Global Volume and Safety of Surgical Care
The World Health Organization (WHO) has estimated that 234 million operations are performed worldwide each year (Weiser and others 2008). The WHO’s analysis establishes three significant findings:
Surgical interventions take place on a massive and previously unrecognized scale in all countries and resource settings.
The inequity in service provision among countries and settings is dramatic.
Little is known about the indications for, and the quality, safety, and outcomes of, surgical care.
Much has since been done to investigate these issues. The WHO has provided guidance for measuring surgical services and capacity through a set of standardized metrics for surgical surveillance (Weiser and others 2009). In addition, a situational analysis tool has been constructed and deployed in a number of countries to help assess surgical capacity (WHO 2010). Yet the logistics of performing surgery are complex and demand standardization. Surgical services must also continuously measure patient outcomes to identify shortcomings, inform improvements, and maintain high levels of quality care.
Mortality Following Surgery
The annual volume of surgery is almost twice that of obstetrical deliveries, and surgical death rates far surpass maternal mortality rates. Global estimates suggest that at least 7 million people suffer complications following surgery each year, including at least 1 million deaths, a magnitude that exceeds both maternal and AIDS-related mortality. As many as 50 percent of these deaths and complications are preventable (Weiser and others 2008). Surgical care is fraught with hazards in every setting; patients face immediate danger from both the technical risks of the procedures themselves and the anesthesia needed to induce insensibility and sedation.
Studies from high-income countries (HICs) confirm high rates of postoperative mortality and high variability in those rates. In the Netherlands, a review of 3.7 million inpatient surgical procedures at 102 hospitals over 15 years reveals a perioperative mortality rate of 1.85 percent (Noordzij and others 2010). A similar inpatient surgical death rate has been noted in the United States, with all-cause postoperative mortality in 2006 estimated to be 1.14 percent to 1.32 percent (Semel and others 2012; Weiser and others 2011). Pearse and others (2012) studied the outcomes of one week of inpatient surgery (excluding cardiac, neurosurgical, radiological, and obstetric procedures) in 498 hospitals in 28 European countries. The in-hospital crude postoperative mortality ranged from 1.2 percent in Iceland to 21.5 percent in Latvia. After adjusting for patient age, the American Society of Anesthesiologists’s score of patient fitness for surgery, the urgency of surgery, the extent of surgery (minor, intermediate, or major), the specialty, and the presence of metastatic disease or cirrhosis, the odds ratios for death following surgery—using the United Kingdom as reference—ranged from a low of 0.44 in Finland to a high of 6.92 in Poland (). Overall, the researchers noted a 4 percent in-hospital crude mortality rate in this sample of more than 46,000 surgical cases, a previously unreported and unexpectedly high number.
Adjusted Odds Ratio for In-Hospital Mortality Following Surgery in 28 European Countries, 2011.
Disparity of Service Provision
Nearly 60 percent of all operations take place in HICs, where 15 percent of the world’s population lives (). Low-income countries (LICs) account for nearly 35 percent of the global population, yet only 3.5 percent of all surgical interventions (Weiser and others 2008). This lack of equity in access to surgical interventions demands further investigation. The extent of unmet need in resource-poor settings remains unclear, but basic surgical services are increasingly recognized as essential for relieving suffering and sustaining health. Surgery is critical for obstetrical emergencies; common congenital conditions, such as clubfoot; traumatic injuries, including orthopedic injuries; and treatment of abscesses, cancers, hernias, and cataracts. Maternal health advocates estimate that an optimum rate of cesarean section is at least 5 percent of all births to avert high rates of death of mothers and children (Dumont and others 2001), but similar minimum criteria have not been proposed for basic surgical services to address the disease burden in a population.
Average National Rate of Surgery for Countries by Category of Health Expenditure, and Total Surgical Volume by Category, 2004.
It has, however, become clear that surgical care is an essential component of effective health delivery systems and vital for enabling long and healthy life. In a population-based survey in Sierra Leone, one of the poorest countries in the world, one in every four respondents reported needs that might benefit from surgical consultation (Groen and others 2012). In addition, almost one in three households had experienced a death within the past year; of these households, one in four had a condition within the week preceding death that likely could have been treated surgically: abdominal distention, bleeding or complications following childbirth, an acute or chronic wound, a mass, or an acquired or congenital deformity. This survey indicates a tremendous unaddressed disease burden that might be mitigated with improved access to surgical care.
Limitations in the Scope of Practice in LMICs
Surgical interventions are performed at much lower frequency in the resource-poor settings of low- and middle-income countries (LMICs) and under more limited circumstances. Typically, rural first-level hospitals and even third-level public hospitals have a high-percentage of urgent cases; these facilities focus on a limited set of interventions, given their resource constraints. For example, studies from LMICs have shown very high ratios of cesarean section compared with other types of surgical procedures. Cesarean section has been found to represent a substantially higher proportion of all surgical interventions in Sierra Leone (42 percent), Zambia (40 percent), Uganda (34 percent), Niger (26 percent), Malawi (23 percent), and Haiti (12 percent) than in Organisation for Economic Co-operation and Development countries (3 percent) (Bowman 2013; Hughes and others 2012; Kushner, Groen, and Kingham 2010). At least one study suggests that the higher the surgical capacity, the lower the observed proportion of cesarean section, indicating that improved skills, materials, and capacity allow the provision of a more comprehensive range of surgical services (Petroze, Mehtsun, and others 2012).
The WHO has identified a set of emergency and essential surgical interventions that all first-level hospitals with surgical capacity should be able to perform. Besides cesarean section, these procedures encompass uterine evacuation, circumcision, wound care, chest drainage, basic laparotomy, amputation, hernia repair, tubal ligation, closed or temporary reduction of fractures, cataract surgery, removal of foreign bodies, and emergency airway management and ventilation. This guideline for essential surgical services is based on the typical capacity of health facilities in remote, resource-constrained settings. These conditions involve relatively straightforward interventions, requiring less complex skills, resources, and postoperative management. Surgical providers at first-level hospitals appear to refer patients to higher-level facilities due to lack of training and experience rather than lack of resources (Bowman and others 2013; Petroze, Nzayisenga, and others 2012). The WHO recommends that referral facilities ensure capacity to provide facial and intracranial surgery, complex bowel surgery, pediatric and neonatal surgery, thoracic surgery, major ophthalmic surgery, and complex gynecologic surgery (WHO 2009c).
Results
Minimal Annual Surgical Rates
Six countries defined the efficiency frontier with a combination of the lowest rates of surgery and the highest life expectancies ( and ). Using this surgical efficiency calculation, minimum annual surgical rates observed at life expectancies of 70, 75, and 80 years were 836, 1,504, and 4,547 operations per 100,000 people, respectively. In 2004, 49 countries had rates of less than 836 per 100,000, and 65 had rates of less than 1,504 per 100,000; the vast majority of countries with rates of less than 836 were LICs. Most countries with rates higher than 4,547 per 100,000 were upper-middle-income (UMICs) or high-income countries (HICs). For LMICs to deliver at least 836 operations per 100,000 people, an additional 10.9 million operations per year (95 percent confidence interval of 3.9 million to 30.7 million) would need to be performed in these settings. To achieve a rate of 1,504 operations per 100,000 people would require an additional 28.4 million (95 percent confidence interval of 11.3 million to 71.2 million) operations annually in these countries.
Variable Mortality Rates
Based on the results of our estimates of surgical mortality, however, increased surgical capacity will exact a substantial toll in postsurgical harm and risk for adverse events. The literature search identified 131 articles that met the inclusion criteria and evaluated either mortality or morbidity from cesarean section, appendectomy, and inguinal hernia repair in LMICs. We summarize these results in , , and . Crude mortality rates following cesarean section ranged from 0.5 per 1,000 operations to 51.3 per 1,000. For appendectomy, the rates of death were 0 to 88.6 per 1,000 operations; and for inguinal hernia repair, rates of death ranged from 0 to 411.8 per 1,000 operations. For comparison, historical death rates following cesarean section in Sweden and the Netherlands are 0.4 and 0.53 per 1,000, respectively (Hogberg 1989; Schuitemaker and others 1997); for appendectomy they are 2.4 and 3.0 per 1,000 (Blomqvist and others 2001; Noordzij and others 2010). The mortality rate for elective inguinal hernia repair in Sweden is 1.1 per 1,000, but the rate rises to 29.5 for emergency operations; overall mortality following inguinal hernia repair is 2.4 per 1,000 in Sweden (Nilsson and others 2007). The death rate following elective inguinal hernia repair in Denmark is calculated to be 2.2 per 1,000; however, for urgent cases it is substantially higher at 70.1, with an overall mortality rate of 5.2 per 1,000 operations (Bay-Nielsen and others 2001).
Published Mortality and Morbidity Rates in Selected Countries Following Cesarean Section.
Published Mortality and Morbidity Rates in Selected Countries Following Appendectomy.
Published Mortality and Morbidity Rates in LMICs Following Inguinal Hernia Repair.
LICs and lower-middle-income countries have rates of death that are orders of magnitude greater than those of HICs and UMICs. Compared with Sweden, a country with historically low death rates following these three operations, cesarean section mortality is at least 2 to 4 times higher in Latin America and the Caribbean, 6 to 10 times higher in South Asia, and 100 times higher in Sub-Saharan Africa. The ranges of mortality rates following both appendectomy and inguinal hernia repair are much narrower, but there is frequently a 40-fold mortality increase in Sub-Saharan Africa, and in some cases more than a 100-fold increased risk of death for the same intervention. If LICs and lower-middle-income countries closed the gap in surgical rates to attain a minimum rate of 836 operations per 100,000 people, but surgical mortality remained at 4 percent—a number well within the range of that in Europe—436,000 people would die annually following surgery in these settings. With rates of surgery reaching 1,504 per 100,000 people in LICs and lower-middle-income countries, a postoperative mortality rate of 4 percent would increase this number to 1.14 million deaths per year. Reducing variability in mortality and bringing postoperative mortality to 1.5 percent would prevent more than 200,000 and 700,000 deaths, respectively, for these two surgical rates.
Although these estimates do not control for comorbidities or other demographic, patient, or facility factors, they suggest tremendous excess mortality following surgical interventions. The excessively high death rates following essential surgical interventions such as cesarean section, appendectomy, and hernia repair indicate that safety concerns are justified and demand attention. The variability in mortality in HICs in Europe and North America has been well established, both among and within countries; it is not surprising that this variability is more pronounced in LICs and lower-middle-income countries. However, the extreme rates of death and disability are so dramatic that health systems in these settings need to adopt strategies to improve survival and reduce complications if surgical interventions are to be acceptable and have a meaningful health impact.
Challenges to Surgical Services in LICs and LMICs
The causes of the disparities in mortality described in the previous section are multifactorial and include access to care, transportation options, behaviors, and attitudes. Financial barriers in particular are substantial, especially for the poor and near poor. Catastrophic health expenditure is a major cause of impoverishment, and surgical care can quickly deplete a family’s financial resources (Kruk, Goldmann, and Galea 2009; Nguyen and others 2013; Van Minh and others 2013; Xu and others 2007). Geographic and transportation barriers present a challenge for populations who live at distances from health centers and first-level hospitals. Delays in care-seeking behaviors are exacerbated when populations lack knowledge of health risks or when poor care has led to severe mistrust in health systems (Gauthier and Wane 2011; Kahabuka and others 2011; Kruk and others 2009; Yaffee and others 2012).
These factors pose difficult challenges when evaluating outcomes of surgical care. Delays in presentation for care translate into higher morbidity and mortality, particularly for surgically treatable conditions. Patients may arrive septic, malnourished, physiologically stressed, dehydrated, and anemic; many may arrive moribund. Yet those with life- or limb-threatening conditions tend to derive the greatest benefit from interventions, but at a cost of poorer overall outcomes. The physiologic insults of surgery and anesthesia are substantial, and patients that arrive with only minimal physiological reserves fare poorly. Thus, even situations that might be considered low risk from a population perspective—such as cesarean section in otherwise young, healthy women or appendectomy in healthy young children—demonstrate high levels of complications and mortality compared with similar conditions in HICs. When addressing issues of harm from surgery, providing constructive improvement strategies to health systems and providers who work under less-than-ideal circumstances and operate on patients with more severe, life-threatening comorbid conditions are important.
As demonstrated throughout this volume, LMICs do not meet the basic surgical needs of their populations. Estimating minimum surgical rates using a surgical rate efficiency methodology indicates that most LMICs fall far short of minimum need. In addition, death rates following common operations are substantial and exceed what would be considered acceptable in HICs. If health systems in LMICs improve their surgical numbers without concurrently improving the safety, quality, and transparency of their services, they jeopardize the health of patients seeking care and risk exacerbating mistrust.
Barriers to Surgical Care
Multiple factors contribute to the risks that surgical patients face in resource-constrained environments. Patients, particularly those whose conditions require urgent surgical interventions, encounter significant barriers to effective and timely care. These conditions can rapidly become fatal, and delays in care are associated with significantly worse outcomes. Emergency surgery carries an added risk of mortality due to the extenuating circumstances of the condition, the inability to adequately plan or prepare for the procedure, the inability to control or modify patient-specific risk factors, the logistical difficulties rallying appropriate human or infrastructure resources, and the challenge of intervening with incomplete information. Accordingly, outcomes are worse for emergency interventions compared with elective or semielective procedures. Emergency operations constitute a higher proportion of operations in resource-limited settings, and any barrier that delays presentation imposes a tremendous burden on patients and the health system.
Delays in care have been categorized into three phases: delays in deciding to seek care; delays in reaching adequate health facilities; and delays in receiving adequate, appropriate, and timely care (Thaddeus and Maine 1994). Because access to and delivery of surgical care presents particular challenges in LICs and LMICs, it is especially important to understand the barriers that contribute to delays in definitive interventional care. These barriers can generally be divided into three dimensions: availability, affordability, and acceptability (Grimes and others 2011; McIntyre, Thiede, and Birch 2009). Each of these dimensions causes delays experienced by patients in need of surgical care ().
Relationship of Barriers to Care and Delays in Obtaining Care.
Availability of Care
Availability refers to the following:
The relationship between the location of health care facilities with the capacity to provide appropriate services, the location of the population needing them, and the transport opportunities available
The ability and willingness of care providers to serve the population in accordance with the type and severity of the presenting condition
The timing and hours of available services and the times patients seek care
The range, quantity, and quality of services provided and the nature and extent of the health needs of people seeking care
A tremendous obstacle to early presentation is the geographic distribution of health facilities with the capacity to recognize and deal with surgical issues (Dye and others 2010; Hang and Byass 2009; Macharia and others 2009; Mock, nii-Amon-Kotei, and Maier 1997; Parkhurst, Rahman, and Ssengooba 2006). Health centers with the sophistication to provide surgical care tend to be located in more populous areas, and LMICs with the lowest surgical volumes frequently have large rural populations. Timely transport to surgical care is critical, yet road and transportation infrastructure can be lacking or intermittent (Macharia and others 2009; Mock, nii-Amon-Kotei, and Maier 1997; Seljeskog, Sundby, and Chimango 2006). Finally, social norms can prevent early presentation; consultation with traditional healers, village elders, or heads of family may delay access to the formal health care system (Briesen and others 2010; Hang and Byass 2009; Mock, nii-Amon-Kotei, and Maier 1997; Parkhurst, Rahman, and Ssengooba 2006; Seljeskog, Sundby, and Chimango 2006).
Once patients do arrive at facilities, the requisite durable and consumable supplies and equipment are often inadequate (Lebrun, Chackungal, and others 2013; Lebrun, Dhar, and others 2014; Macharia and others 2009). The availability of personnel and services is often intermittent, particularly at night. Above all, the human resources for health are frequently lacking. Surgical skill requires education, training, and experience; trained clinicians are not always available or capable of performing specific surgical tasks; the status of anesthesia services, as discussed in chapter 15, is even more dire. Providers’ confidence in their skills is particularly important for the provision of surgical care; recent research has indicated that lack of confidence—due to lack of training, experience, or surgical assistance—may be a primary cause of triage and transfer, as well as a major barrier preventing immediate intervention (Bowman and others 2013; Petroze, Nzayisenga, and others 2012).
Affordability of Care
Affordability, the match between costs of services and the ability of individuals to pay, presents a major challenge because of the following factors:
Price of services at the point of delivery
Direct costs associated with transportation, food, and lodging
Indirect costs, such as lost income or productivity
The ability of individuals to pay also relates to their personal wealth and assets, eligibility for financial support from financing mechanisms, and amount of indirect costs incurred. Transportation costs can be unaffordable, and, when combined with prohibitive out-of-pocket expenses, frequently delays early consultation (Afsana 2004; Mock, nii-Amon-Kotei, and Maier 1997; Nwameme, Phillips, and Adongo 2013). In addition, affordability refers not just to the ability of an individual or family to pay for care but also the potential impact of that payment on the household, and the manner and timing of payment. For example, up-front charges may prevent early assessment and definitive management as families seek to secure necessary funds for payment for services (Kruk, Goldmann, and Galea 2009).
Acceptability of Care
Acceptability refers to the expectations, behaviors, perceptions, and attitudes inherent in medical encounters. Providers’ attitudes are affected by stereotypes, chief complaints, and the manner of presentation. Patients’ attitudes are similarly affected by stereotypes, as well as by perceptions of respect, efficiency of care, and trust in the integrity of the system. Of particular concern with surgical intervention is the personal security of clinicians; deaths following surgery may be blamed directly on surgical providers, and family and community members may seek retribution, regardless of premorbid conditions or cause of demise (Burch and others 2011; Malik and others 2010).
Many of these domains interact in ways that magnify delays. Concerns about financial commitments, compounded by mistrust of the health care system, a lack of transparency, and poor quality, lead to long delays in treatment-seeking behavior. People in many LMICs may justifiably perceive surgical care to be a poor investment of resources.
Anesthesia Safety
The safe provision of anesthesia is a critical consideration in establishing and expanding the capacity for surgical care. Improvements in anesthetic monitoring and techniques have led to dramatic improvements in its safety profile in HICs and UMICs. In many settings with low levels of human resources, however, anesthesia is provided by nonphysician clinicians or technicians, or even by the operating surgeons. Poor training, supervision, and monitoring standards all contribute to high mortality from the administration of anesthesia.
Although the rate of overall deaths due to anesthesia is estimated to be 34 per 1 million anesthetics administered, profound differences exist among countries and settings. Bainbridge and others (2012) report that in low human development index countries, deaths solely attributable to anesthesia are estimated to be 141 per million, compared with 25 per million in high human development index countries. Critically, anesthesia mortality in LMICs continues to be a major problem, with death rates as high as one per 500 (Walker and Wilson 2008). Anesthesia in HICs and UMICs has improved only relatively recently, with changes in monitoring and increased standardization responsible for a 100-fold reduction in mortality over the past 40 years—34 deaths per million instances of anesthetics administered in the 1990s and 2000s, down from 357 deaths per million before 1970 (). Low professional standing, inadequate basic monitoring equipment, and a lack of professional standards all contribute to the current disparity between HICs and other countries.
Meta-Regression for Risk of Death due Solely to Anesthesia, 1939–2009.
Perioperative Safety
Surgical intervention, by its nature, involves risks. High-quality and high-resource systems still fail to provide proven interventions every time for every patient.
In the United States, the failure to adhere to basic WHO standards occurs in 6 percent to 20 percent of operations, indicating substantial room for improvement (Stulberg and others 2010). Individual care standards are being used in pay-for-performance initiatives to help improve quality by linking it to reimbursement. When individual standards are evaluated in isolation, however, they frequently fail to demonstrate improvements in outcomes with improving levels of compliance () (Stulberg and others 2010). Multiple care standards need to be evaluated as a composite whole; partial completion of tasks does not always deliver a partial benefit, and improvements often require total compliance to result in improved outcomes (Nolan and Berwick 2006).
Surgical Care Improvement Project: Infection-Prevention Process Measures.
This all-or-none compliance likely indicates that systems able to achieve high compliance rates with multiple standards-of-care processes are highly organized and functioning efficiently; accordingly, they are able to deliver on difficult-to-measure but essential components of care, such as communication and information transfer (Weiser 2010). In LMICs, lack of compliance may be especially germane—poorly used or misallocated resources constitute a drain on an already stressed health system.
Compliance with care standards in LMICs is frequently poor. In a study in India, Das and others (2012) selected a random sample of health care providers in rural Madhya Pradesh and urban Delhi to receive a visit from a “standardized patient” trained to present one of three scenarios: unstable angina, asthma, or a parent describing dysentery in a child at home. These standardized patients were then debriefed following their clinic visit to assess the quality of care and compliance with care checklists and best practices. Providers in both locations did poorly in asking appropriate questions and performing appropriate examinations (33.7 percent and 31.8 percent in Madhya Pradesh and Delhi, respectively); making the correct diagnosis (12.2 percent and 21.8 percent in Madhya Pradesh and Delhi, respectively); and identifying pertinent clinical issues and making appropriate recommendations for treatment (30.4 percent and 45.6 percent in Madhya Pradesh and Delhi, respectively) (). The rate of unnecessary or harmful treatments exceeded 40 percent in Madhya Pradesh. Despite the range of provider education, from no formal medical education to fully licensed physician, only a small difference was observed in adherence to care standards; no difference was observed in arriving at the correct diagnosis or providing the correct treatment.
Adherence to Checklist of Questions and Examinations for Unstable Angina, Madhya Pradesh, 2010.
Such studies point to troubling discrepancies between what is known and taught about care standards, on the one hand, and actual practice patterns, on the other hand. Adherence to known practice standards is a hallmark of high-quality health organizations; yet, similar to what is found in primary care and general practice settings, surgical delivery in HICs and LMICs alike frequently fails to follow standards of care, despite well-described strategies and techniques for improvement.
Postoperative Care and Safety
In addition to the risks during surgery, patients are at high risk during postoperative recovery. The two most common causes of complications within the first week of surgery are bleeding and infections. Additional causes of delayed morbidities include blood clots, heart attacks, pneumonia, and stroke. Anticipating potential complications, and either preventing them (for example, by prophylaxis for venous thromboembolism) or identifying the signs and symptoms and intervening early and aggressively, are essential to reduce these risks.
An important study established the prominent role of a mature system of postoperative care in managing complications and preventing them from resulting in death (also known as failure to rescue). Ghaferi, Birkmeyer, and Dimick (2009a, 2009b) found that although baseline complications rates were strikingly similar in institutions across the United States, mortality rates following these complications varied dramatically ( and ). These findings confirm earlier research suggesting that the primary difference in outcomes among hospitals is not due to differences in complication rates but to differences in the rates of failure to rescue (Silber and others 1992).
Rates of All Postsurgical Complications, Major Complications, and Deaths after Major Complications, According to Hospital Quintile of Mortality.
Rates of Deaths, Complications, and Death after Major Complications for Five Operations with the Highest Number of Deaths, According to Hospital Quintile of Mortality, 2005–07.
Further research has demonstrated that higher-volume hospitals appear to have a better ability to recognize, intervene, and save patients undergoing high-risk procedures from death and complications following surgery (Ghaferi, Birkmeyer, and Dimick 2011). Complications must be anticipated following high-risk procedures; the ability to recognize, diagnose, and treat complications separates the high performers from the poor performers. The quality of communication and the systems of care, and the skills and capacity of ancillary services—such as availability of intensive care and the presence and experience of specialized services—appear to be important factors for improving outcomes following complications.
Strategies for Improving Surgical Care
Several effective strategies have been identified for improving surgical outcomes in LMICs. These strategies include the adoption and use of basic technologies, the development of monitoring standards, and the use of surgical safety checklists. Organizational and management strategies also appear to be important. Essential to all of these interventions is a mandate to measure the delivery of care and its impact on health. These low-cost interventions, which can dramatically lower postsurgical mortality rates, demand prioritization by health systems seeking to improve access and surgical service provision.
Anesthetic Monitoring and Safety
One of the most important contributions to improved surgical safety has been the development of basic standards of anesthetic monitoring. The Harvard monitoring standards for intraoperative anesthesia care formalized a set of medical standards of practice that have become de facto international standards endorsed by the World Federation of Societies of Anaesthesiologists (Eichhorn and others 1986; WFSA 2008). The standards include the continuous presence of trained anesthesia providers and the uninterrupted monitoring of oxygenation, ventilation, and perfusion. Today, adherence to these standards in HICs is essentially universal; however, this was not the case a mere three decades ago, and it is far from standard practice in many LMICs.
In addition to continuous monitoring techniques, anesthesia delivery systems have been standardized, with safety engineered into the instruments themselves. Inhalational anesthetic machines are now engineered to be redundant; safety features, such as auto-lock mechanisms, prevent lethal hypoxic gas mixtures. Despite the 100-fold plunge in anesthetic-related mortality rates in HICs and UMICs during the past 40 years, anesthetic mortality in LMICs is a major problem due to lack of professional stature, training, and credentialing of anesthesia providers; deficiencies in basic monitoring equipment; and failure to adhere to strict standards of care.
One critical mechanism for anesthesia monitoring is the use of pulse oximetry. Although the continuous monitoring of blood oxygen levels using a pulse oximeter is considered an essential standard, more than 77,000 operating rooms worldwide do not have this basic monitoring device (Funk and others 2010). Pulse oximetry can alert anesthesia personnel to drops in oxygenation before clinical signs become apparent, allowing for corrective actions before hemodynamic instability or lethal arrhythmias occur. In Moldova, an implementation program supplying pulse oximetry equipment in conjunction with provider training on the use of a surgical safety checklist reduced postoperative deaths and complications (Kwok and others 2013). Use of pulse oximetry is highly cost-effective as well, with the cost per DALY averted from anesthetic mishaps due to improved monitoring at US$374 for a standard commercial oximeter and US$115 for a smaller hand held device (Burn and others 2014). A concerted effort is underway through the Lifebox Foundation to provide pulse oximetry monitoring capabilities to every operating theater in the world (http://www.lifebox.org).
Surgical Checklists
Standardization of care is essential because of the tremendous magnitude of interactions and care processes that occur during even simple surgical procedures. Complex patient characteristics, therapeutic options, technical demands, and team dynamics require specific strategies for organizing care protocols and service delivery. The effective use of checklists by teams during surgery has cut mortality rates by up to 50 percent.
In 2008, the WHO codified a set of basic surgical standards into guidelines for safe surgery. Researchers transformed these guidelines into a simple, 19-item checklist to be used during the perioperative period and conducted a multicenter trial assessing the efficacy of this safety tool on postoperative morbidity and mortality (). In a pre- and postanalysis of nearly 8,000 surgical patients, use of this checklist nearly doubled adherence to basic perioperative safety standards, including confirmation of the procedure and operative site, administration of antibiotics, use of pulse oximetry for monitoring, objective airway assessment, and completion of instrument and sponge counts at the conclusion of the operation. Use of the checklist reduced deaths by more than 47 percent and cut complication rates by 35 percent (Haynes and others 2009). This beneficial effect was maintained in a subanalysis of urgent and emergency cases (Weiser, Haynes, Dziekan, and others 2010).
World Health Organization Surgical Safety Checklist.
Several other large, well-designed studies have confirmed the substantial enhancements to surgical safety that checklists provide. Following the introduction of a comprehensive perioperative checklist in six hospitals in the Netherlands, postoperative complications and deaths dropped by 30 percent and 47 percent, respectively; in five control hospitals, no improvements were noted during the same period (de Vries and others 2010). A second study in the Netherlands virtually repeated the original multinational WHO investigation, demonstrating improvements in postoperative mortality that strongly correlated with checklist compliance (van Klei and others 2012).
Just as a pilot’s checklist does not instruct a pilot how to fly a plane, surgical checklists do not dictate how clinicians should deliver care; instead, checklists help confirm critical steps, prompt consideration of extenuating or unusual factors, and stimulate or facilitate team communication. These processes are particularly important in the complex and multidisciplinary environment of surgery. Checklists are often a critical part of crew resource management, a method of team training that promotes shared mental models for care and conduct that has been implemented in many organizations and sectors in which high reliability and fidelity are paramount, such as aviation and nuclear power. This method has been extended to surgical teams; it has been observed, for example, that cardiac surgery teams that consistently work together are more efficient and have better outcomes than those with rotating members (Carthey, de Leval, and Reason 2001; de Leval and others 2000). Because this method is often not possible in urgent circumstances or when human resources are limited, checklists can play an essential role in promoting consistent processes of care. A study conducted at 74 Veterans Administration hospitals in the United States demonstrated significant improvements in mortality compared to controls following a full-day team training program that included implementation and training in the use of checklist-guided briefings and debriefings (Neily and others 2010).
Checklists have become an established standard of surgical care globally (Birkmeyer 2010). Their effectiveness has demonstrated the accuracy of previous estimates suggesting that at least 50 percent of existing surgical mortality is preventable. Checklists are most effective when they are implemented, not as a tickbox exercise, but as a means to reinforce communication, prompt genuine dialogue and discussion of critical information, and facilitate prospective feedback and quality improvement (Weiser, Haynes, Lashoher, and others 2010). Large-scale regulatory mandates alone appear not to be effective in fostering effective adoption (Urbach and others 2014). Implementation has been found to require local champions from all disciplines, support from leadership, monitoring of progress, and involvement of frontline clinicians (such as through team training) and not just administrators. Such an approach has been followed in Scotland, leading to a statistically significant drop in inpatient surgical death rates from 2011 after three years of flat mortality rates. The Scottish government has documented more than 9,000 lives saved (Leitch 2012).
The challenge of conducting multidisciplinary implementation programs in LMICs raises legitimate concerns about ability to scale up such programs globally. However, a follow-up WHO study in Honduras, Moldova, and Zambia confirms the ability to implement and replicate large improvements in safety and outcomes (Kim and others 2012).
Management Practices
Effective and efficient management strategies are an essential component in the smooth functioning of health facilities. Numerous econometric studies have looked at management practices in industry and business and identified characteristics that affect productivity. Two economists from Stanford University and the London School of Economics conducted a series of interviews with midlevel managers from a range of medium-sized manufacturing firms in France, Germany, the United Kingdom, and the United States, using a survey to assess four domains of management: operations, monitoring, targets, and incentives (Bloom and van Reenen 2007). High scores in these domains were strongly related to higher productivity and profitability, as well as to the longevity of the company.
In LMICs, however, multiple factors affect the performance of industry, particularly for the worse. Management practices are suboptimal for various reasons, including lack of knowledge of optimal management practices, reduced competition, high proportion of family ownership, lack of delegation of decision making because of fear or mistrust, reduced incentives, and poorly allocated financing. Bloom and van Reenen (2010) and Bloom and others (2010) note that similarly sized local firms in LMICs were severely lacking in management practices, with correspondingly lower overall productivity. Although economic environments and organizational factors played a role (Bloom, Sadun, and van Reenen 2012), introducing management practices through an intensive consulting process resulted in massive improvements in efficiency and productivity (Bloom and others 2010; Bloom and others 2013).
In health care, the management practices evaluated by Bloom and van Reenen (2010) roughly translate to operations management, quality evaluation, goal-setting, and talent management. Their scoring mechanism has been used to evaluate hospital management practices and its subsequent correlation with patient outcomes across Brazil, Europe, India, and the United States. They find tremendous variability in management practices within countries, as well as a particularly large proportion of poorly managed hospitals in LMICs. A McKinsey study looking at hospitals in the United Kingdom and the United States determined that an increase of 1 point on Bloom’s management practice survey scale is associated with a decrease of 6 percent to 7 percent in 30-day mortality following acute myocardial infarction, an increase in hospital earnings of 14 percent to 33 percent, and an overall improvement in patient satisfaction (Bury and others 2007).
Although there is a paucity of research in the area of hospital management in LMICs, it is reasonable to infer that management practices affect the organizational structure, efficiency, and even safety of the health system. In one of the first studies of this kind, Funk and others (2013) suggest that more robust management practices are associated with enhanced surgical productivity. Unfortunately, many first-level rural and urban referral hospitals in LMICS are likely to be plagued by poor management practices similar to their business and manufacturing counterparts. Such problems lead to waste and poor resource allocation, and potentially even to fraud and abuse. It remains to be seen whether improvements in management translate into improved surgical productivity in these settings and, if so, the mechanisms by which such improvements occur.
One essential mechanism that management uses to enhance the quality of care is the implementation of surveillance and evaluation practices, allowing quality improvement (QI) programs to be targeted to identified weaknesses. These practices range from very simple outcome assessments, such as Morbidity and Mortality Conferences (M&M), to more complex monitoring, such as ongoing surveillance of complications, adverse events, and errors, and use of risk-adjusted mortality. Many hospitals in LMICs have some type of QI activity, even if limited to periodic M&M conferences. Often, the effectiveness of these efforts could be increased by simple measures, such as better recording of problems discussed, more purposeful enactment of corrective action, and monitoring of the outcome of the corrective action. A WHO review of the effectiveness of QI programs for trauma care identifies 36 studies, 34 of which report improvements in patient outcomes (including mortality) or process of care after a new QI program or method is introduced (Juillard and others 2009). Two articles report no change, and no articles report a worsening of any outcome; five articles also report cost savings. Most of the articles were from HICs; two were from Thailand. A summary of the model QI program in Thailand is provided in chapter 3 on trauma care in this volume. The WHO has outlined a multimodal approach to QI processes for trauma systems through the use of morbidity and mortality, preventable death panel reviews, audit filters, and the establishment of trauma data bases and surveillance systems (WHO 2009a).
Measurement Strategies
The measurement of outcomes of intervention, regardless of the service provided, is essential to ensure that the effects of care are aligned with intent and that resources are used efficiently, effectively, and with the least harm to patients. Practitioners, facilities, and health systems require information on surgical capacity, throughput, and results to determine how such service lines perform. Other notable public health successes, such as improvements in maternal and neonatal health, HIV care, and control of poliomyelitis and malaria, have been dependent on surveillance (Ceesay and others 2008; Ronsmans and Graham 2006; WHO 2000, 2005, 2007). Surveillance is equally essential in optimizing access to and the safety of surgical care; the absence of data on surgical delivery and outcomes perpetuates the neglect such therapy receives in resource-constrained settings (Weiser and others 2008, 2009).
The WHO has proposed a set of standardized metrics for surgical surveillance at the national level that have been tested and validated (WHO 2009d), and is included in annex 16B. These metrics include the number of operating rooms in each country, the numbers of trained surgeons and trained anesthetists in each country, the number of procedures performed in operating rooms in each country, the number of deaths on the day of surgery, and the number of in-hospital deaths after surgery (Weiser and others 2009) (). Although each of these metrics has important weaknesses that must be acknowledged, all can be obtained and reported in a straightforward manner.
Standardized Statistics for Surgery: Definitions, Rationale, and Data Sources.
National-level metrics nonetheless require the interest, investment, and commitment of the central government or agency responsible for collecting, analyzing, and disseminating such information. Local efforts at QI should not be limited to crude, population-level data collected to measure health system performance. Several basic metrics must be adopted by facilities and health systems to improve the quality and delivery of care (). These could include the following:
Proposed Facility-Level Metrics for Quality Improvement, in Addition to the Standardized Statistics Described in Table 16.8.
Structural metrics, such as the availability of essential monitors like pulse oximetry; equipment, such as anesthetic machines and autoclaves; and consumable and reusable materials, such as surgical equipment, devices, and antibiotics
Process metrics, such as hours of operation, duration of operations, number of operations per operating room, appropriate administration of antibiotics, and use of and compliance with checklists
Outcome metrics beyond mortality, such as surgical infections and reoperation, and other perioperative complications such as pneumonia, renal failure, heart attack, or stroke
One of the issues of greatest concern is the misuse of such metrics to deny care to the most frail and vulnerable populations. Health systems that manipulate their outcomes by increasing inappropriate services, failing to intervene, and underreporting mortality succumb to perverse, negative incentives that divert essential resources and inhibit care for the sickest patients. Under ideal circumstances, surgical statistics should help health systems improve the delivery and safety of surgical care by creating benchmarks for improvement rather than being used for punishment or comparison across fundamentally different organizations, environments, and populations (WHO 2009b). Attempts at comparisons across systems, countries, and health settings ignore variations in patient condition and complexity of procedure. Hospitals and health systems that wish to evaluate differences between facilities and practitioners must account for the characteristics of the patients, case mix, and urgency—all of which require robust and sophisticated data collection that is frequently beyond the capacity of overworked or underfinanced health systems. However, countries and health systems in LMICs that are able to collect such metrics will have a foundation of information upon which they can improve and sustain surgical care to the betterment of their nation’s health.
Any complete discussion of quality clearly encompasses more than simple measures of mortality and complications. Important outcomes also include, among others, the nuanced measures of functional recovery, control of pain, and satisfaction with care. While meaningful, these issues are beyond the scope of this chapter, as are the potentially important strategies for improving surgical capacity through the use of physician extenders and task sharing, and the aggregation and centralization of cases to take advantage of volume-outcome relationships.
Much work is needed to strengthen surgical systems of care, and the investments are likely to be considerable. Given the barriers to access and delivery of needed surgical services, investments are necessary at the facility and institutional level, as well as for the progressive financing of health protection and communication and transportation infrastructure. Improving anesthetic monitoring and safety, implementing surgical checklists, refining management practices, and instituting measurement and surveillance techniques could dramatically improve care within existing health systems. However, designing, implementing, and scaling these interventions in LMICs will take considerable resources because each strategy for improvement requires training, infrastructure, an information management system, and political will. Even though little is currently known about the actual investment and recurrent costs of introducing and scaling up these strategies, they are likely to be highly cost-effective.
Acknowledgments
The authors would like to thank Charles N. Mock, University of Washington, for his help and guidance with this chapter; Pablo Uribe Leitz, Lydia Maurer, and Joshua Jaramillo, Stanford University, for help with data extraction and analysis; and Marissa Wagner, University of Texas, Houston, and Kimberly Brayton, Stanford University, for their assistance with the initial creation and structuring of the database. They also would like to extend appreciation to Rebecca Kim and Lyen Huang, Stanford University Medical Center; Luke Funk, University of Wisconsin Medical Center; and Kendra Bowman, Brigham and Women’s Hospital, for their willingness to share preliminary data. Kristan Staudenmayer, Stanford University, helped review and proof drafts. Riti Shimkhada and John Peabody, Qure Healthcare, provided insightful feedback on metrics for measurement; and Nick Bloom, Stanford University, offered invaluable insights regarding his work on management practices.