The purpose of this study was to assess the effectiveness of instituting a comprehensive protocol for the treatment of maternal hemorrhage within a large health care system. A comprehensive maternal hemorrhage protocol was initiated within a health care system with 29 different delivery units and with >60,000 annual births. Compliance with key elements of the protocol was assessed monthly by a dedicated perinatal safety nurse at each site and validated during site visits by system perinatal nurse specialist. Outcome variables were the total number of units of blood transfused and the number of puerperal hysterectomies. Three time points were assessed: (1) 2 months before implementation of the protocol, (2) a 2-month period that was measured at 5 months after implementation of the protocol, and (3) a 2-month period at 10 months after implementation. There were 32,059 deliveries during the 3 study periods. Relative to baseline, there was a significant reduction in blood product use per 1000 births (−25.9%; P < .01) and a nonsignificant reduction (−14.8%; P = .2) in the number of patients who required puerperal hysterectomy. Within a large health care system, the application of a standardized method to address maternal hemorrhage significantly reduced maternal morbidity, based on the need for maternal transfusion and peripartum hysterectomy. These data support implementation of standardized methods for postpartum care and treatment of maternal hemorrhage and support that this approach will reduce maternal morbidity.
Maternal hemorrhage remains a major source of maternal morbidity and death. In the United States, the rate of postpartum hemorrhage has increased steadily; in 2004, 3% of all births were complicated by postpartum hemorrhage. Increased reporting or changes in coding practices are likely not a major factor in this increased rate of postpartum hemorrhage, because the nationwide rate of transfusion during admission for labor and delivery nearly doubled during the 8-year period from 1997-2005. Similar trends have been noted in Canada, Australia, and Europe. The increased use of transfusion in the peripartum period has been attributed to many factors; however, the primary reason appears to be due to an increase in the rate of uterine atony and can be explained only partially by changes in obstetrics practice. The incidence of abnormal placentation also appears to be increasing; in New York State, it is the most common cause of obstetric hemorrhage. “Near miss” events, defined as blood loss of ≥1500 mL, occur in approximately 15% of patients who experience postpartum hemorrhage, which suggests that approximately 18,000 women per year in the United States have life-threatening hemorrhage during the course of childbirth. Unfortunately, when subjected to rigorous review, many of these cases demonstrate that there was significant opportunity for improvement that would likely have reduced maternal morbidity.
Obstetricians, anesthesiologists, and obstetrics nurses all have experience with treatment of obstetric hemorrhage. Although most patients respond to therapy, patients frequently are treated by a variety of approaches that frequently vary considerably within a single institution. Further, the frequency at which any 1 provider (nurse or physician) will encounter a major obstetric hemorrhage is low, which suggests that a standardized and coordinated intervention is critical for optimal maternal and neonatal outcome. The Joint Commission and the Society of Maternal Fetal Medicine recommended the adoption of protocols to address maternal death and morbidity that are associated with postpartum hemorrhage.
In 2011, we reported our results after the introduction of a comprehensive patient safety initiative from a single institution that was directed at the treatment of maternal hemorrhage. In that report, we noted a reduction in the severity of maternal hemorrhage, the total number of blood products transfused, and the number of patients who experienced disseminated intravascular coagulation. The purpose of this report was to determine whether the implementation of a similar standardized protocol within large health care system would produce similar results.
Materials and methods
The data that were collected for this report were from an approved ongoing clinical patient safety monitoring program and as part the hospital system’s continuous quality improvement programs.
Data were collected from 29 Dignity Health System hospitals with maternity units. The hospitals vary in size from a small rural center with <200 deliveries per year to large urban hospitals with >6000 births annually. There are approximately 60,000 births annually within the hospital system. In 2010, the hemorrhage protocol was distributed to all hospitals within the system with the recommendation that each facility have an onsite dedicated hemorrhage cart. Compliance with the protocol was not assessed until November 2011. In November 2011, all hospitals were required to have a dedicated perinatal safety nurse and prospectively to collect data that were related to any obstetric hemorrhage and that met the criteria outlined in the protocol ( Table 1 ). Records and processes were audited to assess compliance with hemorrhage protocol recommendations. To be considered compliant, all components that are outlined in Table 1 had to be met. If a single element was not met, the case was considered noncompliant. Compliance with the protocol was assessed monthly; each center’s results were distributed to all sites through monthly perinatal safety conference calls, and the results were posted to an internal website. Physician leaders at each facility were encouraged to participate in the monthly conference calls. To encourage compliance and administrative support, senior administrative leaders were required to present their data at corporate leadership meetings. In addition, 2 perinatal safety nurses made scheduled site visits to audit the accuracy of the data that were submitted. When compliance was less than expected at any facility, the perinatal safety nurse was encouraged to address any specific concerns with the physician staff. If issues or concerns were still present, the lead author made site visits to obstetrics department meetings to review the protocol, rationalization for implementation, and data that supported its use. Hospital administration also attended these meeting. Compliance after site visits uniformly improved. At each site, staff education of all key elements and low fidelity drills were encouraged.
| □Admission hemorrhage risk assessment completed |
| □Correct blood bank request requested, based on risk |
| □Blood and blood clots weighed per protocol |
| □Correct laboratory results obtained for stage 2 and 3 hemorrhage |
| □Were >2 uterotonics given without the medical doctor present |
| □Blood products administered according to protocol |
| Quantity blood products are administered and which ones a |
| Number of peripartum hysterectomies a |
a Additional data that were collected but not used to assess compliance with the protocol.
The protocol was designed to provide graded assessments of patient acuity with standardized interventions based on the acuity status of the patient. It is important to point out that this was not designed or implemented as a stand-alone transfusion policy. Based on our previous experience, it was assumed that early standardized intervention and aggressive early transfusion with component blood therapy would decrease the severity of hemorrhage and use of blood products. The protocol that was used was similar to our previous report with minor modifications ( Figures 1-4 ).
The protocol was initiated at the time of admission to labor and delivery. At that time, an initial risk assessment was completed that was related to the patient’s potential risk for obstetric hemorrhage ( Table 2 ). Patients were then categorized as low, medium, or high risk; based on this admission risk assessment, different levels of “status alerts” were given to the blood bank. The validity of this risk assessment approach has been validated recently by others. This process was done primarily to streamline the initial rapid access to blood products when needed. In the event that there was no immediately available cross matched blood when patients reached stage 3, uncrossed type O-negative or type-specific blood was released for use.
| Low risk (clot to hold) | Medium risk (type and screen) | High risk (type and cross) |
|---|---|---|
| Singleton pregnancy | History of postpartum hemorrhage | Previa/accreta |
| Unscarred uterus | Previous cesarean delivery or uterine surgery | Hematocrit level <30% plus other risk |
| No history of postpartum hemorrhage | ≥5 previous vaginal deliveries | Bleeding on admission |
| ≤4 previous vaginal deliveries | Multiple gestations | Platelets <100,000/μL |
| No history of bleeding disorder | Large uterine fibroid tumors | Bleeding on admission plus symptoms b |
| Chorioamnionitis | Known coagulation defect | |
| Magnesium sulfate use |
a Patients with 2 risk factors in any category were moved to the next higher category
b These patients were assumed to be stage 3 on admission. Symptoms included: heart rate, >110 beats/min; blood pressure, <85/45 mm Hg; oxygen saturation, ≤94%; shortness of breath, confusion, or agitation.
Although each patient’s status was assessed in both the intrapartum and postpartum time periods, protocol interventions were designed primarily to address postpartum hemorrhage. Skills training that was related to recognition of blood loss was used to improve accuracy. Because of recognized inaccuracies in blood loss estimates, even after training, we recommended that a quantitative estimate blood loss be used. This was accomplished by weighing all lap-sponges, bed linens if needed, and fluid in collection systems. Nonblood fluid in delivery collection systems, particularly before delivery of the placenta, was subtracted from the estimated blood loss. Although there is some risk that amniotic fluid is included in the blood loss estimate, this method of assessment has been shown to improve the accuracy of blood loss estimates. The patient’s bleeding status was assessed continuously and assigned a clinical “hemorrhage stage.” which were grouped into 4 categories (stages 0-3). Stage 0 was designated as a normal intrapartum and postpartum course. Stage 1 was defined as bleeding greater than expected for normal vaginal delivery (500 mL) or cesarean delivery (1000 mL). Stage 2 was defined as bleeding that did not respond to conservative treatment, as outlined in Stage 1. Stage 3 was defined as continued bleeding, with actual or expected blood loss to exceed 1500 mL. There were minor changes to the originally published protocol based on facility feedback and recommendations from physicians and nursing staff within the Dignity Health System. Details of each stage are outlined below and in Figures 1-4 .
Patients whose condition was activated to stage 1 after vaginal delivery were treated primarily by nursing staff. This stage was designed to encourage prompt assessment and treatment of uterine atony. If uterine atony was suspected, a single dose of an uterotonic, in addition to oxytocin used for the third stage of labor, could be given after discussion with the physician. If there were a need for a second dose of a supplemental uterotonic agent, the patient’s status was upgraded to stage 2. The details of stage 2 are outlined in Figure 3 , A (vaginal delivery) and B (cesarean delivery). The most significant components of stage 2 was the commitment of additional personnel to assist with patient care; both the patient’s obstetrician and anesthesiologist were required to report to the patient’s bedside. This requirement was believed to be critical and was designed to prevent continued use of therapies that may be either ineffective or would limit timely physician evaluation and treatment. In addition, most interventions that were included in stage 2 require physician assessment. Based on previous experience, aggressive intervention at this stage should limit the number of patients that would progress to “near miss” or life-threatening hemorrhage and/or disseminated intravascular coagulation. To facilitate treatment, an “obstetrics hemorrhage cart” was developed, and it was recommended that there be 1 cart assigned to both labor and delivery and any of the operating room areas where obstetrics patients may be treated. The hemorrhage cart was designed and organized to contain all of the routine and unique supplies necessary for patient treatment. The recommended contents of the cart have been previously decribed. At stage 2, the hemorrhage cart was brought to the patient’s room.
The patient’s status was elevated to stage 3 if the estimated blood loss exceeded or was expected to exceed 1500 mL ( Figure 4 ). The primary goals of stage 3 were to mobilize all necessary resources towards reducing further blood loss and to reduce the risk of the development of disseminated intravascular coagulation. Additional nursing staff were assigned, and considerations for additional physician support (obstetrics, anesthesia, general, or urologic surgeons and interventional radiology personnel) were also suggested, as deemed necessary. At this point, fixed ratios of blood products in a designated “obstetrics hemorrhage pack” were prepared for immediate release from the blood bank. Details of the “obstetrics hemorrhage pack” are outlined in Table 2 . In the event that either a severe placental abruption or an amniotic fluid embolism was suspected, preparation of 10 units of cryoprecipitate was also recommended. Blood (packed red blood cells [pRBCs]) and fresh frozen plasma (FFP) were given in fixed ratios of 3:2. This ratio was increased to 1:1 after the transfusion of the first 6 units of pRBCs and 4 units of FFP. The use of high ratios of plasma to red blood cells has been shown to improve survival in trauma-related hemorrhage and to improve outcomes in obstetrics patients whose condition required transfusion. Treatment was directed towards the goal of maintaining the hematocrit level at >24%, international normalized ratio at <1.5, platelet count at >50,000/μL, and fibrinogen level at >100,000 mg/dL. Additional treatment goals were suggested that included maintaining the patient’s pH at >7.2, base excess at <−5, temperature at >95°F (35°C), and a normal ionized calcium level, because all of these factors are known to influence coagulation.
Postpartum care was also modified for those patients whose condition was activated to stage 2 or stage 3; stage 3 patients were assigned to labor and delivery or transferred to the intensive care unit, depending on their acuity. The decision for intensive care unit admission primarily was based on concerns related to the presence of pulmonary and/or renal compromise. The minimal frequency and duration of vital signs and laboratory assessments were also outlined by the protocol.
For data analysis, the 3 time periods were compared. Time point 1 was a 2-month baseline collection period (November and December 2011). Then, 2 additional postimplementation time periods (April-June 2012 and September-October 2012) were assessed. Compliance with the protocol was assessed monthly based on the 5 components that are given in Table 1 . Comparison of peripartum hysterectomy rates were done with the use of 2011 rates relative to 2012 rates when the hysterectomy was associated with coding that was consistent with hemorrhage or transfusion codes (ie, 666.x, 285.1, 99.04-99.07) ( Table 1 ). Data were analyzed by t test and by comparison of the differences between independent proportions with the use of the online statistical program, Vassar Stats ( www.vassarstats.net ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
