Introduction

Recognition of the integral role of surgery in health delivery on a global basis has grown [1, 2]. The Disease Control Priorities Project (DCPP), a World Bank-funded effort to evaluate health interventions for low- and middle-income countries, identified surgery as a cost-effective strategy to reduce disease burden [3] and continues to estimate the effectiveness and importance of surgical services. Concurrently, The Lancet has convened a Commission on Global Surgery working to improve surgical delivery worldwide [4].

The global volume of surgery is estimated at 312.9 (95 % CI 266.2–359.5) million operations in 2012, but rates of surgery vary dramatically by country and region [5]. Less than 7 % of the global volume of surgery was performed in very low health expenditure countries which account for 37 % of the world’s population, while 60 % of the surgical volume took place in the high-expenditure countries which account for 18 % of the world’s population. Surgical rates ranged from under 100 to over 20,000 operations per 100,000 people annually [5], indicating tremendous variability in the provision of surgical services.

Benchmarking levels of service provisions has provided important guidance for policymakers at national and international levels. For instance, in 1985 maternal health advocates proposed a minimum cesarean delivery rate of 10 % based on limited though convincing perinatal mortality data. This provided a target for health system development which has improved maternal and perinatal outcomes [6–9].

A general lack of understanding of the relationship between surgical care and population health outcomes has limited the ability to set similar benchmarks for surgical capacity beyond those that exist for cesarean delivery. In order to establish a benchmark for surgical service provision, we used available data to estimate minimum rates of surgery associated with a variety of important health indicators including life expectancy (LE) and maternal mortality ratio, and estimated minimum surgical needs based on disease burden. We evaluated the validity of these rates by comparing them to known surgical rates from Chile, China, Costa Rica, and Cuba (“4C”), countries with moderate resources but high health outcomes.

Materials and methods

We used three strategies to identify minimum rates of surgery that are associated with positive health indicators. We first estimated rates of surgery per country for 2012 associated with LE of 74–75 years, obtained from the World Health Organization and the World Bank [10, 11]. We used previously reported country level surgical rate data for 2012 [5]. The mean LE by World Bank income group (high-, upper-middle-, lower-middle-, and low-income) showed a mean LE of 79.3 years for high-income countries, 73.8 years for upper-middle-income countries, 68.2 years for lower-middle-income countries, and 61.1 years for low-income counties. We chose a LE of 74–75 years of age as an optimistic yet reasonable goal and calculated the median surgical rate of countries that had a LE within this range.

For our second strategy, we estimated rates of surgery associated with a maternal mortality ratio (MMR) of less than or equal to 100 deaths per 100,000 live births using 2013 MMR data from the World Bank, with MMR defined as the number of women who die from pregnancy related causes while pregnant or up to 42 days postpartum per 100,000 live births [10]. We chose this MMR as our threshold based on the United Nations Millennium Development Goals of reducing the MMR by three quarters between 1990 and 2015. The MMR in 1990 was 380, thus a reduction of three quarters is an MMR of 95 by 2015 [12]. Once again, we used the previously reported country level surgical rate data for 2012 [5], and obtained the median surgical rate of countries that had an MMR of less than or equal to 100 deaths per 100,000 live births.

For our third strategy, we estimated the minimum need for surgery in the 7 global super-regions based on the prevalence of conditions from previous work by Rose et al. [13]. These prevalence data were obtained from the 2010 global burden of disease (GBD); the data were organized in 119 disease states in accordance with the World Health Organization Global Health Estimate (GHE) [14]. These data represented 187 countries and were organized according to the 21 GBD epidemiologic regions. Predefined values for the incident need for surgery for each GHE disease states were used to calculate minimal need for surgery in each region based on the prevalence of each condition. A detailed description of these methods is described in a separate study [13]. We then aggregated these surgical need data to the seven global super-regions as used by GBD and weighted by population size [15]. We calculated the mean surgical need of the seven global super-regions for 2010.

We then compared our findings to the observed surgical rates of the “4C” countries (Chile, China, Costa Rica, Cuba) identified in The Lancet Global Health 2035 Commission Report [16]. The “4C” countries were used in the report as examples of countries that were classified as low-income or lower-middle-income in 1990 but that had reached high levels of health status by 2011. These four countries illustrated that a “scaling-up” of health sector interventions was possible, and produced better outcomes than economically similar countries and regions that had not done so [16]. We calculated the mean surgical rate for the 4C countries reported for 2012 as an external test of validity.

We used SAS version 9.3 for all statistical analyses.

Results

Based on the first strategy assessing surgical rates and LE, countries with reported LE of 74-75 years (N = 17) had a median surgical rate of 4392 (IQR 2897–4873) operations per 100,000 people annually. Figure 1 demonstrates the relationship between observed surgical rates per country and LE and shows that as surgical rates increase, there is a trend toward an increase in LE (adjusted R-square 0.5148). Based on our second strategy exploring the relationship between surgical rates and MMR, the median surgical rate for countries that had an MMR below 100 (N = 109) is 5028 (IQR 4139–6778) operations per 100,000 people annually. Figure 2 demonstrates the relationship between observed surgical rates per country and MMR and shows that as surgical rate increases, the MMR decreases (adjusted R-Square 0.5556). Based on the third strategy assessing minimum surgical need, the mean surgical rate estimated for the 7 super-regions based on the 21 GBD regions’ prevalence data was 4723 (95 % CI 3967–5478) operations per 100,000 people annually. Estimated need of surgery for the 21 GBD regions and seven super-regions based on prevalence data from 2010 are shown in Table 1. The “4C” countries had a mean surgical rate of 4344 (95 % CI 2620–6068) operations per 100,000 people annually.

Fig. 1
figure 1

Relationship between observed surgical rates in 2012 and corresponding life expectancy for 194 WHO member countries. Adjusted R-square 0.5148. Adapted from Weiser et al.

Full size image
Fig. 2
figure 2

Relationship between maternal mortality ratio and surgical rate in 2012 for 181 WHO member countries (13 of the 194 WHO member countries did not have maternal mortality ratio data). Adjusted R-square: 0.5556

Full size image
Table 1 Estimated need of surgery for the 21 GBD regions and seven super-regions based on prevalence data from 2010, weighted by population
Full size table

Based on 2012 surgical rate estimates, thirteen of the twenty-one GBD regions, accounting for 5.5 billion people (78 % of the world’s population), do not achieve the lowest end of the surgical rate range (Table 2).

Table 2 Estimated volume of surgery that occurred in 2012 and estimated surgical rates based on population for each of the 21 GBD regions (regions in italics are those that do not meet lowest surgical rate of proposed range)
Full size table

Discussion

Despite three diverse strategies, we identified a surprisingly narrow range of surgical rates associated with positive health indicators, including a LE of 74–75, a MMR of less than or equal to 100, and minimal surgical capacity to meet surgical disease prevalence. While there is no ideal strategy for identifying a minimum surgical rate threshold, LE and maternal mortality are reasonable outcome measures to choose, as they are desirable benchmarks for low- and middle-income countries working to improve their health systems. Using disease prevalence to estimate surgical need allows regions to consider and accommodate differences in their disease profiles when planning surgical service provision. The 4C countries also had a surgical rate similar to our three proposed strategies. This similarity helps validate our findings, as the 4C countries have experienced improved health outcomes and have exemplary maternal and neonatal health outcomes.

Using these three strategies along with our validation scheme, we identified that countries with a surgical rate ranging from 4344 to 5028 operations per 100,000 population annually have achieved desirable health outcomes. Although our study findings cannot infer a causal relationship between a minimum surgical rate and these positive health outcomes, it is likely that countries providing these minimum levels of surgical care have strong health systems. Health systems that are able to support surgical provision to this degree are likely able to support other health interventions that improve its population’s longevity. As such, surgical rates and capacity serve as a marker of the strength of a health system.

We chose a LE target of 74–75, which is 13–14 years above the mean LE of low-income countries and approximately 5 years below the mean LE of high-income countries. Countries with a LE of 74–75 had a median surgical rate of approximately 4400 operations per 100,000 people annually. Since this might appear arbitrary, we calculated the median surgical rates for LE of 73–74 years and 75–76 years, which resulted in similar rates of 4584 and 4694 operations per 100,000 people annually, respectively. However, the magnitude of the effect and the direction of the relationship (more longevity leading to more need for surgery, or vice versa) are complicated, and simply increasing surgical rates without attending to other fundamental weaknesses plaguing health systems are unlikely to meaningfully improve life expectancy [11].

There are certain limitations with each strategy. The reported global volume of surgery was based on operations in the operating theater, but standardization was not uniform across countries. Country-specific volumes may vary based on how such operations and procedures are recorded. This study is an ecological analysis, and therefore causal inferences between the relationship of surgical rates and LE and MMR cannot be made [17]. We did not adjust for other confounders, such as GDP or health expenditure, as these are collinear with surgical rates. Nor do we do not account for variation in surgical rates within each country and region, and therefore cannot assess the impact this may have on health outcomes. For example, there are likely differences in surgical rates in urban areas compared to rural areas of countries that are not accounted for in our study. Our assessment of surgical rates associated with surgical disease prevalence is likely an underestimate of the actual surgical need. Because of methodological considerations, many surgical conditions were not included in the analysis assessing surgical need based on disease prevalence [14]. Additionally, resource limited settings may have a higher burden of surgical conditions than might otherwise be expected, as neglected infections require surgical intervention more frequently than they might if addressed in a timely fashion. Despite these limitations, our study highlights important relationships between surgical rates and positive health indicators. As such, surgical rates and surgical capacity serve as a marker of the strength of a health system.

This target range of 4400–5000 operations per 100,000 people annually can be used as a benchmark goal that informs policy aimed at strengthening health care systems and surgical capacity. Maternal health advocates have used benchmarking and have shown the importance of cesarean availability to improve maternal outcomes. To our knowledge this has not been done for surgery, and we hope our findings can support surgical advocacy in the global health arena.

These findings support the argument that minimum rates of surgery are associated with desirable health indicators. Three-fourths of the world’s population, accounting for 5.5 billion people, live in countries that do not meet the lowest surgical rate of our proposed range. Surgery plays an important role in global health and health service delivery, and many countries require increased surgical infrastructure to strengthen overall health systems and health outcomes.