|
Paramedic Intubation Errors: Isolated Events Or Symptoms Of Larger Problems?
Paramedics provide life-saving emergency medical care to patients in the out-of-hospital setting, but only selected emergency interventions have proved to be safe or effective. Endotracheal intubation (the insertion of an emergency breathing tube into the trachea) is an important and high-profile procedure performed by paramedics. In our study population, we found that errors occurred in 22 percent of intubation attempts, with a frequency of up to 40 percent in selected ambulance systems. These findings indicate frequent errors associated with this life-saving technique. These events might be emblematic of larger issues in the structure and delivery of out-of-hospital emergency care.
PARAMEDICS PROVIDE OUT-OF-HOSPITAL emergency medical services (EMS) to patients suffering life-threatening conditions such as cardiac arrest, respiratory failure, and major trauma. Paramedics often must perform endotracheal intubation (ETI), the insertion of a breathing tube into the trachea (windpipe), on these critically ill patients. Intubation is regarded as one of the most important EMS procedures and has been the standard of care in the United States for more than twenty years.1 ETI is a complex procedure. Considerable skill is required to maneuver the breathing tube into the correct position. An improperly placed tube (for example, in the esophagus) can deprive the patient of oxygen and rapidly lead to death. Physician trainees in anesthesiology, critical care medicine, and emergency medicine spend many hours in controlled settings learning and practicing this difficult procedure. In contrast, paramedics undergo only a fraction of the training in ETI that physicians receive. Furthermore, paramedics perform ETI in uncontrolled, stressful, and chaotic out-of-hospital settings, such as the floor of a cramped bathroom or in a mangled car after an automobile crash. Prehospital patients are often critically ill and injured, and it is often impossible to identify basic airway structures. Finally, although ETI is viewed as life-saving and common in EMS systems, most paramedics do not perform it frequently.2 Given these factors, there is great potential for adverse outcomes and errors during ETI—for example, inadvertent oxygen desaturation, heart rhythm disturbances, breathing tube dislodgment, and unrecognized tube misplacement.3 For this paper we sought to identify the prevalence of ETI errors and their associations with patient and EMS system characteristics. The quality of out-of-hospital emergency care and the occurrence of medical errors are both subjects of intense scrutiny and debate in the twenty-first-century health care system.
For this study (which was approved by the University of Pittsburgh Institutional Review Board), we analyzed data from the Prehospital Airway Collaborative Evaluation, Phase II (PACE II). As part of this prospective observational study, rescuers from forty-two advanced-life-support EMS services in Pennsylvania provided information on patient encounters involving ETI. An EMS "service" is a public or private agency that operates a group of ambulances providing emergency care to a designated region or municipality. Of the 1,443 EMS services in Pennsylvania, approximately 400 provide advanced-life-support care. Rescuers completed structured, closed-response data collection forms describing the characteristics of the patient, the clinical nature of the encounter, the course of airway management, adverse events and complications, and initial outcomes.4 Forms were returned to the coordinating center (University of Pittsburgh), where the data were entered into a database and analyzed. In Pennsylvania, paramedics, prehospital nurses, and physicians (both attending and resident level) perform out-of-hospital intubation. Paramedics performed more than 95 percent of intubations in this series. We calculated data return rates by comparing the number of returned data forms with the number of intubations recorded by computer charting programs at each service. We collected data from 1 January 2003 through 30 June 2004. This methodology was based on the data collection strategy developed for Phase I of this evaluation.5 We defined an ETI error as any of the following: breathing tube misplacement (placement in esophagus) or dislodgment from trachea, multiple attempts to perform ETI (four or more efforts to expose the vocal cords), or failed ETI. The investigative team chose these events by consensus because they (1) fulfilled the definition of medical error (failure to carry out an intended task or failure to carry out a task as intended), (2) could be identified using self-reports, and (3) had plausible links to adverse patient outcomes.6 We defined the ETI error rate as the proportion of patients receiving ETI efforts who were exposed to an error. We examined the associations between ETI errors and patients’ characteristics and clinical factors, including age, sex, race, clinical status (cardiac arrest versus nonarrest), nature of call (trauma versus medical), and intubation method (conventional, sedation-assisted, and neuromuscular blockade–assisted). We also examined relationships between errors and EMS characteristics, including the agency’s number of annual intubations, number of annual patient contacts, system configuration (ground versus air medical), personnel mix (all career versus mixed career/volunteer), number of advanced-life-support rescuers, and the service’s mean response and transport times. Service characteristics were determined from profiles provided by the Pennsylvania Department of Health. We examined associations between ETI errors and the population characteristics of each service. Population settings were defined as rural or urban, based on Office of Rural Health Policy definitions for the minor civil division of the ambulance service’s primary station.7 We determined population median income and percentage below poverty from 2000 U.S. census data.8 Using exact binomial confidence intervals, we calculated aggregate and individual service error rates. We evaluated the relationships between ETI errors and patient characteristics and clinical factors using univariate exact odds ratios. We evaluated the association between ETI errors and system characteristics using multivariable regression (generalized estimating equations), adjusting for patient age, clinical status (cardiac arrest versus nonarrest), systolic blood pressure, method of intubation (with or without adjunct medications), number of annual intubations, and number of annual patient contacts, and we accounted for clustering within each service. All analyses were completed using Stata version 8.2.
Exhibit 1  summarizes the characteristics of participating EMS services. Data were available for 1,953 patients receiving ETI, reflecting a data return rate of 68 percent. The mean patient age was sixty-two years, and 1,192 (61 percent) were male. Of these patients, 1,278 (65.4 percent) were in cardiac arrest, and 285 (14.6 percent) suffered traumatic (as compared with medical) conditions.
Any ETI error occurred in 444 patients (Exhibit 2 ), including tube misplacement or dislodgement in 61 patients, multiple ETI attempts in 62, and failed ETI in 359. Error rates within EMS services ranged from zero to 40 percent (Exhibit 3).
Errors were most pronounced for patients younger than six years, nonarrest patients (patients with a pulse), and trauma patients and were less likely for patients receiving neuromuscular blocking agents (Exhibit 4 ).
Error rates were lower for EMS services that performed more intubations annually (more than fifty per year) but higher for services with greater numbers of patient contacts (more than 5,000 per year). ETI errors were not associated with system configuration (ground versus air medical), personnel patterns (all career versus mixed career/volunteer), the number of paramedic-level rescuers, mean response or transport times, or population characteristics (urban/rural setting, median household income, or percentage of population below poverty).
Despite calls and initiatives to reduce medical errors nationally, there have been few efforts to understand errors in out-of-hospital care.9 This part of health care is likely one of the most error-prone areas of medicine.10 Preceding data on intubation errors are limited in scope. For example, Steven Katz and Jay Falk examined tube misplacements in intubated paramedic cases arriving at a single urban trauma center, and James Dunford and colleagues examined physiological disturbances occurring during paramedic intubation assisted by neuromuscular blockade.11 In this multicenter study, we used a broad definition of ETI error to describe the extent to which patients are at risk for one or many errors. We found that almost one of every four patients receiving ETI was exposed to an error. Error rates were variable, with some services reporting error rates as high as 40 percent. We found that ETI errors were not associated with system characteristics, which suggests that these events are not limited to isolated practice or population settings. Therefore, any interventions to reduce errors must be applicable to all EMS services. Clinical implications. These errors have important clinical implications. Breathing tube misplacement or dislodgment can result in inadvertent oxygen delivery to the stomach instead of the lungs.12 Multiple ETI attempts can cause direct injury to airway structures and subject the patient to episodes of oxygen deprivation.13 Failed ETI not only is a process error (failure to successfully carry out an intended procedure) but can result in suboptimal oxygen delivery if alternative airway measures cannot be instituted. We used a composite endpoint in this analysis because ETI errors often overlap; for example, it may be difficult to distinguish tube dislodgement from tube misplacement. Although we did not connect ETI errors with patient outcomes, these events might be partly responsible for the increased mortality observed in selected patients receiving the intervention.14 Study limitations. We limited this analysis to adverse events that met accepted definitions of medical error, that could be identified using self-reports, and that had plausible effects upon patient outcome. We did not design a system to capture other potentially important adverse events such as oxygen desaturation, bradycardia, or airway injury. Furthermore, certain events might not be detectable until examined by physicians at the receiving hospital. Also, our series describes only a portion of the 11,500 intubations performed annually in Pennsylvania.15 Thus, the actual ETI error rate and total number of errors are likely higher. Our analysis is limited by the use of self-reported data. Although the independent identification of clinical events is desirable, this is logistically impossible on a large scale in the uncontrolled out-of-hospital environment. Anonymous self-reporting is a pragmatic alternative. Because of this and the moderate data return rate, we might not have captured all possible adverse events. However, we observed alarming error rates even with these limitations. We surmise that respondent biases would deflate (not inflate) these error rates, and thus our estimate likely represents the "best-case" scenario. Other associations. Two notable secondary associations are that ETI errors were more prevalent in "awake" nonarrest patients but less likely in those receiving neuromuscular blocking agents. Semiconscious and head-injured patients might possess intact protective reflexes that impede the rescuer’s access to the airway. Physicians in the hospital often administer neuromuscular blocking agents to facilitate intubation of these patients. In Pennsylvania the out-of-hospital use of these medications is restricted to air medical rescuers, who receive more advanced airway training than ground-based paramedics do. Other research has demonstrated increased adverse events (including increased mortality) when ground-based paramedics use these drugs for intubation.16 It might be difficult to reduce ETI errors because of the multifaceted nature of the problem. For example, paramedics face multiple barriers to acquiring and maintaining airway management skills. Although anesthesiology and emergency medicine physician trainees must perform 35–200 ETI procedures, paramedic students must perform only five to graduate.17 Once in clinical practice, many paramedics have only limited opportunity to perform the procedure. In a prior study, we found that most paramedics in Pennsylvania perform only one or two per year, and some, none at all.18 Increasing medicolegal concerns have severely curtailed paramedics’ access to operating rooms for initial and continuing ETI training.19 Whether or not to intubate. In the spirit of "first, do no harm," we might consider not intubating at all. This strategy might be appropriate for patients who still have some intact breathing capacity. For those who are apneic (not breathing at all), ETI could be replaced with simpler airway management devices that are easier to insert than a standard endotracheal tube, have lower potential for error, and are preferentially used by ambulance systems in many countries.20 However, this strategy challenges the long-standing but unsubstantiated belief in the United States that ETI is the optimum method for airway management. Policy implications. The patient safety literature highlights the fact that shortcomings in the system—not the failure of individuals—result in medical errors.21 As with hospital errors, the procedural errors observed in our study might be symptoms of larger problems with out-of-hospital emergency care nationally. Historically, federal support fueled the development of the U.S. EMS infrastructure during the 1970s.22 The overarching goal of this widespread effort was to improve survival from cardiac arrest and major trauma; however, the government has done little to evaluate the results. Few EMS studies have enjoyed Public Health Service support.23 Only recently has the National Institutes of Health (NIH) assembled a multicenter Resuscitation Outcomes Consortium to study the out-of-hospital treatment of cardiac arrest and trauma.24 Current federal policies also fail to recognize the priority issues and realities of out-of-hospital care. Since September 11, 2001, millions of dollars have been dedicated to preparing EMS services for bioterrorism, but these events are exceedingly rare.25 EMS as a whole might enjoy greater benefit if these dollars were spent on common conditions encountered every day, such heart attacks and brain injuries. Medicare reimbursement policies undercut the fiscal realities of providing emergency ambulance care.26 In Pennsylvania many services struggle to pay paramedics just over minimum wage, and many have lost professional staff to better-paying jobs at gas stations, department stores, and fast-food restaurants.27 Improvement of EMS care is the responsibility of state, regional, and local bureaus as well as the ambulance agencies themselves. However, quality of care might have a stronger chance of improvement if federal leadership, advocacy, and direction were strengthened. Oversight is now provided by a program office in the National Highway Traffic Safety Administration.28 Movement of EMS leadership to another branch of government (such as the Department of Homeland Security) has been proposed and debated.29 EMS has long been overshadowed by other health care constituencies in Washington, D.C., and enhanced advocacy and leadership could help address EMS needs nationally. The involvement of government as a partner in the science and quality of out-of-hospital care is not without precedent. One of the largest, most successful out-of-hospital research efforts, the Ontario Prehospital Advanced Life Support Study, originated from a Canadian government mandate to justify the cost of adding paramedic-level care to communities in Ontario.30 EMS care in the United States has been essentially unchanged since its conception more than twenty-five years ago. The safety of EMS as a whole could depend upon a reexamination—and potentially a redesign—of the entire system.
Henry Wang (wanghe{at}upmc.edu) is an assistant professor in the Department of Emergency Medicine, University of Pittsburgh School of Medicine, in Pittsburgh, Pennsylvania. Judith Lave is professor and chair of the Department of Health Policy and Management, University of Pittsburgh. Carl Sirio is an associate professor in the Department of Critical Care Medicine, University of Pittsburgh School of Medicine. Donald Yealy is professor and vice chair of the Department of Emergency Medicine, University of Pittsburgh School of Medicine. This study was supported by the Laerdal Foundation for Acute Medicine, Wappingers Falls, New York, and the Pittsburgh Emergency Medicine Foundation, Pittsburgh, Pennsylvania. Respondent incentives were provided by the Pepsi Bottling Group (McKees Rocks, PA), Sheetz Inc. (Altoona, PA), and Krispy Kreme Doughnut Corporation (Winston-Salem, NC). Henry Wang is supported by Clinical Scientist Development Award no. K08 HS013628 from the Agency for Healthcare Research and Quality (AHRQ). The authors acknowledge Douglas Kupas for his invaluable assistance in facilitating the study, and Michelle Morgan, Benjamin Abo, Guy Guimond, and Laura Rosas for their diligent assistance in coordinating the study. They also acknowledge the support of the Pennsylvania Department of Health EMS Office, and the participating sites of the PACE Research Consortium (listed at http://www.emsresearch.org/content.php?content.36).
| ||||||||||||||||||||||||||||||||||||||||||
Top
Study Data And Methods
Study Results
