The study, published in Frontiers in Health Services, was led by Yuki Julius Ng of University Hospital Bristol and Weston Trust in the United Kingdom, together with colleagues from the Norwegian University of Science and Technology, St Olav’s Hospital in Trondheim, the George Institute for Global Health in New Delhi, BARC Hospital in Mumbai, Karolinska Institutet and the University of Global Health Equity in Rwanda. The team asked a deceptively simple question: at the national level, how strongly does the density of the specialist surgical, anaesthetic and obstetric workforce predict how many operations a country actually performs?

To answer it, the researchers carried out a cross-sectional ecological regression using publicly available World Bank data. They extracted the specialist surgeon, anaesthesiologist and obstetrician workforce per 100,000 population and the number of surgical procedures per 100,000 population for every country reporting both indicators. When the two figures came from different years, the workforce density was adjusted for population growth so that each workforce estimate corresponded to the year of its surgical volume. Countries reporting more than 10,000 procedures per 100,000 population were excluded from the primary, policy-focused model, which removed ten very high-volume health systems including Australia, Belgium, Brazil, Denmark and Sweden. A secondary unrestricted analysis including all 67 eligible countries was run to check how these extreme observations shaped the fitted relationship.

The final primary dataset contained 57 countries, of which 37, or 64.9 percent, required temporal adjustment. The team explored several functional forms, including linear, power, exponential and polynomial models, but the logarithmic model best captured the observed pattern. The fitted equation, y = 884 ln(x) + 685, where x is the specialist workforce density and y is the surgical volume, explained 52.4 percent of the variance in surgical volume across countries, with an adjusted R-squared of 0.516. The regression coefficient was highly significant, at 884.12 with a 95 percent confidence interval from 656.6 to 1,111.7 and a p-value below 0.001.

The numbers behind the curve are striking. At a workforce density of 10 specialists per 100,000 population, the model predicted roughly 2,720 procedures per 100,000 population. Doubling the workforce to 20 per 100,000 raised the predicted volume to 3,333.7 procedures, with a 95 percent confidence interval for the fitted mean of 2,917.2 to 3,750.2. Doubling again to 40 per 100,000 produced 3,946.5 procedures, with a confidence interval of 3,470.7 to 4,422.3. In other words, each doubling of the specialist workforce bought a progressively smaller increase in surgical output, a hallmark of the logarithmic relationship.

The researchers quantified this flattening with a model-implied marginal slope analysis, calculating the first derivative of the fitted curve at each workforce density. At a density of 2 per 100,000, the local slope was approximately 439.4 additional procedures per 100,000 for each extra specialist per 100,000 population. By a density of 10, the slope had fallen to about 87.9. The authors stress that this quantity is a mathematical property of the fitted cross-country curve, not a causal estimate of what happens when a single specialist is added to a health system. They also caution that the logarithmic model breaks down at the extreme lower tail: below roughly 0.46 specialists per 100,000, the equation produces physically impossible negative volumes, so four countries in the dataset, Sierra Leone, South Sudan, Chad and Niger, fell into a region where the fitted values cannot be read literally.

Regression diagnostics supported the overall fit while flagging some influential observations. Standardised residuals centred on zero with a maximum of 3.514, and the maximum Cook’s distance was 0.173, above the conventional 4/n heuristic, indicating at least one potentially influential country. The normal probability plot showed general correspondence with some departure in the upper tail, and the maximum studentized deleted residual was 4.023. These diagnostics were considered alongside the results rather than used as automatic exclusion criteria, and the authors emphasise that the cross-sectional design means the relationship reflects association, not causation, capturing the combined influence of workforce availability and broader health-system capacity such as infrastructure, operating theatres and service organisation.

The income-group pattern in the data was stark. Of the 13 high-income countries in the primary model, 10, or 76.9 percent, performed above the fitted trend, with Nauru, the Isle of Man and the Cayman Islands exceeding even the upper confidence limit. Among the 21 upper-middle-income countries, 6, or 28.6 percent, exceeded the prediction, including Fiji, Costa Rica, Mongolia, Armenia, Tuvalu and Belarus. Only 3 of 17 lower-middle-income countries, Sri Lanka, Samoa and Bhutan, performed above the trendline, and every low-income country in the model fell below it, many by a wide margin. Sri Lanka and Bhutan are particularly intriguing outliers, achieving more surgery than predicted despite constrained resources, and the authors suggest they merit closer investigation as possible models of efficiency.

The findings challenge the assumption of constant productivity embedded in the original Lancet Commission benchmarks, which were calculated on the premise that output would not change as the workforce grew. Real-world evidence supports the diminishing-returns pattern: in Sierra Leone, doubling the surgical workforce between 2012 and 2017 produced only a 15.6 percent increase in surgical volume, while in the United Kingdom up to 56 percent of the surgical workforce reported barriers to operating room access linked to burnout, anaesthetic shortages and infrastructure limits. The authors frame the pattern through the concept of the surgical transition: as health systems mature, additional workforce capacity is increasingly absorbed into complex, resource-intensive and subspecialised care that requires larger teams, more anaesthetic support and more perioperative resources per case, rather than simply more procedures.

The study also exposes deep inequities that national averages conceal. Mongolia outperformed the fitted model thanks to strong infrastructure, yet around 80 percent of its procedures occur in the capital. Mexico fell below its predicted volume of 727 procedures per 100,000 despite meeting workforce targets, with health-system fragmentation a suspected but unmeasured factor. In Pacific Island nations such as Palau and the Cook Islands, high specialist densities coexist with only one anaesthetist, showing how density alone can mislead. The authors argue for tiered surgical volume targets as intermediate milestones, for equity-focused indicators capturing geographic access and facility readiness, and for financing reforms, from New Zealand’s package-of-care model to fully costed National Surgical, Obstetric and Anaesthesia Plans, so that workforce expansion is matched with operating theatres, anaesthetic services, blood products and predictable funding. Task-sharing is part of the answer, with non-specialist providers performing up to 90 percent of obstetric surgeries in some settings, but reports from Uganda show more than a third of surgical facilities lack essential equipment such as oxygen or anaesthesia machines. The authors conclude that the nonlinear relationship should be treated as exploratory rather than a validated prediction framework, and that future work with harmonised longitudinal data must incorporate health-system capacity before such models guide policy.

Subject of Research: Global surgery workforce

Article Title: The correlation between surgical, anaesthetic and obstetric workforce and surgical volume: a modelling study

Article References: Ng, Y. J., van Duinen, A. J., Bolkan, H. A., Gadgil, A., Patil, P., & Roy, N. (2026). The correlation between surgical, anaesthetic and obstetric workforce and surgical volume: a modelling study. Frontiers in Health Services, 6, Article 1920090. https://doi.org/10.3389/frhs.2026.1920090

Image Credits: AI Generated

DOI: 10.3389/frhs.2026.1920090

Keywords: global surgery, surgical volume, health workforce, Lancet Commission on Global Surgery, logarithmic regression, health systems, low-income countries, surgical indicators, correlation, between, surgical, anaesthetic