Does Unconscious Bias Affect How Pediatricians Manage Their Patients?

BACKGROUND: Both implicit bias-referred to as unconscious bias-and explicit bias affect how clinicians manage patients. The Implicit Association Test (IAT) has incremental predictive validity relative to self-reports of unconscious bias. Few studies have uniquely specified the impact of unconscious bias in pediatric practices. OBJECTIVE: We aimed to assess the influence of unconscious bias on decision-making in the faculty in a pediatric academic center using the IAT, in addition to and separately applying clinical vignettes with racial and socioeconomic class associations in both tools as it relates to clinicians' race, gender, years in practice, education achieved by the clinician's parents, and language spoken. METHODS: We conducted a prospective quality control evaluation of faculty in an academic center's pediatrics department. An anonymous online tool was used to gather IAT responses, clinical vignette responses, demographics, and explicit bias questions. RESULTS: Of 295 faculty members (73% females), 230 completed the questionnaire, at least in part. Faculty reported on the explicit bias questions, neutral feelings when comparing the demands of educated vs noneducated patients, African American vs European American patients, and patients in the upper vs lower socioeconomic class. Of the approximately two-thirds who answered the IAT, faculty showed preference for European American and upper socioeconomic class. However, the clinical vignettes revealed no differences in how faculty responded to patients based on race or socioeconomic status when stratified by factors listed above, except physicians who favor upper socioeconomic class over lower socioeconomic class were more likely to give a detailed explanation of options if the patient's parent was upper socioeconomic class (P = 0.022). CONCLUSIONS: Pediatricians exhibit racial and socioeconomic unconscious bias that minimally affects decision-making, at least based on vignette responses.

A prospective cohort study of catheter-related thrombosis in cancer patients treated with 1 month of anticoagulation after catheter removal.

The duration of anticoagulation in patients with catheter- related thrombosis (CRT) is not standardized. This is a multiinstitutional prospective pilot study in patients with cancer and upper extremity CRT. Patients received therapeutic enoxaparin for 1 month after catheter removal. Incidence of recurrent thrombosis, hemorrhage, and postthrombotic syndrome (PTS) using the modified Villalta scale, and functional limitation using the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire were assessed at months 1, 3, and 6 after catheter removal. Clopper-Pearson exact 95% confidence intervals (CI), Pearson correlations, and Skillings-Mack, and Wilcoxon signed ranks tests were done. Twenty-seven patients from three institutions were enrolled. Feasibility outcomes were not met. Seventy percent (n = 19) of the cohort had hematologic malignancies. Excluding two patients who were still on enoxaparin at study withdrawal, the median total duration of therapeutic enoxaparin was 32 [interquartile range (IQR) 30-52] days in the remaining 25 patients. During the 6 months after catheter removal, the incidence of recurrent thrombosis was 0% (n = 0/20, 95% CI 0-17%) and major hemorrhage was 5% (n = 1/20, 95% CI 0.13-25%). One patient (5%, 95% CI 0.13-25%) had PTS in the affected arm at any visit, and none had severe PTS. Higher PTS scores were associated with higher DASH scores. DASH scores at month 6 were significantly lower compared with month 1 (P = 0.0066). No deaths occurred. A multicenter pilot study of treatment with anticoagulation for 1 month after catheter removal did not meet feasibility outcomes but we found no recurrent thrombosis and a low incidence of PTS.

Duration of labor induction in nulliparous women with hypertensive disorders of pregnancy and maternal and neonatal outcomes.

OBJECTIVE: The objective of this study was to quantify the association between duration of labor induction in nulliparous women with hypertensive disorders of pregnancy and maternal and neonatal morbidity. METHODS: This was a secondary analysis of a multicenter cohort study of 228,438 deliveries in 19 U.S. hospitals. The analysis included nulliparous women ≥18 years old with singleton gestation diagnosed with hypertensive disorders of pregnancy and undergoing induction of labor for that indication. Duration of labor induction, defined as time from admission to delivery, was examined by 4 h intervals from <12 h to ≥24 h in relation to maternal and neonatal composite outcomes. Maternal composite outcome included operative vaginal delivery, chorioamnionitis, blood transfusion, intensive care unit admission, placental abruption, 3rd or 4th degree perineal laceration, endometritis, postpartum hemorrhage, or venous thromboembolism. Neonatal composite outcome included neonatal intensive care unit (NICU) admission, respiratory distress syndrome, 5-minute Apgar score ≤7, seizure, infection, intrapartum meconium aspiration, intracranial hemorrhage, shoulder dystocia, and neonatal death. The trends in proportions of outcomes that occurred at different intervals were examined by Cochran-Armitage trend test. Relative risks were calculated with <12 h as the reference category and potential confounders adjusted by log-binomial or Poisson regression. Possible correlations within centers were taken into account using generalized estimating equations. RESULTS: A total of 3,990 women met inclusion criteria. The median labor duration was 19.8 h (interquartile range 12.9 h-27.9h), with 849 (21.3%) lasting <12 h and 1,426 (35.7%) >24 h. The frequency of composite maternal outcome was not associated with labor duration; however, the rates of chorioamnionitis (p < .001) and postpartum hemorrhage (p < .001) increased as labor duration increased. The frequency of composite neonatal outcome was greater with increasing labor duration (p < .001). After multivariable adjustment, duration of labor induction was associated with increased risks of maternal composite outcome after 24 h (aRR 1.39, 95% CI 1.20-1.62) and neonatal composite outcome after 24 h (aRR 1.32, 95% CI 1.11-1.56). CONCLUSIONS: In nulliparous women with hypertensive disorders of pregnancy, duration of labor induction was associated with increased risks for maternal and neonatal morbidity after 24 h.

Risk of pulmonary emboli after removal of an upper extremity central catheter associated with a deep vein thrombosis.

Standard treatment of catheter-associated upper extremity deep vein thrombosis (UE-DVT) is anticoagulation, although catheters are often removed for this indication. The optimal time for catheter removal and whether the act and/or timing of catheter removal is associated with pulmonary embolism (PE) remain unknown. A retrospective cohort study was performed at 8 participating institutions through the Venous thromboEmbolism Network US. Patients with hematologic malignancies and central venous catheter (CVC)-associated UE-DVT were included from 1 January 2010 through 31 December 2016. The primary outcome was objectively confirmed PE within 7 days of UE-DVT diagnosis in anticoagulated patients comparing early (≤48 hours) vs delayed (>48 hours) catheter removal. A total of 626 patients were included, among whom 480 were treated with anticoagulation. Among anticoagulated patients, 255 underwent early CVC removal, while 225 had delayed or no CVC removal; 146 patients received no anticoagulation, among whom 116 underwent CVC removal alone. PE within 7 days occurred in 2 patients (0.78%) with early removal compared with 1 patient (0.44%) with delayed or no CVC removal (P > .9). PE or any cause of death within 7 days occurred in 3 patients in both the early removal (1.18%) and delayed/no removal (1.33%) groups (P > .9). In patients treated with CVC removal only (no anticoagulation), there were no PEs but 3 deaths within 7 days. In patients with hematological malignancy and CVC-associated UE-DVT, early removal of CVCs was not associated with an increased risk of PE compared with delayed or no removal.

Treatment of catheter-related thrombosis in patients with hematologic malignancies: A Venous thromboEmbolism Network U.S. retrospective cohort study.

INTRODUCTION: Optimal treatment of catheter-related thrombosis (CRT) is uncertain in patients with hematologic malignancy. We aimed to evaluate the treatment strategies, outcomes, and predictors of recurrent venous thromboembolism (VTE) associated with catheter-related thrombosis (CRT) in patients with hematologic malignancy. METHODS: We performed a multicenter retrospective cohort study of eight institutions through the Venous thromboEmbolism Network US. Patients with hematologic malignancies with documented CRT were identified using ICD-9 and ICD-10 diagnostic codes. Semi-competing risks proportional hazard regression models were created. RESULTS AND CONCLUSIONS: Of the 663 patients in the cohort, 124 (19%) were treated with anticoagulation alone, 388 (58%) were treated with anticoagulation and catheter removal, 119 (18%) treated with catheter removal only, and 32 (5%) had neither catheter removal nor anticoagulation. 100 (15%) patients experienced a recurrent VTE event. In the 579 patients who had catheter removal, the most common reason for catheter removal was the CRT [392 (68%)]. For subjects who received any anticoagulation (n = 512), total anticoagulation duration was not associated with VTE recurrence [1.000 (0.999-1.002)]. After adjustment patients treated with catheter removal only had an increased risk of VTE recurrence [2.50 (1.24-5.07)] and death [4.96 (2.47-9.97)]. Patients with no treatment had increased risk of death [16.81 (6.22-45.38)] and death after VTE recurrence [27.29 (3.13-238.13)]. In this large, multicenter retrospective cohort, we found significant variability in the treatment of CRT in patients with hematologic malignancy. Treatment without anticoagulation was associated with recurrent VTE.

Increases in oxygen saturation following discharge from Fontan palliation - an indicator of resolution of pulmonary arteriovenous malformations?

BACKGROUND: Pulmonary arteriovenous malformations in single ventricle congenital heart disease are poorly understood. Previous studies investigating pulmonary arteriovenous malformations predominantly focus on patients with heterotaxy syndrome and interrupted inferior caval vein. It is unknown if development and resolution of pulmonary arteriovenous malformations are similar for patients with and without heterotaxy syndrome. METHODS: In this retrospective single-institution study, we identified patients with a history of single ventricle congenital heart disease and Fontan palliation. We then matched patients with heterotaxy syndrome (intact and interrupted inferior caval vein) and non-heterotaxy hypoplastic left heart syndrome. To compare development of pulmonary arteriovenous malformations, we identified the frequency of positive diagnoses pre-Fontan. To compare resolution of pulmonary arteriovenous malformations, we recorded oxygen saturation changes for 12 months following Fontan. RESULTS: A total of 124 patients were included. Patients with heterotaxy and interrupted inferior caval vein were more likely to have a pre-Fontan contrast echocardiogram performed (p < 0.01) and more likely to be diagnosed with pulmonary arteriovenous malformations pre-Fontan (p < 0.01). There was no difference in oxygen saturation prior to Fontan, yet all patient groups had increased their oxygen saturations in the first year after Fontan discharge. CONCLUSIONS: Pulmonary arteriovenous malformations are variably diagnosed prior to Fontan palliation; however, all study groups had increased oxygen saturations after Fontan discharge, potentially indicating resolution of pulmonary arteriovenous malformations in all groups. The prevalence of pulmonary arteriovenous malformations pre-Fontan is likely underestimated. A quantitative, systematic approach to diagnosis and follow-up of pulmonary arteriovenous malformations is needed to better understand susceptibility and pathophysiology.

Rituximab and antimetabolite treatment of granulomatous and lymphocytic interstitial lung disease in common variable immunodeficiency.

BACKGROUND: Granulomatous and lymphocytic interstitial lung disease (GLILD) is a life-threatening complication in patients with common variable immunodeficiency (CVID), but the optimal treatment is unknown. OBJECTIVE: Our aim was to determine whether rituximab with azathioprine or mycophenolate mofetil improves the high-resolution computed tomography (HRCT) chest scans and/or pulmonary function test results in patients with CVID and GLILD. METHODS: A retrospective chart review of clinical and laboratory data on 39 patients with CVID and GLILD who completed immunosuppressive therapy was performed. Chest HRCT scans, performed before therapy and after the conclusion of therapy, were blinded, randomized, and scored independently by 2 radiologists. Differences between pretreatment and posttreatment HRCT scan scores, pulmonary function test results, and lymphocyte subsets were analyzed. Whole exome sequencing was performed on all patients. RESULTS: Immunosuppressive therapy improved patients' HRCT scan scores (P < .0001), forced vital capacity (P = .0017), FEV1 (P = .037), and total lung capacity (P = .013) but not their lung carbon monoxide diffusion capacity (P = .12). Nine patients relapsed and 6 completed retreatment, with 5 of 6 of these patients (83%) having improved HRCT scan scores (P = .063). Relapse was associated with an increased number of B cells (P = .016) and activated CD4 T cells (P = .016). Four patients (10%) had pneumonia while undergoing active treatment, and 2 patients (5%) died after completion of therapy. Eight patients (21%) had a damaging mutation in a gene known to predispose (TNFRSF13B [n = 3]) or cause a CVID-like primary immunodeficiency (CTLA4 [n = 2], KMT2D [n = 2], or BIRC4 [n = 1]). Immunosuppression improved the HRCT scan scores in patients with (P = .0078) and without (P < .0001) a damaging mutation. CONCLUSIONS: Immunosuppressive therapy improved the radiographic abnormalities and pulmonary function of patients with GLILD. A majority of patients had sustained remissions.

Echocardiographic Identification of Pulmonary Artery Flow Reversal: An Indicator of Adverse Outcomes in Single Ventricle Physiology.

Individuals with single ventricle congenital heart disease (CHD) undergo multiple staged surgical palliations. Staged single ventricle palliation with a superior cavopulmonary connection (SCPC) in infancy followed by a Fontan in early childhood relies on passive, unobstructed pulmonary blood flow and normal pulmonary vasculature. We hypothesized that patients with echocardiographic identification of retrograde flow in a branch pulmonary artery (PA) after SCPC or Fontan are at increased risk for adverse outcomes. We conducted a retrospective chart review of patients seen at Children's Wisconsin from 1999 to 2019. Inclusion criteria included a history of single ventricle congenital heart disease and surgical palliation with a superior cavopulmonary connection (SCPC). We created two cohorts based on transthoracic echocardiographic identification of branch PA flow patterns: those with color Doppler-defined pulmonary artery flow reversal (PA reversal cohort) and those with normal anterograde flow (Non-reversal cohort). We identified 21 patients in the PA reversal cohort and 539 patients in the Non-reversal cohort. The PA reversal cohort had increased hospital length of stay after SCPC palliation (p < 0.001) and decreased transplant-free survival (p = 0.032), but there was no difference in overall survival (p = 0.099). There was no difference in hospital length of stay after Fontan (p = 0.17); however, the PA reversal cohort was significantly less likely to progress to Fontan palliation during early childhood (p = 0.005). Echocardiographic color Doppler identification of branch PA flow reversal in patients with single ventricle physiology is a high-risk indicator for adverse short- and long-term outcomes.

Rates of Venous Thromboembolism and Central Line-Associated Bloodstream Infections Among Types of Central Venous Access Devices in Critically Ill Children.

OBJECTIVES: Central venous access devices, including peripherally inserted central catheters and central venous catheters, are often needed in critically ill patients, but also are associated with complications, including central-line associated bloodstream infections and venous thromboembolism. We compared different central venous access device types and these complications in the PICU. DESIGN: Multicenter, cohort study. SETTING: One hundred forty-eight participating Virtual PICU Systems, LLC, hospital PICU sites. PATIENTS: Pediatric patients with central venous access placed from January 1, 2010, to December 31, 2015. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Patient and central venous access device variables postulated to be associated with central-line associated bloodstream infection and venous thromboembolism were included. Data were analyzed using Pearson chi-square test or Fisher exact test for categorical variables, Mann-Whitney U test for continuous variables, and logistic regression and classification trees for multivariable analysis that examined significant predictors of venous thromboembolism or central-line associated bloodstream infection. Analysis included 74,196 first lines including 4,493 peripherally inserted central catheters and 66,194 central venous catheters. An increased rate of venous thromboembolism (peripherally inserted central catheter: 0.93%, central venous catheter: 0.52%; p = 0.001) (peripherally inserted central catheter: 8.65/1,000 line days, central venous catheter: 6.29/1,000 line days) and central-line associated bloodstream infection (peripherally inserted central catheter: 0.73%, central venous catheter: 0.24%; p = 0.001) (peripherally inserted central catheter: 10.82/1,000 line days, central venous catheter: 4.97/1,000 line days) occurred in peripherally inserted central catheters. In multivariable analysis, central venous catheters had decreased association with central-line associated bloodstream infection (odds ratio, 0.505; 95% CI, 0.336-0.759; p = 0.001) and venous thromboembolism (odds ratio, 0.569; 95% CI, 0.330-0.982; p = 0.043) compared with peripherally inserted central catheters. CONCLUSIONS: Peripherally inserted central catheters are associated with higher rates of central-line associated bloodstream infection and venous thromboembolism than central venous catheters in children admitted to the PICU.

Chronic pain persists in adults with sickle cell disease despite regular red cell transfusions.

BACKGROUND: Pain affects over 50% of adults with sickle cell disease (SCD), and this pain is largely managed outside of the hospital. While chronic transfusion therapy is used to decrease the rate of acute pain events in patients with SCD, less is known about its impact on the day-to-day experience of pain. To address this knowledge gap, we provided pain diaries to patients with SCD receiving chronic transfusion. PATIENTS AND METHODS: A convenience sample of chronically-transfused adults with SCD successfully completed a diary over the course of at least 2 transfusion events. Patients receiving simple transfusions and red cell exchanges were included. Pain was rated on a scale of 0 to 10 each day, and patient laboratory values, co-morbidities, and hospital utilization were also obtained using the electronic medical record. The mean pain scores pre- and post-transfusion were evaluated using both a random effects-expectation maximization regression tree analysis and a generalized linear mixed regression model. RESULTS: Ten subjects (63%) in this cohort were defined as having chronic pain, while the remaining four (27%) subjects had episodic pain. Despite chronic transfusion and a suppressed HbS% (22.5% (16.5-25.9)), 10 patients (63%) continued to report nearly daily pain, and on almost 70% of diary days, the pain was significant (≥5/10). When the relationship between HbS% and reported pain intensity was examined, no association was found. DISCUSSION: These results suggest that, even with regular transfusions and a low HbS%, daily pain persists in many adults with SCD.

Cardiovagal modulation in pediatric functional gastrointestinal disorders.

BACKGROUND AND OBJECTIVES: Though reduced cardiovagal modulation accompanies adult IBS, adolescents with functional gastrointestinal disorders (FGID) have not been studied. We aim to investigate whether adolescents with FGID have reduced cardiovagal modulation. METHODS: After 10-minute supine rest, we recorded ECG for 5-minute supine and 5-minute standing without support in healthy and FGID-affected adolescents. After analysis with Kubios 2.2 for high-frequency (hf) and low-frequency (lf) heart rate variability (HRV), Wilcoxon signed-ranks test compared individual paired supine and standing HRV data, while Kruskal-Wallis and Mann-Whitney U tests compared HRV across groups. RESULTS: A total of 50 FGID subjects (90% females; median age 17 years [range 11-21]) and 22 healthy comparison group (HC) (59% females; median age 14 years [range 10-18]) participated. Both absolute and relative supine hfHRV exceeded standing in both groups. Absolute supine lfHRV was higher than standing in FGID patients and not in HCs, while relative supine lfHRV power was actually lower in both groups. Compared to HC, FGID group showed significantly lower absolute HRV, while relative HRV did not differ between groups. CONCLUSIONS: Cardiovagal modulation is lower in adolescents with FGID. This difference impacts these subjects significantly. Whether this finding reflects a cause or a consequence of FGID is unknown.

The association of tea consumption with bladder cancer risk: a meta-analysis.

The association between tea consumption and bladder cancer has been confirmed in several animal studies, but one epidemiological study in 2001 showed no association between them. In order to provide an accurate assessment of this, we conducted a meta-analysis on tea consumption and bladder cancer risk. Studies were identified by a literature search in PubMed from January 1980 to March 2012 and the reference lists of relevant studies. Random effect models were used to calculate summary relative risk estimates (RR) and their corresponding 95% confidence intervals (CI) based on high contrast to low intake values. Twenty-four publications (6 cohort studies and 18 case-control studies) based on consumption of overall tea, black tea, and green tea to bladder cancer risk were included in this analysis. For overall tea, the summary RR indicated no association between tea consumption and bladder cancer (RR= 1.09, 95%CI: 0.85-1.40). In subgroup analyses, we found a moderate increase of bladder cancer risk in smoking group (RR= 1.77, 95%CI: 1.04-3.01). In the black tea group, no statistically significant association was observed (RR= 0.84, 95%CI: 0.70-1.01). Interestingly, in the subgroup of sex, a protective effect was observed between tea consumption and bladder cancer risk in female (RR= 0.61, 95%CI: 0.38- 0.98). For green tea group, there was no relationship associated with bladder cancer risk (RR= 1.03, 95%CI: 0.82- 1.31). In conclusion, our data suggest that high overall tea intake in smokers increased the risk of bladder cancer, and high black tea intake in female may reduce the risk of bladder cancer.