50,000 Reasons for Transforming Labor & Delivery Unit Culture
image credit DANI BRUBAKER
It is eerily quiet at two thirty in the morning. As I exit the women’s hospital, automatic glass doors slide shut behind me and cut off the supply of cool air, but I barely notice the enveloping July heat. What strikes me is the quietness. I am also aware of a feeling I have not been able to shake for much of the night. It manifests as a pit in my stomach and a warm tightness around my chest; the trembling, halting sense of relief when a mother evades death during childbirth by a hair’s breadth, but certitude is momentarily out of reach because the entire sequence of events has not fully processed.
The feeling is familiar. It has resurfaced with each article published by Nina Martin, Renee Montagne and the team at ProPublica in “Lost Mothers,” the award-winning series that launched the maternal health crisis in the United States into the mainstream. It has followed each successive news report and social media post cataloguing some of the nine hundred pregnancy and childbirth-related deaths occurring each year at a rate higher than any other industrialized country. It is enmeshed in collective outrage evident in the growing multimedia compendium that simmers because two-thirds of all maternal deaths are preventable and black women are dying two and a half times more than white women.
Besides setting off a robust discourse, these needless deaths have seen outrage channeled into triumphant action. March for Moms is leading a highly visible advocacy campaign targeting the spectrum of maternal health concerns. The Preventing Maternal Deaths Act, signed into law a few days before Christmas, will fund maternal mortality review committees in every state and tribal nation. These steps are an important response to the crisis at a macro level. But, it is not immediately obvious how such initiatives translate at the point of interface between patient and provider in hospitals where each year fifty thousand women experience a severe maternal morbidity, such as a blood transfusion or hysterectomy, according to the CDC. Little in the national discourse on maternal mortality in the United States is focused on addressing latent characteristics within health facilities in which preventable adverse events are allowed to progress to death.
In California, where Kira Johnson died following a routinely scheduled repeat cesarean section in June 2016, a maternal review board - the California Maternal Quality Care Collaborative (CMQCC) - has been in place since 2006. More than two hundred hospitals in the state are voluntary members of CMQCC and ninety percent of births occur in member facilities. The CMQCC is credited with the sustained decrease in maternal deaths in that state even as the rest of the country has seen an upward trend through its two core functions. First, collecting and analyzing data to generate facility level insights that help member organizations improve quality of maternal health services. Second, developing evidence-based toolkits to guide providers in clinical management of obstetric complications. The success of the CMQCC belies the fact that retrospective review of cases negates the possibility of impacting any one patient’s outcome in real time and that complex elements of an organization’s culture influence adherence to guidelines. These key disadvantages are a critical lens through which the unfortunate circumstances surrounding Ms. Johnson’s cascade to death can be viewed.
Ms. Johnson developed a surgical site hematoma in the hours following surgery at Cedars-Sinai Medical Center, a member of the CMQCC. Records obtained from the Medical Board of California and court documents match the account given by her husband, Charles Johnson IV. He recalled Ms. Johnson being assessed by multiple providers on multiple occasions. Despite abnormal findings compounding an initial diagnosis of hematuria (blood in urine), significant delays in mounting an effective response were encountered along the trajectory of care. More than twelve hours after frank blood first turned urine in the bag attached to her foley catheter pink, she suffered a succession of cardiac arrests while on the operating table as futile attempts to save her life were made. The vacillating response followed by last minute heroic, but unsuccessful, intervention implies latent organizational characteristics within the labor and delivery unit at Cedars-Sinai interacted in exactly the right magnitude and sequence to allow Ms. Johnson to succumb to surgical complications.
Linda Aiken, the eminent nurse researcher at the University of Pennsylvania, and her colleagues have described and studied the phenomenon known as failure-to-rescue, which they define as the “failure to bring patients with complications back from the brink of death.” Failure-to-rescue is best understood, conceptually, through its theoretical underpinnings. More than three decades of scientific research have established that internal organizational systems - the structures, processes and values that comprise nurses’ work environment, but often unseen by patients - are important determinants of quality of care and patient outcomes. Whether a unit schedules a sufficient number of nurses to effectively manage the existing workload, for example, or has processes in place for the same nurses to address concerns with hospital administrators affects their ability to prioritize patient surveillance. A subset of hospitals in the United States has mastered optimal organizational systems that consistently ensure quality work environments for nurses and concomitant superior patient outcomes. Eileen Lake, another nurse scientist at the University of Pennsylvania, identified and validated five domains that characterize these exemplary health facilities: 1) nursing participation in hospital affairs; 2) nursing foundations for quality of care; 3) leadership support of nurses; 4) adequacy of staffing and resources; and 5) collegial nurse-physician relationships. In seminal research, Aiken et al revealed significant relationships between Lake’s nurses’ work environments measures and patient outcomes. Assigning a nurse one additional patient increases the likelihood of failure-to-rescue in surgical patients by seven percent. Conversely, a ten percent increase in the proportion of baccalaureate prepared nurses decreases the likelihood of patients cascading to death by five percent. A one standard deviation increase in the composite work environment score is associated with an eight percent decrease in the odds of mortality. So important is the quality of work environments that in facilities where nurses report subpar work environments, improvements in nurse staffing confer zero benefits.
On that July night, I had not heard about Ms. Johnson’s case and I was not thinking about nurses’ work environments in the theoretical sense. I was scheduled to leave the hospital hours earlier when my shift officially ended and there were more people on the sidewalk plus a steady stream of motorists to navigate around. But, we had a near-miss. A healthy, young woman walked into the building now behind me to have a baby and nearly died following what was supposed to be a routine repeat cesarean section. Instead of sending my worn scrubs down the laundry chute when the clock told me it was almost midnight, I was pushing my barely conscious patient back into the operating room for an emergency exploratory laparotomy.
“Luisa*!” I had called out to a half-full waiting room at about five-thirty that afternoon.
She was the last scheduled cesarean section for the day. On standing up to respond, she revealed exquisite floral details on a strapless navy floor-length patio dress. From her heavy sigh and shifting gaze, I knew instinctively that she was nervous. So when I introduced myself, I reassuringly wrapped my arm around her bare shoulders and said, with a smile, that everything was going to be alright. I could tell that she believed me because almost immediately our conversation turned into a friendly and animated banter as we walked to the admissions suite where I was going to prepare her for surgery. We spoke about her first son while I strapped the fetal monitor on to her gravid abdomen for one last check of baby’s heartbeat before delivery. Her husband, Diego, clasped her right hand in both of his when she half-smiled, half-grimaced as I eased an eighteen gauge angiocath into a plump vein on her left forearm. With a bolus of crystalloid fluid rapidly infusing through the large bore intravenous access, I entered pre-operative data in the electronic medical record. Luisa was unremarkable in every way, which in our world is actually a good thing. The fact that this was going to be her second cesarean section did not raise significant alarm. In the operating room, when the baby was born, we all laughed as we cheered, “it’s a boy!,” because the entire team had known beforehand that it was going to be another boy.
In our department, one nurse cares for scheduled obstetric surgery patients from admission through transfer to the mother-and-baby unit. Therefore, Luisa and I were still a pair in the post-anesthesia care unit when her condition began to change. The first time she mentioned feeling slightly light-headed, she was in a semi-reclined position, with her newborn son wearing only a diaper lying prone directly on the skin on her chest. Sitting on a wheeled stool positioned at the foot of the stretcher, I was closely watching her vital signs, which registered on a monitor attached to the wall above her head. Simultaneously, the readings populated automatically on her recovery record where I promptly validated them on a computer terminal located right next to my stool. So far, they were all normal and did not deviate more than twenty percent from figures obtained on admission. I stood up, walked to the head of the stretcher, slowly released the lever at the top and lowered it by gently pushing down to allow an increase in blood returning from her legs to the central circulation. Then I lifted the pile of blankets keeping mother and newborn warm to palpate the fundus of her uterus, which was firm at the level of the umbilicus. The maternity pads between her legs were stained, but not saturated, with lochia rubra. Assured by objective and subjective signs that Luisa’s recovery was proceeding normally, I smiled and turned my attention to a scene I will never grow tired of: a content and peacefully sleeping newborn infant nestled on his mother’s bosom.
Almost an hour into Luisa’s post-anesthesia recovery, I recorded her first low blood pressure. Eighty over forty. Heart rate normal. Respirations normal. Recalling her complaint of light-headedness less than a half hour earlier, I immediately took another blood pressure, but this time it was normal. As a precaution, I relayed my observations to the anesthesiologist. We both agreed that lack of overt signs of excessive bleeding meant it was more likely that she was “behind” on fluid. “Behind” is medical jargon for hypovolemia, which the anesthesiologist treated by ordering a 500 milliliter intravenous bolus of lactated ringers solution. Luisa was already receiving in her main line an intravenous infusion containing a synthetic form of the hormone Oxytocin to prevent excessive bleeding. Oxytocin possesses antidiuretic properties with potential to exacerbate hypovolemia particularly when administered rapidly. Therefore, I piggybacked fresh tubing connected to a one liter bag of lactated ringers solution to the main intravenous line and programmed the medication infusion pump to deliver half.
Before the initial bolus of lactated ringers was complete, Luisa registered another low blood pressure. Neon pink digits on the monitor flashed eighty over forty. Like the last abnormal blood pressure reading, this one was also followed by normal values. When I called the anesthesiologist a second time, he asked me to run the entire liter of lactated ringers and I immediately adjusted the settings on the infusion pump to comply with the new order. I also informed the surgeon, who happened to be in-house tying loose clinical and administrative ends. Meanwhile, Luisa’s light-headedness, which seemed to subside when the head of the stretcher was positioned at a lower angle, began to bother her again. She held the newborn son she was breastfeeding in her right arm, placed the back of the left hand over her forehead and shut her eyes. I was pained to interrupt what was until now a successful first latch, but we both agreed that she could use the rest. I swaddled the baby in several standard issue blue and pink striped blankets then placed him in the plastic bassinet, which was wheeled to the well-baby nursery by another nurse so I could remain vigilant at Luisa’s bedside. Customarily, new fathers accompany their newborn infants to the nursery to observe the baby’s admission process, but Diego declined. Perhaps sensing, but not articulating an unsettling feeling that his wife was not well, he preferred to remain at her side.
The next set of events progressed rapidly. The surgeon and anesthesiologist arrived to assess Luisa almost simultaneously. Standing to the patient’s left, the anesthesiologist and I considered in our differential diagnosis the likelihood of internal bleeding given vital sign changes, complaint of dizziness and the lack of overt hemorrhage. However, the surgeon standing on the opposite side of the stretcher demurred. “I’m pretty sure she was dry when I closed,” he said, emphasizing a high degree of certainty that he had achieved hemostasis before tying off the last suture to close the wound. Unbeknownst to the surgeon as he was declaring hemostasis, complex intrinsic compensatory mechanisms that activated automatically to counter hypovolemia were already reaching their limit and struggling to sustain Luisa’s blood pressure. The one liter bolus of lactated ringers solution did not successfully expand intravascular blood volume. Luisa’s blood pressure dipped even further to seventy over thirty. Unlike previous instances when repeat measurements registered normal, readings now lingered at lower levels and were accompanied by acceleration in heart rate beyond one hundred and twenty beats per minute in response to the decreased cardiac output. I inserted a second large bore intravenous access in her right forearm and started a wide open infusion of another bag of lactated ringers solution while another nurse confirmed that the blood bank was urgently readying packed red blood cells for possible transfusion.
By now the small group of nurses and obstetric technicians was beginning to assemble in the previously deserted recovery room. Fluorescent lights that had been turned off when operations wound down earlier were promptly switched back on as a hive of activity resumed at the first inkling of trouble. We prepared for the next logical steps. Thinking ahead, the charge nurse ordered an operating room to be prepared because signs indicated that surgical intervention was all but certain. “Just in case, make sure you open and count a hysterectomy tray,” she told the obstetric technician and nurse assigned to prepare sterile instruments in the event of an exploratory laparotomy. Luisa still exhibited classic signs of hemorrhage – low blood pressure, elevated heart rate and dizziness – but, showed no overt signs of excessive bleeding. A telltale pool of blood did not form between her legs and when the surgeon performed a bimanual exam, his gloved hand emerged bloodstained, but without any blood clots or large collection. “The lower uterine segment is well contracted,” he said. Before we could weigh the risks against the benefits of conservative management, which the surgeon initially recommended, a senior obstetrician who was on call that night arrived. After a brief presentation of the case, he took one look at Luisa who was now as pale as the white sheets she lay on, one look at the vital signs on the monitor above her head and said ‘we have to bring her back.”
Later, when the events began to digest, I wondered how word of Luisa’s condition had reached the senior obstetrician. His directive to return to the operating room, fortified with authority bestowed by his seniority, made all the difference for Luisa. It meant we did not have to spend precious moments making a case for surgical intervention. Then, if the surgeon disagreed, despite the validity of our case, taking additional time to invoke the chain of command by calling a physician administrator at home so they could act as an arbitrator. In reality, clinical management of complications like Luisa’s rarely follows a linear path. And conflicts routinely arise. How obstetric teams navigate through these differences in opinion can mean the difference between life and death. Poor teamwork and communication are implicated in three-fourths of all perinatal deaths. Thankfully, in Luisa’s case, unlike Ms. Johnson’s, redundancies in the system worked. Among the phone calls the charge nurse made while coordinating resources was to the senior obstetrician who then executed an effective response. In other instances, such phone calls have met a sharp retort. Frontline providers know when a patient is being mismanaged to the point they are exposed to harm and who the problem providers are. More than ninety percent of three thousand physicians, midwives and nurses responding to a survey endorsed that they had observed colleagues taking dangerous shortcuts, missing competencies, being disrespectful and exhibiting performance problems that undermined patient safety. But, fewer than thirteen percent of the same respondents stated they shared their concerns. In another study of more than two thousand physicians, nurses and midwives, more than fifty percent of maternal fetal medicine specialists and more than one third of generalist obstetricians and midwives “reported difficulty getting clinical concerns heard.”
The Harvard obstetrician-gynecologist, Neel Shah, correctly points to shifting patterns in maternal mortality in the United States, which indicate that two-thirds of deaths take place more than a week after delivery due to a chronic etiology. The data suggests that greater resources should be dedicated to community interventions where majority of deaths occur. Given the prevalence of severe maternal morbidity stands at fifty thousand cases per annum, each with the potential to proceed unencumbered to death when conditions are suitable, strategic plans should include forward-facing transformations of work environments in which maternal care is delivered as crucial antecedents of an organizational culture that prioritizes safety. Obstetric units are dynamic in nature. They are characterized by frequent fluctuations in patient census and acuity, which occur in a safety critical environment where error could easily result in death or serious injury. Yet, the degree to which best safety practices are ingrained in the culture of labor and delivery units across the country is unknown.
After the immediate crisis ended and Luisa was deemed to be in stable condition, I handed her care over to the night staff. The surgeon remained in the hospital the rest of the night, in close proximity just in case her condition deteriorated. Across the street, my car was in the garage where I parked it what seemed like a few days ago. In reality, sixteen hours had passed and now it was the only car on the entire rooftop. The brilliant white glare from the lamp directly above the hood was in stark relief to darkness so dark and quietness so quiet it was as if I had entered another realm. I was certain in the quietness of that moment it might be impossible for an uninvolved person to imagine the messy battle waged by the young woman who confronted death just a few steps away and emerged victorious. Not on such a quiet night. There was not a single sign that something was seriously awry. Even after I turned the key in the ignition, shifted into first gear and eased my left foot off the clutch, the hum of the engine did not drown the quietness.
*Luisa and Diego are not the couple’s real names.