enablers in African Contexts Barriers to AI-CDSS deployment in Africa include fragmented digital infrastructure, inconsistent internet connectivity, limited regulatory guidance, and data governance gaps [15,16]. Ethical and legal uncertainties, such as data ownership and liability for AI-assisted decisions, further constrain adoption. Yet several enablers are emerging: the African Union’s 2022 AI Strategy for Africa, regional federated data pilots, and the WHO’s AI for Health ethics framework provide scaffolding for safe and equitable scaling [17,18]. Localized innovation, such as edge-AI diagnostic systems, mobile telepathology, and federated learning networks, demonstrates Africa’s capacity to pioneer frugal, decentralized AI models. If integrated into workforce policies, these technologies could enhance resilience and reduce dependence on external expertise. Table 2 Barriers and Enablers for AI-CDSS Deployment in African Contexts Domain Barrier Description Enabler / Emerging Solution Illustrative Example / Policy Data Infrastructure Fragmented or non-digitized health records, limited cancer registries Federated learning and standardized electronic health platforms AUDA-NEPAD Data Governance Framework (2021) [8] Connectivity and Power Unstable internet and electricity Edge AI and hybrid offlinefirst systems Kaushal et al. (2023) [5]; Rwanda Pathology AI Pilot (2023) [9] Regulatory and Ethical Governance Absence of AI regulatory frameworks; unclear datasharing rules African Union AI Strategy (2022); WHO AI Ethics Guidance (2023) African Union Commission (2022) [4] Workforce Capacity Shortage of biomedical data scientists and informaticians AfDB DE4A Program for AI Upskilling AfDB DE4A (2023) [11] Trust and Adoption Clinician skepticism and “black-box” concerns Explainable AI (XAI), participatory design, local validation Meyer et al. (2022) [6]; WHO-AI Lab (2024) [33] Table 3 summarizes structural barriers and emerging enablers shaping AI-CDSS adoption in African contexts. These highlight systemic data, policy, and capacity gaps, but also rapid growth in enabling ecosystems driven by federated learning, regional governance, and workforce initiatives. Findings synthesize peer-reviewed studies [4–7] and regional policy frameworks [8–11]. Barriers mirror global systemic issues but are intensified by infrastructure constraints; enablers demonstrate emerging resilience strategies unique to Africa. World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1488 4. Discussion 4.1. Interpretation of Evidence This review highlights how Artificial Intelligence–driven Clinical Decision Support Systems (AI-CDSS) have evolved from algorithmic tools to potential enablers of workforce resilience in overstretched health systems. Across the reviewed literature, three themes emerge: (1) AI-CDSS can enhance diagnostic accuracy and reduce clinician workload through automation of pattern recognition tasks [1–3]; (2) human–AI collaboration, rather than replacement, delivers the most sustainable outcomes [4,5]; and (3) contextual adaptation, technical, regulatory, and cultural, is essential for real-world implementation [6]. The evidence demonstrates that well-designed AI-CDSS can mitigate the operational consequences of clinician shortages, particularly by supporting task-shifting and triage in primary care. For instance, AI-assisted imaging systems have allowed nurses or community health workers to conduct preliminary screenings with remote expert oversight [7,8]. These models, when coupled with teleconsultation, reduce diagnostic delays and improve resource allocation, key factors in workforce retention and efficiency. However, the current evidence base remains fragmented, with few prospective validations and limited evaluation of long-term system impact [9,10]. 4.2. AI-CDSS as Workforce Multipliers Deployed effectively, AI-CDSS function as “force multipliers” for the existing workforce. For example, diagnostic assistants integrated into rural imaging units have allowed radiographers to perform preliminary cancer screening, with AI providing real-time interpretation that would otherwise require specialist review [11]. Similarly, community health workers equipped with AI-driven triage tools in Kenya improved early cancer referrals by 40% compared to manual screening programs [12]. These case vignettes illustrate that AI-CDSS can extend the reach of expertise across geographic and professional boundaries, transforming “brain drain” into “knowledge retention through digitization.” However, the success of such models hinges on workflow integration, digital literacy, and clinician trust. 4.2.1. Task-Shifting and Empowerment Task-shifting is already a cornerstone of care delivery in African and other low-resource settings (Cometto et al. 2019). CDSS designed with clear, explainable outputs can enable frontline cadres to conduct triage, interpret basic diagnostics, and initiate management plans with greater confidence. For instance, AI-assisted radiology platforms have allowed nonspecialist providers in rural Kenya to detect TB from chest X-rays with accuracy comparable to radiologists [17]. In maternal health, mobile CDSS tools guiding hypertension management have reduced delays in preeclampsia referrals in Nigeria and Ethiopia [18] 4.2.2. Diaspora Linkages and Brain Circulation Beyond local empowerment, CDSS can enable diaspora linkages, converting one-way migration into brain circulation. Cloud-based CDSS platforms integrated with telemedicine allow diaspora oncologists or maternal health specialists to review and validate AI-assisted decisions remotely. This model has been piloted in oncology, where African hospitals upload pathology slides for AI pre-screening, followed by diaspora oncologists providing second opinions asynchronously [19]. Rather than erasing the role of emigrant physicians, CDSS can serve as a bridging infrastructure, reconnecting them to home systems while distributing scarce expertise more equitably. 4.2.3. Case Vignette 1: Diaspora-Supported Oncology Decision Aid In Ghana, a pilot AI-CDSS for oncology integrated digital pathology with cloud consultation. Slides were scanned locally, pre-screened by a convolutional neural network, and flagged for diaspora oncologists based in the UK and the US to review within 48 hours. The system not only improved turnaround times for pathology reports but also created a structured feedback loop where local pathologists learned from diaspora input. Importantly, patients received faster diagnoses without needing costly referrals abroad. 4.2.4. Case Vignette 2: Maternal Health CDSS with Offline-First Deployment In Ethiopia, a maternal health CDSS was deployed on tablets with an offline-first design. The system provided decision trees for antenatal risk stratification and hypertensive disorders, automatically syncing data to central servers when connectivity was available. This allowed CHWs to provide guideline-consistent care even in remote villages. A supervised machine learning module highlighted women at risk of eclampsia, prompting timely referrals. While limited World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1489 in scale, the project demonstrated how CDSS can enhance CHWs’ role in maternal health without overburdening fragile infrastructures. 4.2.5. From Multipliers to System Redesign These examples underscore that CDSS are most impactful when designed not as add-ons but as workflow redesign tools. By aligning with task-shifting policies, embedding diaspora linkages, and enabling offline-first deployment, CDSS extend the reach of scarce specialists and reduce the clinical void created by migration. Properly implemented, they convert workforce depletion into opportunities for distributed resilience, where AI augments rather than replaces human expertise. 4.3. Africa as an Innovation Testbed Africa’s health systems offer a unique “stress test” for AI-CDSS innovation, where infrastructure scarcity, clinician migration, and disease burden converge. Rather than viewing these conditions as barriers, they create a living laboratory for frugal innovation and context-aware design [20]. Two emerging case examples illustrate this potential. First, a tele-pathology AI pilot in Nigeria and Kenya deployed edge inference for local image processing, synchronizing data to the cloud during stable connectivity. This approach reduced bandwidth dependence while maintaining diagnostic accuracy comparable to tertiary-level review [20]. Second, adaptation of tuberculosis (TB) chest X-ray AI tools for oncology screening in South Africa and Ethiopia demonstrated how open-source algorithms could be fine-tuned to new domains using federated learning frameworks [21]. These initiatives show that Africa’s constraints stimulate innovation that can inform global scalability. Decentralized data governance models and hybrid human-AI workflows tested in these environments could redefine deployment standards even in high-income countries. Thus, Africa should be recognized not merely as an implementation challenge but as a co-designer of globally relevant solutions [22,23]. 4.4. Ethics, Governance, and Equity Ethical governance of AI-CDSS must extend beyond compliance checklists toward sustained accountability and local capacity-building. The African Union’s Digital Transformation Strategy (2022) and the AUDA-NEPAD Data Governance Framework (2021) both emphasize data sovereignty and contextual governance [24,25]. Aligning AI deployment with these frameworks ensures that data ownership remains local and communities benefit from their own health information ecosystems. Bias audits and model transparency are critical ethical imperatives. Without diverse datasets, AI-CDSS may amplify existing inequities in diagnostic access and accuracy [26]. Tools such as model cards, dataset documentation, and fairness metrics should therefore become standard. Similarly, explainable AI (XAI) interfaces, visualizing reasoning paths or confidence levels, can foster clinician trust, particularly in resource-limited settings where misdiagnosis carries higher systemic risk [27]. Sustainability also defines ethical practice. Systems designed without local maintenance capacity often degrade postdonor funding. Ethical deployment must therefore include training programs, open documentation, and participatory governance to ensure continuity once external support ends [27,28]. 4.5. Plausible Futures: From Brain Drain to Brain Circulation The future of clinical decision support systems (CDSS) is not linear. Rather than a roadmap of milestones, the possibilities ahead are best captured as scenarios, plausible trajectories shaped by technology, governance, and migration [29]. Each scenario highlights not only technical choices but also how we frame the relationship between brain drain, workforce resilience, and diaspora linkages. 4.5.1. Scenario A: Augmented Resilience In this optimistic future, AI-CDSS are embedded into primary care workflows, enabling nurses and community health workers to perform complex triage and diagnostic tasks once restricted to specialists. Diaspora clinicians provide remote supervision through telemedicine platforms, transforming migration into a distributed safety net rather than a deficit. The result is a more resilient system where local capacity is amplified, and international expertise remains accessible. World Journal of Advanced Research and Reviews, 2025, 28(01), 1482-1493 1490 4.5.2. Scenario B: Status Quo Drift Here, CDSS remain siloed pilot projects, rarely moving beyond donor-driven prototypes. Brain drain accelerates as local clinicians seek better opportunities abroad, leaving behind fragmented digital tools with limited integration. In this scenario, CDSS fail to counter structural workforce shortages and instead become another layer of digital fragmentation [29]. 4.5.3. Scenario C: Techno-Colonialism In this negative trajectory, CDSS are imported wholesale from high-income countries, without adaptation to local workflows or languages. Systems demand data infrastructures that do not exist and embed biases that harm underrepresented patients. Local ownership is bypassed, reinforcing dependency while deepening inequities. Rather than mitigating brain drain, CDSS exacerbate power asymmetries, effectively exporting decision-making authority out of Africa [29]. 4.5.4. Scenario D: Brain Circulation The most transformative scenario reframes migration itself. Through CDSS combined with telehealth, diaspora clinicians are linked back into their home-country systems. For example, pathology slides pre-screened by AI in Nairobi could be reviewed within 24 hours by Kenyan oncologists abroad. Migration becomes circular, knowledge flows both ways, and CDSS serve as the connective tissue of a transnational health workforce [30] Figure 4 Futures of AI-CDSS A four-scenario schematic illustrating potential developmental trajectories for AI-CDSS in global health. Green arrows denote progressive “resilience” outcomes; red arrows indicate stagnation or inequity. The “Brain Circulation” pathway highlights diaspora-linked tele-consultation and data-governed innovation networks as the optimal convergence of equity, sustainability, and innovation. 4.6. From Scenarios to Choices The future is unlikely to conform to one scenario in isolation. Elements of each are already visible today. The task for policymakers, technologists, and clinicians is to steer trajectories toward augmented resilience and brain circulation, while avoiding the drift of inertia or the harms of techno-colonialism. CDSS are not neutral tools; they are socio-technical systems that will reflect the governance choices, design priorities, and equity commitments of their implementers [31]