Brain-drain, the large-scale emigration of skilled professionals from lower-income countries to wealthier ones, is not a new story in West Africa. Ghana has trained nurses, physicians, and allied health professionals for decades only to see many of them recruited by health systems in the United Kingdom, the United States, Canada, and elsewhere in the high-income world. What is new, the researchers argue, is the pace. In the years following the COVID-19 pandemic, the migration of health workers from Ghana to high-income countries has increased to what the authors describe as alarming levels, threatening the country’s ability to achieve universal health coverage, the commitment that every person should be able to obtain the health services they need without financial hardship.

The protocol is explicit about why the pandemic appears to have acted as an accelerant. COVID-19 placed extraordinary strain on health systems in low- and middle-income countries, exposing chronic shortages of equipment, inadequate compensation, and difficult working conditions. For many Ghanaian health workers, the pandemic years intensified economic difficulties and dimmed already-poor job prospects at home. The authors also point to strained relations between health workers and leadership as a factor likely to feature prominently in the literature, alongside the systemic layer that underpins many of these grievances: corrupt practices within Ghana’s healthcare system that shape hiring, posting, promotion, and pay.

Methodologically, the study follows an established and transparent recipe. The review is guided by the framework developed by Arksey and O’Malley, one of the foundational methodological frameworks for scoping reviews, which allows researchers to systematically map the breadth of literature on a topic rather than answer a single narrow effectiveness question. Findings will be reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews, known as PRISMA-ScR, a checklist that standardizes how such reviews are documented so that readers can judge the completeness and rigor of the search and selection process.

The evidence base will be drawn from four major bibliographic databases: PubMed, Scopus, Web of Science, and CINAHL, the latter being the leading index for nursing and allied health literature. Because relevant policy documents, reports, and grey literature may not appear in those databases, the team will run supplementary searches in Google Scholar and Google. The search window is deliberately bounded, covering articles published between 2021 and 2026, a period chosen to capture the post-pandemic wave of migration rather than the longer historical record. Two research questions anchor the entire exercise: how prevalent is post-COVID-19 brain-drain of healthcare workers from Ghana, and which factors, including corruption, contribute to it.

Once records are gathered, the screening process will apply pre-determined inclusion criteria to full-text articles, a step designed to filter out studies that do not directly address the research questions or meet quality thresholds. Relevant data from the included articles will then be extracted and charted, a scoping-review technique in which key information from each study is organized into a structured form that allows patterns to emerge across the literature. The analysis itself will be qualitative: thematic analysis and synthesis will be performed on the charted data to generate findings upon which conclusions are drawn. This means the review will not pool numbers statistically, as a meta-analysis would, but will instead identify recurring themes, such as salary inadequacy, workplace mistreatment, or corrupt recruitment practices, across the body of published evidence.

The stakes of this mapping exercise are considerable. Ghana’s health system already operates under workforce constraints that predate the pandemic, and every departing nurse or physician represents not only a lost pair of hands but also a lost investment. Training a single health professional consumes significant public resources, and when that professional’s first years of service are delivered in a foreign health system, the country that paid for the training effectively subsidizes the wealthier destination. The authors frame this dynamic as a direct threat to Sustainable Development Goal 3, the United Nations target of ensuring healthy lives and promoting well-being for all, because a health system cannot deliver quality care without an adequate, retained workforce.

What distinguishes this protocol from much of the existing migration literature is its insistence that corruption be treated as a first-order variable rather than background noise. The authors state that many of the factors driving health workers out of Ghana remain rooted in corrupt practices in the country’s healthcare system, and they argue that addressing healthcare worker migration from Ghana means addressing these systemic factors directly. In practical terms, that could mean examining how favoritism in postings, informal payments, or opaque promotion pathways erode morale and push skilled workers toward the exit. By explicitly including corruption among the factors to be charted, the review is positioned to produce evidence that speaks to governance reform as much as to workforce policy.

The publication of a protocol, rather than a completed review, is itself a meaningful scientific practice. Registering the search strategy, inclusion criteria, and analytical approach in advance reduces the risk that researchers unconsciously steer a review toward convenient conclusions, and it invites scrutiny from other scholars before any findings exist. The article, which was received in March 2026, accepted in July 2026, and published as an open-access review on 13 July 2026, received no external funding, and the authors note that no ethical approval was required because the study collects no primary data, relying entirely on already-published literature.

For Ghana, the implications of the eventual findings could reach well beyond academia. If the synthesis confirms that economic hardship, poor job prospects, strained leader-worker relations, and corruption are the dominant drivers of the post-pandemic exodus, policymakers will have a consolidated evidence base for designing retention strategies that go beyond pay raises alone. The authors’ central argument is blunt: retaining the healthcare workforce that universal health coverage depends on requires confronting systemic problems, including corruption and related issues, rather than treating each departure as an individual career choice. As high-income countries continue to recruit aggressively from lower-income nations, the countries losing their health workers will increasingly need evidence of exactly this kind to defend their most critical asset, the people who deliver care.

Subject of Research: Post-pandemic emigration of healthcare workers from Ghana and its contributing factors, including corruption

Article Title: Factors affecting post-pandemic brain-drain of healthcare workers from Ghana: a scoping review protocol

Article References: Attigah, D. K., Acquah, P. K., Oppong, J. D., Donkor, L. A., Bio, J. Y., Agyemang, E. D., Boateng, J., Arhin, J., Prah, E., Fynn, H. P. K., Buabeng-Odoom, B., Nartey, K. T., Jingbeja, E., & Ansah, E. W. (2026). Factors affecting post-pandemic brain-drain of healthcare workers from Ghana: a scoping review protocol. Discover Social Science and Health. https://doi.org/10.1007/s44155-026-00451-1

Image Credits: AI Generated

DOI: 10.1007/s44155-026-00451-1

Keywords: Ghana, brain-drain, healthcare workers, COVID-19, corruption, scoping review, universal health coverage, health workforce migration, University of Cape Coast, PRISMA-ScR, Sustainable Development Goal 3, health policy

The research team built their analysis on a reproducible systematic mapping review conducted under the PRISMA-ScR standard, the accepted protocol for systematic reviews in health and computer science research. After a rigorous screening pipeline, 259 primary studies formed the core analytical corpus, with 294 studies in the released living-review corpus. The authors are unusually candid about the limits of their own method: primary-arm records were screened by a single audited automated pass, while human double-screening with an inter-rater agreement of kappa equal to 0.65 was applied only to a supplementary arm. They explicitly describe broad-map attribute shares as provisional heuristic estimates and stress that the corpus is a carefully constructed mapping sample, not a census of the entire field. That kind of methodological honesty is rare in a literature often criticized for overclaiming.

The survey’s first major contribution is a crisp conceptual boundary called the dependent-step test. The test demarcates genuine autonomous agentic evaluation from ordinary static natural-language-processing evaluation and from prompt engineering. In essence, an evaluation qualifies as agentic only when later steps in a task depend on the outcomes of earlier steps, so that errors compound and the agent must recover, replan, or abandon a strategy mid-trajectory. A model that answers a thousand independent trivia questions is being tested very differently from one that must book a flight, notice the payment failed, diagnose why, and retry with a corrected form. The dependent-step test gives researchers a principled way to decide whether a benchmark is actually measuring agency or merely repackaging static question answering.

The second contribution is a meta-taxonomy that places every agent benchmark in a three-pillar coordinate system: capability, meaning what the benchmark measures; scoring paradigm, meaning how performance is judged; and environment topology, meaning where the agent operates. This framework matters because two benchmarks with identical headline scores may probe entirely different faculties. One may reward factual recall in a sandboxed text environment, while another measures tool orchestration in a live operating system where a single mistyped command can cascade into failure. By locating each benchmark along these three axes, the taxonomy lets researchers compare evaluations like coordinates on a map rather than as isolated leaderboard entries, exposing blind spots where entire capability regions remain untested.

Third, the authors paired an evidence map over the full corpus with a purposive, hand-verified, venue-verified landscape matrix of seventeen prominent benchmarks, achieving an inter-rater reliability of kappa equal to 0.77, a level conventionally regarded as substantial agreement. This matrix functions as a deep-dive companion to the broad map: where the corpus-wide analysis offers breadth with provisional estimates, the seventeen-benchmark matrix offers verified depth on the evaluations that most actively shape the field’s self-image. The contrast between the two instruments is itself instructive, showing how much confidence a reader should place in each kind of claim.

The survey’s most striking finding emerges from its critical analysis of that verified matrix. The authors identify a structural trilemma facing agent evaluation: data contamination, non-determinism, and execution cost. Data contamination occurs when benchmark tasks leak into training data, inflating scores without genuine capability. Non-determinism arises because agentic trajectories involve stochastic model behavior and environment interactions, so the same agent can score differently across runs. Execution cost reflects the computational expense of running agents through long, tool-using trajectories, which discourages the repeated trials needed to tame non-determinism. Among the seventeen prominent, verified benchmarks, the study found that none reports evidence that all three threats are jointly controlled, and none reports a complete standardized run-cost record, a 0 out of 17 result that should unsettle anyone quoting leaderboard numbers.

The implications of that 0 out of 17 finding ripple outward. When contamination is uncontrolled, a high score may reflect memorization rather than reasoning. When non-determinism is unquantified, a single reported run may be a lucky draw rather than a stable estimate of ability. When cost is unrecorded, results cannot be reproduced at reasonable expense, and the community cannot even assess whether an evaluation is practical to repeat. Together these gaps mean that many celebrated comparisons between competing agents may be statistically fragile, and that the field’s rapid narrative of progress rests on measurements whose error bars are largely unknown. The trilemma is structural because addressing any one threat tends to worsen another: more repetitions to handle non-determinism raise cost, and larger task sets that resist contamination raise it further.

The transparency of the review process itself deserves attention as a model for the field. The authors released a companion repository containing their pipeline, including a build script that reproduces every reported count against a committed HTTP response cache, yielding byte-identical, MD5-verified outputs with zero live API calls. They document per-source query status, flagging databases where only counts could be verified and one major index that remained unresolved for lack of credentials. They even ran a capture-recapture diagnostic between OpenAlex and Crossref, then declined to treat its output as a recall estimate because the two indices are heterogeneous targeted searches rather than independent random samples, a violation of the estimator’s core assumptions. A probe-set validation instead confirmed that canonical benchmarks were captured completely, with Scopus corroborating the result. This is bibliometrics done with unusual care.

Looking forward, the survey proposes a five-direction roadmap that reframes progress not as the invention of yet more individual benchmarks but as a shift toward standardized, reliability-first, cost-aware evaluation infrastructure. In practice that means community-agreed protocols for reporting contamination checks, variance across repeated runs, and the compute cost of every evaluation, so that a benchmark score arrives with the same kind of measurement metadata that a physics experiment or clinical trial would demand. The roadmap effectively asks the field to grow up: to treat agent evaluation as a measurement science with known instruments, calibrated error, and reproducible procedures, rather than as a proliferation of leaderboards each with its own unstated assumptions.

For an industry pouring billions into agentic AI products, the stakes could hardly be higher. Autonomous agents are beginning to write code, manage workflows, and interact with real systems on behalf of users, and deployment decisions increasingly cite benchmark performance as evidence of safety and competence. If, as this survey shows, no prominent benchmark currently demonstrates joint control over contamination, non-determinism, and cost, then the numbers guiding those decisions are weaker than they appear. The study does not claim agents are failing; it claims we cannot yet reliably tell. Turning evaluation from a competitive spectacle into trustworthy infrastructure, the authors argue, is the prerequisite for knowing what today’s agents can actually do, and the open repository accompanying the paper offers the community a concrete place to start.

Subject of Research: Evaluation and benchmarking of large language model agents

Article Title: Large language model agent evaluation and benchmarking: a systematic survey, meta-taxonomy, and critical research roadmap

Article References: Nageshwaran, V., Ezekiel, S., Tran, T. T., & Lakshmi Narasimhan, V. (2026). Large language model agent evaluation and benchmarking: a systematic survey, meta-taxonomy, and critical research roadmap. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11678-4

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11678-4

Keywords: large language models, AI agents, benchmarking, agent evaluation, meta-taxonomy, systematic review, data contamination, non-determinism, execution cost, PRISMA-ScR, evaluation infrastructure, Artificial Intelligence Review