Catholic Healthcare Partners CHP
Crew resource management CRM
Hospital Corporation of America HCA
Hypoxic-ischemic encephalopathy HIE
Intensive care unit ICU
Institute of Medicine IOM
Nulliparous term singleton vertex NTSV
Situation, background, assessment, recommendation SBAR
Health care quality has been defined by the Institute of Medicine (IOM) as, “the degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge.” Under the IOM framework, safety is part of the larger concept of health care quality. In this chapter, we will describe ways to enhance patient safety and to attempt to measure and achieve quality care in obstetrics.
Patient safety has become an increasing focus of the attention of the medical community. Although the reasons for the increased focus are doubtlessly multifactorial, one important contributor has been the recognition of the number of preventable adverse events that occur. The relatively high number of medical errors has been documented in the Institute of Medicine (IOM) report To Err is Human . That report, published in 1998, noted that between 44,000 and 98,000 people die in hospitals each year as the result of medical errors. This report, correspondingly, noted the importance of error reduction as a means to improved health care and patient outcomes.
Because obstetric admissions are the leading cause of hospitalization for women in the United States, accounting for over 4 million hospital discharges each year, pregnant women are at particular risk of encountering a medical error. Emphasizing the importance of obstetric safety further is the fact that each obstetric admission has the potential to affect the health not only of a single patient but of both a mother and her infant.
Frequency of Preventable Obstetric Adverse Events
Many studies that have sought to determine the frequency of preventable obstetric adverse events have used retrospective designs and studied in detail cases in which adverse outcomes have occurred. The overall frequency of preventable adverse events on a given obstetric unit cannot be determined from these studies. This is particularly true given that adverse events, such as administration of a medication to which the patient is stated to be allergic, may occur that do not result in an actual adverse outcome (e.g., anaphylactic shock). However, these studies do give insight into the frequency with which that adverse outcome may have been preventable when such an outcome occurs.
Geller and colleagues analyzed cases of morbidity and mortality that occurred at their institution. These cases included those with severe and “near-miss” morbidity. It should be emphasized that so-called near-miss morbidity is not so named because a woman nearly misses having morbidity but because she nearly misses a mortal event. Thus in their framework, maternal morbidity has a spectrum of severity, and none is more severe than near-miss morbidity. In their study, morbid and mortal events were often found to be preventable, and a preventable event was defined as one that could have been avoided by any action or inaction on the part of the health care provider (e.g., mismanagement of patients, failure or delay in diagnosis), the system (e.g., failure in communication), or the patient (e.g., noncompliance). Moreover, women who had the most adverse outcomes were more likely to have had preventable events: 16% of cases with severe morbidity were judged preventable versus 46% of near-miss events and 41% of mortal events.
Berg and colleagues focused exclusively on maternal deaths that occurred in North Carolina. They studied the circumstances that surrounded the pregnancy-related deaths of 108 women between 1995 and 1999. Deaths were considered potentially preventable if (1) preconception care and counseling could have improved outcome; (2) the patient had not adhered to medical advice; (3) the structure and functioning of the health care system was suboptimal; or (4) clinical care was not satisfactory. In their study, 40% of deaths were considered to be preventable, although this frequency significantly differed on the basis of the primary underlying cause of death. For example, 93% of deaths related to hemorrhage, but only 22% of deaths related to cardiomyopathy, were thought to be preventable.
White and associates examined 90 cases that resulted in litigation and in which the claims had been closed. It should be noted that these cases included both gynecology and obstetrics cases, although the majority were related to obstetrics. In their study, 78% of these cases were thought to have a contributing factor that was potentially avoidable. Clark and colleagues also analyzed claims data and focused only on those claims related to perinatal care. Their findings were quite similar to those of White and associates; specifically, they noted that 70% of cases were potentially preventable and related to the care the patient received in the hospital.
Forster and coworkers did attempt to quantify the frequency of adverse events that occur on an obstetric unit through the use of a prospective study. These investigators placed an observer in the labor and delivery unit during the weekday hours of a 6-week period. This observer was trained to ascertain poor patient outcomes, procedural errors, and unsafe working conditions. Cases of concern that were ascertained by the observer were then further evaluated by a multidisciplinary team. The primary outcome of the study was the occurrence of a “quality problem,” defined as the occurrence of either an adverse event—that is, an adverse outcome due to health care management, as opposed to progression of natural disease—or a potential adverse event (i.e., “defective processes that have a high likelihood of causing harm”). Of the over 400 patients who were cared for during the study period, 5% experienced a quality problem (i.e., 2% had an adverse event and 3% had a potential adverse event). Sixty-six percent of the adverse events that occurred were judged to be due to errors in health care delivery.
Factors That Contribute to Preventable Adverse Events
The literature that exists would suggest that multiple contributors are involved in the occurrence of adverse obstetric events. Data from the Joint Commission, for example, illustrate the contributions that multiple factors make in the occurrence of major adverse maternal and perinatal outcomes. For example, their analysis of major maternal adverse events revealed multiple root causes that included lack of adequate communication, training, staffing, and patient assessment. Other investigators similarly have found that preventable adverse events or potential events cannot be traced to one simple and easily remediable cause but are multifactorial and due to a complex interplay of factors. Geller, White, Forster and their colleagues all have demonstrated the many different factors present at both the provider and systems level and that have contributed to adverse events.
Despite the many factors that have been implicated in the occurrence of preventable adverse events, it is worth noting that communication and “systems” issues that transcend simple individual error have consistently been found to be predominant etiologies in the occurrence of these events. Systems issues is a term that refers to problems that stem not from one individual’s actions but from the interconnected relationships of people and institutional policies. With regard to the sentinel events in maternal care analyzed by the Joint Commission, a communications issue was judged to be a root cause in over 80% of cases. This frequency far outstripped the next most frequent factors, competency and patient assessment, which were present in fewer than 40% of cases. In their review of closed claims, White and associates noted that inadequate communication among providers was the single most common preventable factor associated with the claim. Similarly, in their prospective study, Forster and colleagues noted that “systems” issues were the most common reason that their trained observer was alerted to further assess for the possibility of a quality problem.
Approaches to Improve Obstetric Safety
The previous discussion highlighting the different factors that contribute to patient safety suggests that efforts to improve safety may require multiple different approaches. The potential need for a multifaceted approach is further suggested by the different levels within an organization at which these factors can manifest. Specifically, key components required for the prevention of adverse events occur at (1) the individual level, such as the level of education or training provided to workers; (2) the group level, as with team effectiveness and communication; and (3) the structural level, like the standardization of processes within an organization. Correspondingly, attempts to enhance patient safety within obstetrics have utilized, in general, several different types of modalities. The theoretic underpinning to believe that these modalities should be effective—as well as the evidence supporting their use, with a focus on obstetrics—is discussed below.
Checklists and Protocols
Because of the complexity of health care processes and the corresponding potential for errors in communication, one approach to improve patient safety has been the introduction of standardized approaches to patient care. These approaches have taken the form of protocols, mandatory items for completion to lead the user to a predetermined outcome, and checklists, a list of action items or criteria arranged in a systematic manner that allow the user to record the presence or absence of the individual items listed to ensure that all are considered or completed. Although both are concerned with standardization, checklists provide explicit lists of items, actions thought to act as a cognitive aid due to grouping related items in an organized fashion to improve recall performance.
Pronovost and colleagues demonstrated the potential utility of checklists in their study regarding catheter-related bloodstream infections in the intensive care unit (ICU). In this study, checklists that detailed five key actions required during any central catheter placement were introduced in 108 ICUs throughout the state of Michigan. Of note, these key actions were evidence based; also, the checklists were not merely posted, they were supported by local leaders versed in the supporting evidence, who provided feedback with regard to optimal methods for implementation. Not only did the incidence of catheter-related bloodstream infections significantly decline by 3 months after implementation (incidence rate ratio, 0.62; 95% confidence interval [CI], 0.47 to 0.81), but this incidence continued to decline until 18 months (incidence rate ratio, 0.34; 95% CI, 0.23 to 0.50) after implementation, which was the end point of data collection.
A checklist to guide the administration of oxytocin was developed by Clark and associates, who then implemented this checklist in a private hospital setting. The outcomes of the 100 women prior to checklist implementation were then compared with those of the 100 women who received oxytocin after the implementation. No difference was seen in the duration of labor, duration of oxytocin administration, or operative delivery after implementation, although the maximum dose of oxytocin used was significantly lower, as was the frequency of newborns with one or more complications ( P = .049).
The potential benefits of a standardized approach to the evaluation and management of preeclampsia has been demonstrated by Menzies and colleagues. After establishing a set of best practices, this group of investigators introduced these practices for preeclampsia management at British Columbia Women’s Hospital. Among women with preeclampsia who were managed after the standardized approach was put into place, 0.7% experienced the composite end point of maternal adverse outcomes, which was an 86% reduction compared with the preintervention frequency of 5.1% ( P < .001). Adverse perinatal outcomes were also reduced, although this finding did not reach statistical significance (odds [OR], 0.65; 95% CI, 0.37 to 1.16).
Nevertheless, a word of caution regarding checklists and protocols is warranted: The mere existence of one of these on a unit cannot be assumed to automatically result in improved care. As demonstrated by Pronovost and associates, the presence of a checklist may enhance care when its components are evidence based and when its use is championed by members of the organization. Moreover, it is ideal to demonstrate—using either traditional scientific or quality improvement study designs—that its introduction is associated with improvement in the care provided or outcomes achieved. Conversely, checklists or protocols can be present but still not translate into any tangible change in health care whatsoever. One set of investigators, for example, observed that the introduction of surgical safety checklists in Ontario, Canada, was not associated with improvements in operative complications or mortality. Bailit and coworkers have demonstrated a similar finding in obstetric care. In their analysis, obstetric units had similar outcomes related to postpartum hemorrhage or shoulder dystocia regardless of whether they had protocols for these events.
Simulation refers to the recreation of an actual event that has previously occurred or could potentially occur. Simulation may be used to enhance patient safety in that an action or procedure can be repeated, thereby improving the execution of that action or procedure without ever exposing providers or patients to harm. Essentially, simulation of events is an opportunity for health care workers to prepare and train for interventions. Although simulation may have benefits for any type of obstetric procedure (e.g., vaginal delivery), it has often been studied in events such as shoulder dystocia and eclampsia. In these occurrences, simulation may be particularly helpful not only for the novice but even for experienced professionals, who can maintain their skills in managing unpredictable and uncommon events.
Data from multiple studies suggest that simulation may enhance several aspects of shoulder dystocia management, which includes communication among the team members, performance of the maneuvers necessary to relieve this problem, and documentation of the event. Moreover, these enhancements have been demonstrated for providers during both their residency training and postresidency experiences.
In a randomized trial, Deering and colleagues demonstrated that those residents who were assigned to train for a shoulder dystocia using an obstetric birth simulator were significantly more likely to utilize maneuvers in a timely and correct fashion in a subsequent simulation than those residents who did not undergo initial simulation training. Furthermore, when judged by a blinded observer, those residents who had undergone simulation training scored higher on measures of overall performance and preparedness.
A study that examined presimulation and postsimulation training outcomes among both residents and attending physicians was performed by Goffman and coworkers. In this study, participants underwent a simulation of a shoulder dystocia event followed by a debriefing that included (1) a brief lecture about shoulder dystocia, (2) a review of the basic maneuvers and a basic algorithm for management of shoulder dystocia, (3) a discussion on optimizing team performance during an obstetric emergency, (4) a review of the key components of documentation, and (5) a review of the digital recording of the simulations and discussion of provider performance. During a second, unanticipated shoulder dystocia simulation, providers demonstrated significant improvements in communication, use of maneuvers, and overall performance.
Similar results were obtained by Crofts and colleagues in their multicenter comparison of presimulation and postsimulation outcomes. This study additionally attempted to ascertain whether high-fidelity simulation using a mannequin with a high degree of biofidelity would enhance performance to a greater extent than low-fidelity simulation using a simple doll-like mannequin by randomizing participants to one of these options. Regardless of randomized group, performance during a simulated shoulder dystocia was improved subsequent to simulation training. Although some measures of performance (e.g., total applied force) were improved to a greater extent in those participants who had undergone high-fidelity training, many other measures (e.g., peak force, use of maneuvers) were no different between the high- and low-fidelity groups. Improvement in performance of simulated shoulder dystocia events has been demonstrated to persist up to 12 months after training.
The previous studies have examined outcomes during simulation events but have not demonstrated improvements in actual clinical outcomes related to simulation training. The few observational studies of clinical outcomes that do exist, however, seem to support the assertion that exposure to simulation may improve outcomes related to shoulder dystocia. In one study at a single institution, investigators examined outcomes both before and after initiation of a shoulder dystocia training program that utilized training in and practice of obstetric maneuvers. After the training program, a significant increase was reported in the frequency with which appropriate maneuvers were used for shoulder dystocia, and a significant reduction was reported in neonatal injury at birth after shoulder dystocia (9.3% to 2.3%; relative risk [RR], 0.25; 95% CI, 0.11 to 0.57). Grobman and associates also used simulation to help train providers to respond to a shoulder dystocia. They did not simulate specific maneuvers but instead emphasized the coordinated response and communication of the team once a shoulder dystocia was diagnosed. After simulation was used in this way, the frequency of brachial plexus palsy at discharge among neonates who had experienced a shoulder dystocia dropped from 7.6% to 1.3% ( P = .04).
Studies that have evaluated simulation training for eclamptic seizure management have demonstrated improvements in response as measured during additional simulations. For example, after having staff in a single tertiary care center participate in simulated eclamptic events with a debriefing session, Thompson and colleagues found that patient resuscitations in posttraining simulation of eclampsia were better managed. Of note, these simulations occurred in the actual delivery unit, as opposed to a separate simulation laboratory, and these researchers used their initial simulations not only for training but also for elucidating site-specific barriers to an optimal response (e.g., an inefficient paging system). This ability of simulation to identify systems-level barriers to best practice has been noted by others as well.
Ellis and colleagues also examined pretraining and posttraining outcomes during eclampsia simulations. In their study, training was associated with an increase in the appropriate and timely completion of desired tasks (e.g., magnesium sulfate administration). Of note, although all participants underwent simulation training, they also were randomized to different training sites (i.e., simulation center vs. hospital unit) and to inclusion of teamwork training in their educational process. Neither site nor supplementary teamwork education was associated with additional improvement.
Other Obstetric Events
Other obstetric events that have been examined in the context of simulation include breech delivery, postpartum hemorrhage, and cord prolapse. In one study focused on breech delivery, after undergoing a simulation of an imminent term vaginal breech, residents took part in a training session with the simulator on the proper techniques for vaginal breech delivery. After this training, residents repeated the standardized simulation. Simulations both before and after training were videotaped and judged by an investigator who was blinded to training status. Scores for skill and safety were significantly greater for residents who participated in the simulated breech delivery after the training. With regard to postpartum hemorrhage, Toledo and colleagues used simulated examples of blood loss volumes to improve care providers’ accuracy with regard to blood loss estimation.
Siassakos and colleagues assessed whether clinical outcomes associated with cord prolapse improved after introduction of an obstetric emergency training program that included cord prolapse drills. They found that after these drills, a significant reduction was reported in diagnosis-to-delivery interval (25 to 14.5 minutes, P < .001), although no significant difference was found in low Apgar scores or rate of admission to the neonatal intensive care unit (NICU).
Draycott and coinvestigators also examined clinical outcomes after introduction of a training session that included fetal heart tracing education and drills in shoulder dystocia, postpartum hemorrhage, eclampsia, twin deliveries, breech deliveries, adult resuscitation (including cardiopulmonary resuscitation), and neonatal resuscitation. After this training, the frequency of hypoxic-ischemic encephalopathy (HIE) decreased by approximately half at their institution (27.3 to 13.6 per 10,000 births, P = .03). This decrease did not seem to be related to other preexisting trends in the frequency of HIE or to changes in the population during the period of study.
Enhancement of Communication
Rather than focus on simulation of specific obstetric events, some investigators have emphasized improvement of communication processes and teamwork in general. One approach to this is through the establishment of a specific team trained to function and be engaged specifically at the time of obstetric emergencies. One institution, for example, created a “Condition O” team, according to the precepts outlined for rapid-response systems, including (1) case detection that triggers a medical crisis team response, (2) a medical crisis team response available at all times, (3) an evaluation and process improvement system, and (4) an administrative structure to support the system. After implementation of this system, the investigators reported several quality improvement interventions that were able to be introduced due to case reviews associated with “Condition O” being invoked, although improvements in particular outcomes were not yet demonstrated.
A different approach, in contrast to the one above, is the training for and implementation of standardized communication processes throughout a unit or institution . One often-cited approach to this type of intervention is crew resource management (CRM). This training began—and has continued—in the airline industry, where it was noted that accidents were primarily related to lack of coordination and poor teamwork. CRM seeks to engender effective communication through standardized language, situational awareness, briefing and debriefing, and a leveling of hierarchy that allows all team members to voice concerns over safety. Some evidence from both observational longitudinal studies and randomized trials suggests that communication training can improve teamwork and communication among providers.
In one longitudinal observational study, the introduction of a CRM course at a single institution was noted to be associated with a reduction in high-severity “lawsuits, claims, and observations” (i.e., when the insurance carrier reserves money) by 62%. However, in the one randomized trial that was performed—and in which this CRM course, the MedTeams Labor and Delivery Team Coordination Course, was evaluated—investigators reported that institutions randomized to the intervention were no more likely to reduce the primary measure of obstetric adverse outcomes (mean adverse outcome index) than institutions that received no intervention at all.
Some have asserted that the complexity of obstetric care is such that improvements in safety will come not through the use of one particular intervention but with a multifaceted approach that ultimately results in a fundamental culture change within an institution. Correspondence with principles of high-reliability organizations have helped to guide some of these efforts. Longitudinal observational studies have suggested that the introduction of such programs is associated with outcome improvements. Examples of such programs have been reported by several organizations:
Catholic Healthcare Partners (CHP) instituted a program at 16 sites that included interdisciplinary education, routine medical record reviews to monitor ongoing adherence to appropriate practice, and standardization of forms to encourage adherence with best practices. After introduction of this program, birth trauma rates decreased from 5.0 to 0.17 per 1000 births, and the number of adverse obstetric occurrences (specified birth-related event or injury that might lead to a claim) decreased by 65%, from 7.2 to 2.5 per 1000 births.
Seton Family Hospitals began a program at four sites that included standardization of forms and care processes, introduction of routine communication processes such as situation, background, assessment, recommendation (SBAR); active surveillance of adverse outcomes with temporally proximate feedback; and interdisciplinary obstetric crisis simulation with high-fidelity mannequins. Subsequent to the introduction of this program, birth injury rates declined from a 2-year average of 0.3% to 0.08%, and the average length of stay for infants admitted to the NICU for birth injury declined by 80%.
Yale–New Haven Hospital obstetric department began a program of outside expert review of adverse events, anonymous event reporting, protocol standardization, creation of a patient safety nurse position and patient safety committee, and training in team skills and fetal heart monitoring interpretation. Introduction of this program was associated with a significant reduction ( P = .01) in their outcome of choice, the mean quarterly adverse outcome index. They have reported that medical liability claims and payments also have been reduced subsequent to the introduction of this program.
Hospital Corporation of America (HCA) instituted a program that included standardized processes, active peer review and feedback, and the empowerment of any member of the health care team to halt care considered possibly dangerous. Subsequent to this program’s initiation, a significant reduction was seen ( P < .001) in the number of professional liability claims per 10,000 deliveries.
It seems clear that at least some of the adverse outcomes that occur in obstetrics are preventable and that the root causes of these preventable events, although multiple, often involve communication and teamwork. Theory supports various approaches to error reduction, many of which have been used in obstetrics. Whereas some but not all studies have documented improvements in outcomes related to patient safety initiatives, these studies have been predominantly observational and have used different outcomes—such as provider satisfaction and perception, demonstration of skill acquisition in simulated scenarios, actual care processes, and actual patient outcomes—to assess improvement. Further work is necessary to understand the best combination of practices and practice methodologies (e.g., simulation on site versus simulation in a laboratory) that most effectively improve care.