Introduction

Over the past several years there has been a growing interest in the role of surgery in global public health [1–4]. Inspired by the work of economist Dean Jamison [5] and the chapter “Surgery” in the Disease Control Priorities in Developing Countries, 2nd edition [6] on the cost-effectiveness of surgical care in the developing world, a Global Burden of Surgical Disease Working Group (GBSDWG) was formed in 2008 [7]. This group of surgeons, anesthesiologists, emergency physicians, and public health experts met for the first time in Seattle on April 16–17, 2008 to begin a dialogue that explores the role of surgery in public health, priority areas of research, and how best to advocate for support of global surgical programs.

The overwhelming consensus of attendees at this meeting pointed to a critical lack of data concerning the true portion of the global burden of diseases that could be alleviated by surgical expertise. Current methodology was inadequate for measuring the impact of surgical care. During these discussions, several key problems were identified that have impeded accurate measurement of the global burden of surgical conditions: First, basic definitions related to surgical conditions have not been clarified. Second, surgery is a procedure-oriented specialty; and previous global burden of diseases estimates have not been approached from an intervention perspective. Third, the large number of surgical diagnoses and treatments and wide range in patient ages complicate the analysis.

In this article, we propose definitions for assessing surgery from a public health perspective. We also describe a conceptual framework within which we can estimate the burden of surgical conditions and the unmet need for surgical care.

Methods

At the 2008 GBSDWG meeting, a committee was designated to work on surgical definitions. The committee included five surgeons, two anesthesiologists, an emergency room physician, and a health economist. All members of the committee had global health experience, and six members had advanced degrees in public health. Over the subsequent year, the committee worked together to develop a consensus opinion on definitions and methodology for assessing surgical care from a public health perspective.

Results

Table 1 summarizes our proposed definitions for analyzing surgical care from a public health perspective. The definitions were crafted with the recognition that not all surgical conditions require a procedure; nonphysicians or general doctors, rather than surgical specialists, often provide surgical care in low-income countries; and multiple surgical sequelae can result from a single surgical condition.

Table 1 Terminology for analyzing surgical care from a public health perspective
Full size table

Two parameters are described for assessing the public health impact of surgical conditions and surgical care. The severity of a surgical condition is represented as a surgical disability weight (DW). As used in the Disease Control Priorities Project [8], surgical disability weight is a health state valuation expressed on a ratio scale between 0 (full health) and 1 (states equivalent to death).

To measure the impact of surgical care, we propose a parameter called the “value of surgical care” (VSC), which is the relative ability to prevent or reverse a surgical disability. In a manner similar to disability weight, the value of surgical care is expressed on a ratio scale (between 0 and 1) according to how effectively the intervention reverses the disability. Surgical care is curative when the value of surgical care matches the disability weight (VSC = DW) or partially restorative when the value of surgical care is less than the disability weight (VSC < DW). In situations where surgical care results in a disability greater than the initial condition, the complication is calculated as a separate surgical condition with its own disability weight.

The method we describe assigns a DALYFootnote 1 value to each surgical condition and a DALY averted value to each surgical intervention. Using this approach, it is possible to measure quantitatively the burden of surgical conditions and the impact of surgical care.

Figure 1 presents a conceptual framework for estimating the burden of surgical conditions and the unmet need for surgical care. The framework is based on the premise that surgically related DALYs within a population are determined by the incidence of surgical conditions and the quantity and quality of surgical care. For analysis purposes, a cumulative incidence function of surgical DALYs is proposed. Cumulative surgical DALYs can be calculated from age-specific data (Appendix 1).

Fig. 1
figure 1

Conceptual framework for estimating the burden of surgical conditions and the need for surgical care. Surgical disability and premature death in a population is a time-dependent function that relates to the incidence of surgical conditions and the quantity and quality of surgical care. The potential, actual, and unpreventable curves represent the age-specific cumulative disability-adjusted life-years (DALYs) with different levels of surgical care. See Table 2 for the definitions describing the need for surgical care and the burden of surgical conditions

Full size image

As shown in Fig. 1, average cumulative DALY curves can vary depending on the level of surgical care. In the best case scenario (when surgical care is ideal), the only surgical disability and premature death in a population is that which is unpreventable. In the worst case scenario (when surgical care is nonexistent), surgical disability and death reaches the maximum possible. Because surgical care is never completely absent, nor ever ideal, the actual surgical DALY curve lies somewhere between these two curves.

Table 2 summarizes terminology for describing the need for surgical care within a population. The met need refers to the surgical care that has been provided, and unmet need refers to potentially treatable disability and death due surgical conditions. The unmeetable need is the disability and premature death that is unavoidable, even with the best surgical care. The burden of surgical conditions is defined as the total disability and premature death that would occur in a population should there be no surgical care. The burden of surgical conditions is expressed as DALYs, met need for surgical care as DALYs averted, and unmet need as potential DALYs avertable. Appendix 2 shows an example of how these parameters can be calculated using population-based data.

Table 2 Proposed terminology for describing and measuring burden of surgical conditions and need for surgical care
Full size table

Discussion

There is an ongoing debate about the relative importance of surgical care in global public health. Definitions of the burden of surgical conditions and the impact of surgical care are critical to clarifying the relative priority of surgery within global health. To date, neither the global burden of surgical conditions nor the effect of surgical treatment has been quantified with existing measures [9]. As major disparities in surgical care exist between high and low-income countries, an estimate of regionally specific disability-adjusted life-years that can be averted by surgical interventions is also needed [6].

In this article, we outline a conceptual framework for estimating the burden of surgical conditions and unmet need for surgical care. Our definition for surgical conditions is broader than the definition used in the Disease Control Priorities Project, where surgical conditions were defined as “any treatment that includes suture, incision, excision, manipulation, or other invasive procedure that usually, but not always, requires local, regional, or general anesthesia” [6]. The main consideration in broadening this definition is the fact that surgical conditions do not always require a surgical procedure. Examples include the care of most head injuries and nonoperative management of blunt abdominal injuries (e.g., splenic injury in a child). A clear advantage of the broader definition is that it more accurately reflects the surgical workload, which may be critical when limited resources are being allocated. While the ratio of nonoperative to operative surgical care is likely to vary by specialty, limited data suggest that it may be substantial. In a community-based study of pediatric surgical conditions in West Africa, only 46% of children presenting with a surgical condition required a surgical procedure [10]. Research is needed to determine what this ratio might be in other areas of surgery.

To be consistent with the latest Global Burden of Disease (GBD) study, we use the term surgical sequelae to describe any abnormality that results from a surgical condition or its treatment [11]. As used in the current GBD study, sequelae refer to the combination of health states that result from particular causes. This terminology arose from confusion in past GBD studies as to why something was a cause versus sequela versus risk factor. We do not use the term “surgical cause” in any of our definitions, as the definition of a surgical condition is based on a clinical problem rather than an established etiology. The latest round of the GBD study, funded by the Bill and Melinda Gates Foundation, is scheduled to be completed by November 2010. The final study is expected to produce specific DALY, YLL, and YLD estimates for more than 220 diseases/injuries and 40-plus risk factors by age range and sex for 21 regions of the world.

We recognize that there may be confusion about what constitutes surgical conditions and sequelae and that the availability of surgical providers may alter rates of surgical consultation. However, it should be feasible to define surgical conditions and sequelae using the International Classification of Disease (ICD) system. The ICD is a coding system of diseases and signs, symptoms, abnormal findings, complaints, social circumstances, and external causes of injury or diseases as well as therapeutic interventions, as classified by the World Health Organization (WHO) [12]. As of 2008, the ICD-10 coding system includes 68,064 diagnoses and 86,917 procedure codes [13]. The increased detail in the ICD-10, compared to the ICD-9, should make it easier to reach a consensus on which codes represent surgical conditions.

Given the large number of diagnoses and procedures in ICD-10, it may be unrealistic to gather information (or estimate the burden) on the full spectrum of surgical conditions and/or interventions. A more practical approach might be to focus on a subset of these conditions or procedures. For example, the short list of “essential” surgical procedures (outlined by the WHO) could be expanded from facilities-based data gathered in low-income countries as well as published material. Procedures such as cesarean section, abscess drainage, laparotomy, and fracture care are likely to account for a significant percentage of interventions in most settings.

Recognizing that in many parts of the world surgical procedures are not done by fully trained surgeons, we use the term “surgically trained provider” rather than surgeon. In many other low-income countries, general practitioners may perform surgical procedures. Twenty-five Sub-Saharan countries utilize nonphysician clinicians, and almost half of them perform minor surgical procedures [14]. In Uganda, a study of five general hospitals reported that more than 5000 surgical procedures were performed annually by general practitioners [15]. With only nine orthopedic surgeons for than 20 million people, Malawi has trained orthopedic clinical officers to deliver most of the musculoskeletal services for the country [16]. In addition, 90% of the cesarean sections at the district hospital level in Malawi are carried out by surgically trained clinical officers [17]. In Mozambique, técnicos de cirurgia have performed major surgery in district hospitals since 1989 [18]. They perform 92% of emergency obstetric care and 65% of major general surgery at the district hospital level. Other examples of surgically trained providers include nonphysician anesthesia providers [19]. A surgically trained provider implies a level of expertise in making a diagnosis, formulating a treatment plan (including the decision whether an operation is necessary), performing a procedure, and recognizing and treating any complications.

Our method for estimating the burden of surgical conditions and the unmet need for surgical care is based on the concept of “cumulative incidence.” Cumulative incidence refers to the number of new cases that occur in a population over a period of time and is expressed in terms of the number of people at risk in the population at the beginning of the study. A key feature of cumulative incidence analysis is that an average age-specific cumulative risk curve can be plotted using population-based data. This is done under the assumption that age-specific incidence rates remain constant in the future. “Lifetime risk” of disease is a variation of the cumulative incidence concept and has been used by clinicians, researchers, and policymakers to assess the burden of a wide variety of diseases [20–26].

We use disability weights and values for surgical care to calculate cumulative surgical DALYs. This approach allows a DALY value to be assigned to each surgical condition and surgical intervention. In the case of rendered surgical care, the units of measure are expressed as DALYs averted. This approach is fundamentally the same as that used by McCord and Chowdhury [27] and Debas et al. [6], with the exception that their calculations were based on the percentage of averted risk. We believe that the “value of surgical care” concept allows a much more detailed analysis of surgical services. One concern is that it might be difficult to assign disability weights and values of surgical care to the large number of surgical conditions and interventions. Developing an approach that utilizes an average value for disability weight and value of surgical care could circumvent this problem.

The terminology we use to describe met and unmet need for surgical care is based on definitions used in obstetric need studies [28]. In these studies, obstetric need was estimated using census data, number of females at reproductive ages, and birth rates. Unfortunately, these methods are not as easily applied to the much broader spectrum of surgical conditions. Nevertheless, using terminology that is consistent with the obstetric literature should lessen any confusion.

An important feature of our proposed methodology is that the results will be comparable to other burden of disease studies. “DALYs” have been selected as the measure of burden of surgical conditions and “DALYs averted” as a measure of the impact of surgical care. This should be helpful in determining where surgical care fits among other global health priorities.

We recognize that the DALY continues to evolve in response to criticism and that this summary measure does not take into account contextual variables. Whereas the impairment associated with a particular disease process may be the same throughout the world, the disability associated with each condition will necessarily be determined by social and cultural variables in each environment. Also, disability weights have not yet been calculated for many of the surgical diseases. We suggest that research be undertaken to revise calculations for disability weights associated with surgical conditions, aimed at achieving a consensus opinion from health professionals (and possibly lay people) in both high- and low-income environments.

Finally, even with a strategy to measure the burden of surgical conditions and the unmet need for surgical care at the global level, there are still major obstacles to obtaining the needed estimates. Foremost is a profound lack of data on the incidence and prevalence of surgical conditions. This is especially the case in low- and middle-income countries, where hospital data are of limited value because of access problems and comprehensive surgical databases do not exist. Ideally, prospective community-based surveys will be undertaken in low- and middle-income countries to acquire the needed data.

Conclusions

We have described a strategy for measuring the impact of surgical care from a public health perspective. Using this methodology, it should be feasible to estimate the burden of surgical conditions and unmet need for surgery in a wide variety of populations.