| کلیدواژههای انگلیسی مقاله |
Simulation learning, Cesarean section, Obstetric, IntroductionTrauma complicates 6-7% of all pregnancies and is the leading cause of non-obstetric maternal morbidity and mortality. An analysis conducted by WHO reported about 73% of all maternal mortalities worldwide were attributed to true obstetric etiologies with hemorrhage as the leading cause of death ( 1, ). Identification of hypovolemic shock during pregnancy poses challenges due to the unique physiological adaptions, including increased plasma volume, stroke volume, and cardiac output ( 2, ). Shock may not manifest until the total blood volume depletion reaches 30-35%, as compared to 15% in a non-pregnant patient ( 3, - 5, ). Clinicians are unable to rely on vital signs, physiological changes, or indications of end organ damage until patients are in a critical condition. The use of vital signs alone is impractical. Thus, accurate estimation of cumulative blood loss can forewarn impending hemodynamic compromise. Estimation methods are non-standardized, but a few commonly used techniques include visual, gravimetric, and colorimetric methods. The visual method involves estimation by the clinician through visual assessment of blood contaminated surgical equipment like surgical sponges and suction containers. Gravimetric methods involve subtracting blood-soaked surgical equipment from their dry weights to indirectly measure the blood loss. This value is then combined with the amount of liquid present in collection containers ( 6, ). Lastly, colorimetric estimation uses technology platforms to calculate the blood loss from photographs of soaked surgical materials. The algorithm filters out non-blood, colored components in the photographs through colorimetric analysis and determines hemoglobin mass present ( 6, , 7, ). Clinically, during simulation, quantitative methods like gravimetric and colorimetric have been shown to improve blood loss estimation ( 7, - 12, ). For instance, Lilley et al. showed gravimetric estimation of blood loss offered a significantly lower mean percent error (4.0&,plusmn 2.7%) in estimation compared to visual estimation (34.7&,plusmn 32.1%) ( 12, ). Gerdessen et al. performed a meta-analysis which compared modalities of blood loss estimation and found that colorimetric techniques offered the highest degree of accuracy in blood loss estimation. This is supported by another study which reported visual and gravimetric methods having a higher degree of bias when compared to colorimetric ones ( 7, ). Medical literature attempts to prove one modality superior to another however, these studies show conflicting results and state the evidence to be insufficient ( 13, ). In addition, each method possesses its own flaws for measurement and is vulnerable to inaccuracy ( 7, ). For instance, the addition of amniotic fluid in obstetric procedures as well as external blood loss in trauma situations can confound gravimetric analysis. A major barrier to implementation of colorimetric analysis is access to an artificial-intelligence technology.Historically, visual estimation is the most practical method of measurement, especially in obstetric emergencies, where quantitative measures cannot feasibly be applied due to time and physical constraints ( 13, - 15, ). However, the inaccuracy of visual blood loss assessment has been illustrated in several studies ( 8, , 16, - 18, ). Stafford et al. compared the accuracy of visually estimated blood loss (EBL) to calculate blood loss (CBL) for both vaginal and cesarean deliveries. The results found EBL to be significantly inaccurate when compared to CBL for both modes of delivery ( 18, ). These findings remain true in a recent simulation study that showed inaccuracies in visual EBL in a video series of emergent and non-emergent injuries ( 19, ). Most studies have shown that at lower volumes visual estimation has a tendency to overestimate blood loss ( 8, , 19, ). On the other end, at higher volumes, visual estimation tends to underestimate blood loss ( 4, , 20, ). Supporting this tendency of underestimation, Lertbunnaphong et al. found visual techniques underestimated blood loss compared to the gravimetric drape method, as the visual method missed 65.4% of post-partum hemorrhage diagnoses ( 20, ). Another supporting study reported that the participants overestimated smaller blood volumes between 50ml to 200ml and underestimated blood volumes greater than 400ml. This study also reported a greater percent error in estimation as blood volumes increased ( 19, ). One study disapproved the inaccuracy of the visual method and showed it to be equal to CBL when approximating the volumes &,lt 500mL ( 21, ). Additional factors such as professional experience and training show conflicting results, with some showing no improvement in accuracy ( 10, , 22, , 23, ). Regardless of these shortcomings, the decision to transfuse in obstetric emergencies has shown no significant difference between various modalities of blood loss estimation ( 9, , 24, , 25, ). Our simulation is novel as it utilizes a programmed learning didactic approach. Programmed Learning/Instruction is the process of arranging the material to be learned into a series of sequential steps to help the learner form mental associations between familial and unfamiliar concepts. As an educational technique, the learner is presented a logical sequence of materials, with multiple content repetitions. Immediate feedback is given, which serves as reinforcement of the content. A systematized method of blood loss estimation may increase awareness that a massive obstetric hemorrhage has occurred, thereby allowing for earlier intervention and improved communications during trauma situations&,nbsp and transition of care.&,nbsp Thus, the objective of this study was to design and implement a simulation exercise depicting obstetric scenarios of hemorrhage to train anesthesiologists, ancillary staff, and surgeons to estimate blood loss visually more accurately. We hypothesized that a programmed learning didactic simulation exercise would improve the provider&,rsquo s estimation of blood loss in a clinical setting. In order to test this hypothesis, we compared the percent error of blood loss estimation for cesarean deliveries after the simulation exercise to the period immediately prior. Methods Part A, The Programmed Learning Simulation We conducted a prospective study to assess the effect of implementing a PLS exercise on quantification of blood loss in an operation setting. This study was reviewed by the Institutional Review Board of our affiliated university, New York Medical College, Valhalla, New York, and received an exemption. The PLS was part of regularly scheduled continuing medical education activities, thus the sampling was a non-randomized convenience method. The programmed learning simulation (PLS) exercise was administered to 88 learners consisting of 24 third-year medical students, 19 obstetrics and gynecology physicians, 22 ancillary staff members (operation room nurses and technicians), 10 trauma surgeons, and 14 anesthesiologists at Richmond University Medical Center, Staten Island, New York. For the residents and attending physicians, participation was mandatory as part of their continuing medical education. Nurses, ancillary staff, and medical students were highly encouraged to participate and did so voluntarily. The participants were told that their answers may be used for quality improvement/research purposes however, there was no formal consent process. The PLS was conducted jointly by attending physician leaders in the ob/gyn and trauma surgery departments. The didactic activity took place over the course of three sessions spanning a ten-day time period. For each session, the stations were recreated with fresh artificial blood to avoid spoilage. Participants were expected to maintain confidentiality and requested not to share information regarding the simulations or answers to the simulation station estimates with their colleagues. The PLS exercise was designed to teach the providers how to visually quantify blood loss using a series of thirteen simulation stations. Each station depicted a simulated amount of blood loss for a clinical scenario. For this PLS exercise, artificial blood was prepared using non-validated mixture light corn syrup, water, and red food coloring with the first two ingredients in a 5,1 ratio and drops of coloring added until the desired hue was achieved. However, several more contemporary recipes are available based on the desired quantity and consistency ( 26, ). All operating room supplies were obtained from the labor and delivery floor of Richmond University Medical Center. The ninety-minute PLS exercise was executed in two parts, followed by a debriefing. For Part 1 (twenty-six minutes), the participants visited each of the thirteen stations (two minutes each) and were asked to visually quantify the amount of simulated blood loss depicted. They solely used visual cues such as the number of soaked laps, appearance of operation field, etc. however, they were not permitted to touch or move any of the objects. The thirteen stations were arranged in a circuit design. Two to three participants were positioned at each of the thirteen stations along the circuit. The participants were instructed to work individually and not to communicate the answers with each other. After the two-minute period elapsed, an alarm was sounded, and participants were instructed to move to the next station along the circuit loop, while at each station, participants documented their answers on a response sheet that was collected at the end of Part 1. The response sheet solicited demographic information (department, level of experience) and had thirteen blank numbered lines for free form documentation of answers to each simulation station. During Part 1 of the PLS activity, each simulation station was viewed without its corresponding paired object.Following Part 1 and after collection of the response sheets, paired objects along with a card revealing the correct answer were placed at each station. This commenced the transition to Part 2 (thirteen minutes) where participants, in groups or two-three people, revisited each of the thirteen stations for an additional one minute. Participants were free to discuss the station, paired objectives, and original answers among the group members at this time. Like Part 1, an alarm bell was sounded to signal when it was time to move to the next station along the circuit. The final debriefing session (approximately fifty minutes) consisted of open discussion and didactic teaching session. The debriefing was conducted by physician leaders in the ob/gyn department and trauma surgery departments who facilitated the simulation. During this session, the answers to each of the thirteen simulations were discussed. Additionally, formal techniques for visually estimating blood loss were presented to the participants ( 6, ). Each person&,rsquo s performance was measured by scoring of the response sheet that was collected at the end of Part 1. By providing immediate feedback to the learner during the debriefing session, we were consistent in maintaining the core tenets of programmed learning simulation. Statistical Analysis for Part A All statistical analyses were carried out using International Business Machine, Statistical Package for the Social Sciences, Version 26.0 (IBM SPSS). In order to assess normality, data points were plotted and observed. Outliers were not excluded from analysis as these extreme data points reflect trends in gross overestimation or underestimation of blood loss is well documented in the literature ( 4, , 8, , 19, - 20, ). We utilized both percent error and absolute value of percent error as they account for different aspects of error in estimation. The percent error calculation accounts for bidirectional error in either overestimation or underestimation. However, it does not account for the magnitude of error as the positive values cancel the negative ones. Thus, we also report the absolute value of the percent error as it accounts for the absolute magnitude of error in either direction. The percent error of estimated blood loss was calculated for each participant response using theFollowing formula ( 17, ), % Error for blood estimation = ( Estimated Blood Volume - Actual Blood Volume ) Actual Blood Volume &,times 100 Descriptive statistical analysis of percent error (mean, range) was calculated for each of the thirteen stations. The absolute value of percent error estimation was calculated by taking the absolute value of the percent error calculation. A pooled subgroup analysis for all thirteen stations was performed to assess the in&,#64258 uence of years of experience (0-5 years, 6-10 years, and greater than 10 years), size of hemorrhage (Small, 30 ml, 60 ml, and 100 ml Medium, 150 ml, 250 ml, and 500 ml Large, 1000 ml, 1500 ml, and 2000 ml), and occupational status (medical student, resident, obstetric surgeon, general/trauma surgeon, anesthesiologist, and ancillary staff- OR nurses or technicians) on the accuracy of estimation. Univariate analyses for continuous variables were compared using one-way ANOVA. In order to minimize the chance of making at Type 1 error when making multiple group comparisons, we applied a Bonferroni correction (&,alpha new = &,alpha original / number of groups) to an alpha original of 0.05. Thus, for the three group comparisons (years of experience and size of hemorrhage), a p-value of &,lt 0.02 (0.05/3) was considered statistically significant and for 5-way comparison (professional grouping) a p &,lt 0.01 (0.05/5) was considered significant after application of a Bonferroni correction. When differences between three or more groups were statistically significant, post-hoc analysis was performed to detect a significant differences between the groups. Part B, Effect of the PLS in a Clinical Setting In order to test the effect of the PLS activity in a clinical setting, we compared the error of blood loss estimation for cesarean deliveries after the simulation exercise to the period immediately prior. Cesarean deliveries performed during the two-month period prior to the simulation exercise were included in the pre-interventional group and cesarean deliveries that occurred in the two-month period after the simulation exercise were included in the post-interventional group. Exclusion criteria were emergency cesarean deliveries (fetal distress, placental abruption, cord prolapse etc.), patients transfused with blood products intraoperatively or post-operatively during the first 24 hours, or those patients that refused a post-operative Complete Blood Count (CBC). A de-identified database grouped by the month when the cesarean sections occurred was maintained. At our institution, it is routinely practiced for the estimated blood loss (EBL) for each cesarean delivery to be determined by an informal joint consensus of the attending surgeon, resident, and anesthesiologist. The final documentation EBL is done by the resident in the medical record.&,nbsp This practice was the same both prior to and after the PLS. The staff were not told that their EBL values were being monitored by the obstetric department after the PLS activity to avoid observational bias. The participants in Part B included the 19 ob/gyn physicians, 22 ancillary staff members, and 14 anesthesiologists who took part in the PLS exercise. The calculated blood loss (CBL) was obtained by multiplying the maternal blood volume by the percent change in hematocrit ( 18, ),Calculated Blood Loss = (Maternal Blood Volume) x (% change in hematocrit), Where the maternal blood volume was determined for a standardized formula ( 18, ), Maternal Blood Volume = 0.75 x [(maternal height in inches x 50) + (maternal weight in pounds x 25)] and the percent change in hematocrit was calculated from the following formula ( 17, ),&,nbsp % change in hematocrit = ( pre delivery hematocrit - post-delivery hematocrit ) pre delivery hematocrit The pre-delivery CBC, used to assess the hematocrit, was obtained at admission to labor and delivery, prior to the cesarean delivery. The post-operative CBC was performed the morning after the index cesarean delivery, as done routinely at our institution on all post-cesarean patients. Maternal weight and height were determined at admission to Labor and Delivery.The percent error in EBL was calculated using the CBL as a basis of comparison yielding a percent error value ( 17, ), % Error for blood estimation = ( Estimated Blood Loss - Calculated Blood Loss ) Calculated Blood Loss &,times 100 Statistical Analysis for Part B Descriptive statistical analysis of percent error (mean, range) was calculated for each pre- and post-interventional period. The absolute value of percent error estimation was calculated by taking the absolute value of the percent error calculation. Comparison between the pre- and post-intervention data sets was analyzed using student&,rsquo s t-test. A p-value &,lt 0.05 was considered statistically significant. Results Part A A total of 88 clinicians, consisting of 24 medical students, 19 ob/gyn physicians, 22 ancillary staff members, 10 surgeons, and 14 anesthesiologists participated in the PLS exercise.&,nbsp Years of experience among providers were, 0-5 years (n=39), 6-10 years (n=9), and greater than 10 years (n=40). &,nbsp Table 1, depicts mean percent error and absolute value of mean percent error for all the thirteen stations. The results of each station in Table 1, are presented as a mean aggregate of all learners that participated in the simulation.Station #Artificial Blood Volume (ml)Mean Estimate, ml (Range)Mean % Error (Range)Absolute Value of Mean % Error (Standard Deviation)Station 1100120 (10 &,ndash 500)19.94 (-90 &,ndash 400)68.13 (87.12)Station 23053 (5 &,ndash 250)75.76 (-83.33 &,ndash 733.33)110.98 (150.71)Station 36041 (5 &,ndash 150)-30.97 (-91.67 &,ndash 150)53.88 (32.70)Station 4150139 (20 &,ndash 500)-7.42 (-86.67 &,ndash 233.33)54.55 (38.30)Station 51,500603 (50 &,ndash 2,500)-59.81 (-96.67 &,ndash 66.67)63.60 (24.87)Station 6250206 (20 &,ndash 800)-17.55 (-92 &,ndash 220)60.70 (40.49)Station 71,0001,495 (70 &,ndash 3,500)45.94 (-93 &,ndash 250)70.24 (66.80)Station 82,0001,781 (100 &,ndash 4,500)-10.93 (-95 &,ndash 125)39.84 (28.88)Station 9500365 (50 &,ndash 2,000)-27.07 (-90 &,ndash 300)51.84 (41.73)Station 10250267 (50 &,ndash 1,500)6.70 (-80 &,ndash 500)54.11 (65.02)Station 11500226 (20 &,ndash 1,000)-54.86 (-96 &,ndash 100)57.59 (25.68)Station 12150183 (20 &,ndash 900)21.67 (-86.67 &,ndash 500)68.18 (75.72)Station 13150349 (20 &,ndash 1,500)-30.26 (-96 &,ndash 200)50.40 (36.05)EBL, Estimated Blood Loss |