Large Academic Healthcare System Research Articles

Abstract WP382: Using Natural Language Processing Algorithms to Identify Stroke Cases and Stroke Subtypes From Neuroimaging Reports

Background/Objective: The long-term goal of our research is to develop automated, accurate methods for conducting stroke outcome surveillance across large populations. To facilitate this, we developed natural language processing (NLP) based machine learning algorithms to classify stroke cases and sub-type strokes using neuroimaging reports. We report on the performance of our algorithms. Methods: Our population of interest included patients with stroke symptoms presenting to the emergency room of our large academic healthcare system. We randomly sampled 332 probable stroke cases. A trained neurologist validated stroke diagnoses using neuroimaging reports. Data preprocessing included cleaning and normalizing the reports into a standardized format. We trained and tested machine learning algorithms using the formatted reports. The NLP-based algorithms predicted a stroke diagnosis (binary classification) and a stroke type diagnosis (multiclass classification) using n-grams of length 1 (i.e., ‘stroke’, ‘hemorrhage’) through 3 (i.e., ‘no mass effect’), term-frequency weighting, and feature dimensionality reduction via truncated singular value decomposition (SVD). We report algorithm performance using the area under the receiver operating characteristic curve (AUC-ROC). Classification methods we tested included Multinomial Naïve Bayes, Logistic Regression, Random Forest Classifier, and Support Vector Machine (SVM). Results: The highest performing algorithm for both stroke and stroke sub-type classification contained 1 to 2 n-grams, no term-frequency, and 200 SVD components. For the stroke case detection, SVM achieved the best AUC-ROC of 95.6%. For stroke sub-type detection, SVM also yielded the highest AUC-ROC of 93.5% for no stroke, 92.3% for ischemic stroke, 91.9% for intraparenchymal hemorrhage, and 94.8% for subarachnoid hemorrhage. Conclusions: We report very promising results in our pilot study using machine learning algorithms to classify stroke cases from neuroimaging reports. Feature selection on this pilot data revealed a subset of words that are highly effective in categorizing stroke. Future work will focus on algorithm improvement, finer grained stroke sub-type stratification, and multi-label phenotyping.

1857. Implementing Antibiotic Stewardship in Urgent Care Centers

BackgroundAntibiotic stewardship (AS) has historically focused on inpatient facilities and primary care clinics; many antibiotics (ABx) are prescribed in urgent care clinics (UCCs). However, few centers have described implementing AS in such settings. We sought to reduce total ABx use in our UCCs as well as specifically decrease azithromycin use.MethodsWe conducted this study in four UCCs owned by a large community-based academic healthcare system in northern Delaware. The UCCs average >65,000 visits annually and include 38 providers (physicians, physician assistants and nurse practitioners). A new electronic health record was implemented in October 2016; ABx utilization data are not available prior to this time. Beginning in January 2017, all providers received in-person education on guideline-recommended management of common infectious diseases, including bronchitis, sinusitis, and pharyngitis. The lead physician performed chart audits and provided group and individual education and feedback via email and telephone. Individual ABx utilization rates were not provided, but documentation of rationale for ABx need was emphasized. Patient education included ABx links on the check-in website, posters in waiting and examination rooms, and patient education materials embedded within each discharge packet, with an emphasis on providing evidence-based care rather than “denying ABx.” We calculated number of total ABx prescriptions (Rx) and of azithromycin Rx per 100 visits per month, and calculated rate ratios comparing January 2017 (pre-intervention) to January 2018 (post).ResultsDuring the 16-month intervention period, total ABx use declined from 67 Rx per 100 visits to 44/100 visits (rate ratio, 0.55, 95% CI 0.37–0.80) and azithromycin use declined from 13 Rx/100 visits to 5/100 visits (RR 0.32, 95% CI 0.10–0.88). Seasonal variability was apparent (figure). ConclusionA multifaceted educational approach positively impacted provider behaviors and patient expectations, and did not rely upon providing ABx utilization data (either clinic- or individual-level). Ensuring leadership support of providers if patients expressed dissatisfaction and standardized messaging and tools were critical for managing patient expectations.Disclosures All authors: No reported disclosures.

1848. Evaluation of Antibiotic Prescribing Practices in Outpatient Clinics for the Treatment of Skin and Soft-Tissue Infections

BackgroundAmbulatory visits for the treatment of skin and soft-tissue infections (SSTIs) have doubled over the past decade and are one of the most common reasons for infection-related visits to outpatient clinics. However, there is limited data evaluating antibiotic prescribing in this population. We aimed to assess the management of SSTIs in adult patients in order to target interventions to improve antibiotic utilization and optimize outcomes.MethodsThis retrospective study included patients within a large academic healthcare system presenting to 38 clinics. Patients were included if they had a diagnosis of a SSTI (ICD-10 for cutaneous abscess, cellulitis, and local SSTIs) in 2016. The primary outcome was to evaluate prescriber compliance to institutional guidelines based on antibiotic selection and duration of therapy. Patient allergies and culture results were used when determining compliance to first-line (trimethoprim–sulfamethoxazole for cutaneous abscess, or cephalexin +/- trimethoprim–sulfamethoxazole for cellulitis and local SSTIs) and second-line recommendations (doxycycline for cutaneous abscess, or clindamycin for cellulitis and local SSTIs). Duration of therapy of 5–7 days was considered compliant.ResultsA total of 2,170 adult encounters for the treatment of SSTIs were included; 1,588 with cellulitis, 413 with local infection and 169 with cutaneous abscess. The overall compliance rate for appropriate therapy, including drug selection and duration, was 64.9% (see Figure 1). Unnecessarily long duration of therapy resulted in an extra 1,637 days of antibiotic therapy. Compliance with drug selection occurred more frequently with physicians (40.3%) compared with residents (33.9%) and Advanced Practice Providers (APP) (25.1%).ConclusionCompliance with an institutional SSTI guideline for antibiotic selection and duration of therapy is suboptimal in outpatient clinics. Stewardship interventions for SSTIs should target both drug selection and duration, and APPs as an important provider group in outpatient settings.Figure 1.Compliance Stratified by Infection TypeInfection typeOverall ComplianceCompliance with Drug SelectionCompliance with Duration of TherapyCutaneous abscess29%55%57%Cellulitis78%82%60%Local infection of skin/soft tissue16%85%60%Disclosures All authors: No reported disclosures.

Developing primary care teams prepared to improve quality: a mixed-methods evaluation and lessons learned from implementing a microsystems approach

BackgroundHealth systems in the United States are increasingly required to become leaders in quality to compete successfully in a value-conscious purchasing market. Doing so involves developing effective clinical teams using approaches like the clinical microsystems framework. However, there has been limited assessment of this approach within United States primary care settings.MethodsThis paper describes the implementation, mixed-methods evaluation results, and lessons learned from instituting a Microsystems approach across 6 years with 58 primary care teams at a large Midwestern academic health care system. The evaluation consisted of a longitudinal survey augmented by interviews and focus groups. Structured facilitated longitudinal discussions with leadership captured ongoing lessons learned. Quantitative analysis employed ordinal logistic regression and compared aggregate responses at 6-months and 12-months to those at the baseline period. Qualitative analysis used an immersion/crystallization approach.ResultsSurvey results (N = 204) indicated improved perceptions of: organizational support, team effectiveness and cohesion, meeting and quality improvement skills, and team communication. Thematic challenges from the qualitative data included: lack of time and coverage for participation, need for technical/technology support, perceived devaluation of improvement work, difficulty aggregating or spreading learnings, tensions between team and clinic level change, a part-time workforce, team instability and difficulties incorporating a data driven improvement approach.ConclusionsThese findings suggest that a microsystems approach is valuable for building team relationships and quality improvement skills but is challenged in a large, diverse academic primary care context. They additionally suggest that primary care transformation will require purposeful changes implemented across the micro to macro-level including but not only focused on quality improvement training for microsystem teams.

Normal Variant Occipital Pneumatization.

Evaluate prevalence and pattern of occipital pneumatization (OP). Individuals with imaging evidence of OP on computed tomography (CT). High resolution CT scans of the temporal bone. The prevalence and pattern of OP on 1000 CT scans performed at a large academic healthcare system. OP had a prevalence of 11.8% (n = 118) while occipital condyle pneumatization occurred in only 0.3%. Occipital air cells were right-sided in 27.1% (n = 32), left-sided in 51.7% (n = 61), and bilateral in 21.2% (n = 25), and OP was contiguous with the occipitomastoid suture. In cases of unilateral OP, the contralateral jugular foramen demonstrated relative enlargement (p = 0.006), but a direct association could not be established. The occipital bone is an accessory site of skull base pneumatization in 11.8% of the population undergoing temporal bone CT, while air cell extension into the occipital condyle occurs more rarely. These occipital air cells have a left-sided predilection and are always in direct proximity to the occipitomastoid suture, suggesting developmental egress from the mastoid. OP is a normal variant and lacks features of craniocervical pneumatization, which is a rare disorder.

Interfacility Transport of the Pregnant Patient: A 5-year Retrospective Review of a Single Critical Care Transport Program

Introduction: Interfacility transport of the pregnant patient poses a challenge for prehospital providers as it is an infrequent but potentially high acuity encounter. Knowledge of clinically significant events (CSEs) that occur during these transports is important both to optimize patient safety and also to help enhance crew training and preparedness. This study evaluated a critical care transport program’s 5-year longitudinal experience transporting pregnant patients by ground and air, and described CSEs that occurred during the out-of-hospital phase of care. Methods: This study was a retrospective review of pregnant patients transported by a single critical care transport system into and within a large academic healthcare system. Patients who were pregnant, and were transported from a referring facility to one of the 2 receiving centers within Johns Hopkins Health System between January 1, 2012 and December 31, 2016 were included in this study. The primary outcome of interest was the occurrence of a predefined clinically significant event (CSE) during transport, while a secondary outcome of interest was the indication for transfer. Results: During the study period 1,223 pregnant patients were transported by our critical care transport service. There were 1,101 patients who met inclusion criteria; 693 (62.9%) of whom were transported by ground and 408 (37.1%) who were transported by rotor wing aircraft. The top 3 indications for transfer comprised 71.4% of all patients and included; preterm labor, hypertensive disorder of pregnancy, and other maternal life threatening disorder. The most common events that occurred across all transports were: exacerbation of hypertensive disease requiring intervention (4.5%), hypotension (1.3%), and altered mental status (0.2%). Conclusions: Incidence of CSEs during the interfacility transport of pregnant patients within our critical care transport system is low (6.0%). Knowledge of the clinically significant events that occur during EMS transport is a vital component of ensuring system quality and optimizing patient safety. This data can be used to augment and focus provider education and training to mitigate and optimize response to future events.

Assessing patient risk of central line-associated bacteremia via machine learning

Central line-associated bloodstream infections (CLABSIs) contribute to increased morbidity, length of hospital stay, and cost. Despite progress in understanding the risk factors, there remains a need to accurately predict the risk of CLABSIs and, in real time, prevent them from occurring. A predictive model was developed using retrospective data from a large academic healthcare system. Models were developed with machine learning via construction of random forests using validated input variables. Fifteen variables accounted for the most significant effect on CLABSI prediction based on a retrospective study of 70,218 unique patient encounters between January 1, 2013, and May 31, 2016. The area under the receiver operating characteristic curve for the best-performing model was 0.82 in production. This model has multiple applications for resource allocation for CLABSI prevention, including serving as a tool to target patients at highest risk for potentially cost-effective but otherwise time-limited interventions. Machine learning can be used to develop accurate models to predict the risk of CLABSI in real time prior to the development of infection.

Administrative Costs Associated With Physician Billing and Insurance-Related Activities at an Academic Health Care System

Administrative costs in the US health care system are an important component of total health care spending, and a substantial proportion of these costs are attributable to billing and insurance-related activities. To examine and estimate the administrative costs associated with physician billing activities in a large academic health care system with a certified electronic health record system. This study used time-driven activity-based costing. Interviews were conducted with 27 health system administrators and 34 physicians in 2016 and 2017 to construct a process map charting the path of an insurance claim through the revenue cycle management process. These data were used to calculate the cost for each major billing and insurance-related activity and were aggregated to estimate the health system's total cost of processing an insurance claim. Estimated time required to perform billing and insurance-related activities, based on interviews with management personnel and physicians. Estimated billing and insurance-related costs for 5 types of patient encounters: primary care visits, discharged emergency department visits, general medicine inpatient stays, ambulatory surgical procedures, and inpatient surgical procedures. Estimated processing time and total costs for billing and insurance-related activities were 13 minutes and $20.49 for a primary care visit, 32 minutes and $61.54 for a discharged emergency department visit, 73 minutes and $124.26 for a general inpatient stay, 75 minutes and $170.40 for an ambulatory surgical procedure, and 100 minutes and $215.10 for an inpatient surgical procedure. Of these totals, time and costs for activities carried out by physicians were estimated at a median of 3 minutes or $6.36 for a primary care visit, 3 minutes or $10.97 for an emergency department visit, 5 minutes or $13.29 for a general inpatient stay, 15 minutes or $51.20 for an ambulatory surgical procedure, and 15 minutes or $51.20 for an inpatient surgical procedure. Of professional revenue, professional billing costs were estimated to represent 14.5% for primary care visits, 25.2% for emergency department visits, 8.0% for general medicine inpatient stays, 13.4% for ambulatory surgical procedures, and 3.1% for inpatient surgical procedures. In a time-driven activity-based costing study in a large academic health care system with a certified electronic health record system, the estimated costs of billing and insurance-related activities ranged from $20 for a primary care visit to $215 for an inpatient surgical procedure. Knowledge of how specific billing and insurance-related activities contribute to administrative costs may help inform policy solutions to reduce these expenses.

Provider Perceptions of the Organization's Cultural Competence Climate and Their Skills and Behaviors Targeting Patient-Centered Care for Socially At-Risk Populations.

As part of a cultural competence needs assessment study at a large academic health care system, we conducted a survey among 1,220 practicing physicians to assess their perceptions of the organization's cultural competence climate and their skills and behaviors targeting patient-centered care for culturally and socially diverse patients. Less than half of providers reported engaging in behaviors to address cultural and social barriers more than 75% of the time. In multivariable logistic regression models, providers who reported moderate or major structural problems were more likely to report low skillfulness in identifying patient mistrust (aOR: 2.01; 95% CI: 1.23-3.28, p<0.01), how well patients read and write English (aOR: 1.63; 95% CI: 1.03-2.57, p=0.03), and socioeconomic barriers (aOR: 2.14; 95% CI: 1.14-4.01, p=0.01), than providers who reported only small or no structural problems. Improved structural support for socially and culturally complex medical encounters is needed to enhance care for socially at-risk patients.

Advancing LGBT Health Care Policies and Clinical Care Within a Large Academic Health Care System: A Case Study

ABSTRACTCulturally competent health care is especially important among sexual and gender minority patients because poor cultural competence contributes to health disparities. There is a need to understand how to improve health care quality and delivery for lesbian, gay, bisexual, and transgender (LGBT) veterans in particular, because they have unique physical and mental health needs as both LGBT individuals and veterans. The following article is a case study that focuses on the policy and clinical care practices related to LGBT clinical competency, professional training, and ethical provision of care for veteran patients in the VA Boston Healthcare System. We apply Betancourt et al.’s (2003) cultural competence framework to outline the steps that VA Boston Healthcare System took to increase cultural competency at the organizational, structural, and clinical level. By sharing our experiences, we aim to provide a model and steps for other health care systems and programs, including other VA health care systems, large academic health care systems, community health care systems, and mental health care systems, interested in developing LGBT health initiatives.

“Think aloud” and “Near live” usability testing of two complex clinical decision support tools

ObjectivesLow provider adoption continues to be a significant barrier to realizing the potential of clinical decision support. “Think Aloud” and “Near Live” usability testing were conducted on two clinical decision support tools. Each was composed of an alert, a clinical prediction rule which estimated risk of either group A Streptococcus pharyngitis or pneumonia and an automatic order set based on risk. The objective of this study was to further understanding of the facilitators of usability and to evaluate the types of additional information gained from proceeding to “Near Live” testing after completing “Think Aloud”. MethodsThis was a qualitative observational study conducted at a large academic health care system with 12 primary care providers. During “Think Aloud” testing, participants were provided with written clinical scenarios and asked to verbalize their thought process while interacting with the tool. During “Near Live” testing participants interacted with a mock patient. Morae usability software was used to record full screen capture and audio during every session. Participant comments were placed into coding categories and analyzed for generalizable themes. Themes were compared across usability methods. Results“Think Aloud” and “Near Live” usability testing generated similar themes under the coding categories visibility, workflow, content, understand-ability and navigation. However, they generated significantly different themes under the coding categories usability, practical usefulness and medical usefulness. During both types of testing participants found the tool easier to use when important text was distinct in its appearance, alerts were passive and appropriately timed, content was up to date, language was clear and simple, and each component of the tool included obvious indicators of next steps. Participant comments reflected higher expectations for usability and usefulness during “Near Live” testing. For example, visit aids, such as automatically generated order sets, were felt to be less useful during “Near-Live” testing because they would not be all inclusive for the visit. ConclusionsThese complementary types of usability testing generated unique and generalizable insights. Feedback during “Think Aloud” testing primarily helped to improve the tools’ ease of use. The additional feedback from “Near Live” testing, which mimics a real clinical encounter, was helpful for eliciting key barriers and facilitators to provider workflow and adoption.

Temporal trends and risk factors for healthcare-associated vancomycin-resistant enterococci in adults

Published data regarding temporal trends in vancomycin-resistant enterococci (VRE) prevalence within specific regions or healthcare systems are scarce. To characterize temporal trends and risk factors for healthcare-associated infections caused by VRE. The study included all adult discharges occurring from 2006 to 2014 with an enterococcal infection from three hospitals in a large academic healthcare system. Bivariate analyses were used to identify statistically significant factors associated with vancomycin-susceptible or -resistant infection. Statistically significant variables were included in a final logistic regression model. Trends assessed whether the proportion of enterococcal infections resistant to vancomycin changed over time. The sample included 10,186 adults with first-time healthcare-associated enterococcal infection. Significant risk factors (P≤0.05) for VRE in the final logistic regression model included: tertiary 1 hospital, intensive care unit length of stay, higher Charlson Comorbidity Index, previous immunosuppressive or chemotherapeutic medications, previous hospitalization, renal failure, malignancy, longer length of stay prior to infection, taking an antibiotic prior to infection, being female, and having an infection in winter or spring. Between 2006 and 2014, the rate of resistance varied from 37.1 to 42.9% but there were no significant differences in the proportion resistant to vancomycin over time (P=0.36). Research targeted at risk factors is important to decrease the amount of VRE infections.

Craniocervical Pneumatization: Estimation of Prevalence and Imaging of Treatment Response.

Estimate the prevalence of craniocervical pneumatization (CCP) and describe successful treatment of this condition with clinical and radiologic correlation. Individuals with documented CCP on computed tomography (CT). CT scans of the head, temporal bone, face, neck, and cervical spine. Cessation of habitual Valsalva maneuver (VM) and insertion of pressure-equalization (PE) tubes. The prevalence of CCP on CT examinations performed during two decades in a large academic healthcare system. Documentation of symptomatic and imaging improvement following treatment of CCP. Radiology database review identified two cases of CCP out of a total of 636,854 head and neck CT scans (0.00031%) or 43,553 temporal bone CT scans (0.0046%). Both CCP patients were symptomatic (aural fullness, hearing loss, vertigo) and practiced habitual VM. One patient showed decreased CCP 4 months following cessation of VM, whereas the other patient improved symptomatically and demonstrated reversion of CCP to normal bone on magnetic resonance imaging (MRI) within 3 years following PE tube placement. Acquired CCP is exceedingly rare and, when successfully treated, reversion to normal bone may accompany symptom resolution.

Point-of-Care Testing in a Large Academic Healthcare System

From the Department of Pathology and Laboratory Medicine, University of North Carolina, Chapel Hill, NC. Reprints: Nichole Korpi-Steiner, PhD, DABCC, FACB, University of North Carolina, Clinical Box 7525, Room 1075, East Wing, Chapel Hill, NC 27599. E-mail: [email protected]. The author declares no conflict of interest.

A Comprehensive CT Dose Reduction Program Using the ACR Dose Index Registry

PurposeThe purpose of this article is to demonstrate the role of the ACR Dose Index Registry® (DIR) in a dose reduction program at a large academic health care system. MethodsUsing the ACR DIR, radiation doses were collected for four common CT examination types (head without contrast, chest with contrast, chest without contrast, and abdomen and pelvis with contrast). Baseline analysis of 7,255 CT examinations from seven scanners across the institution was performed for the period from December 1, 2011, to March 15, 2012. A comprehensive dose reduction initiative was guided by the identification of targets for dose improvement from the baseline analysis. Data for 14,938 examinations from the same seven scanners were analyzed for the postimplementation period of January 1, 2013, to July 1, 2013. ResultsThe program included protocol changes, iterative reconstruction, optimization of scan acquisition, technologist education, and continuous monitoring with feedback tools. Average decrease in median dose-length product (DLP) across scanners was 30% for chest CT without contrast, 29% for noncontrast head CT, 26% for abdominal and pelvic CT with contrast, and 10% for chest CT with contrast. Compared with average median DLP in the ACR DIR, the median institution-wide CT DLPs after implementation were lower by 33% for chest CT without contrast, 32% for chest CT with contrast, 26% for abdominal and pelvic CT with contrast, and 6% for head CT without contrast. ConclusionsA comprehensive CT dose reduction program using the ACR DIR can lead to substantial dose reduction within a large health care system.

Substance use and mental diagnoses among adults with and without type 2 diabetes: Results from electronic health records data

BackgroundComorbid diabetes and substance use diagnoses (SUD) represent a hazardous combination, both in terms of healthcare cost and morbidity. To date, there is limited information about the association of SUD and related mental disorders with type 2 diabetes mellitus (T2DM). MethodsWe examined the associations between T2DM and multiple psychiatric diagnosis categories, with a focus on SUD and related psychiatric comorbidities among adults with T2DM. We analyzed electronic health record (EHR) data on 170,853 unique adults aged ≥18 years from the EHR warehouse of a large academic healthcare system. Logistic regression analyses were conducted to estimate the strength of an association for comorbidities. ResultsOverall, 9% of adults (n=16,243) had T2DM. Blacks, Hispanics, Asians, and Native Americans had greater odds of having T2DM than whites. All 10 psychiatric diagnosis categories were more prevalent among adults with T2DM than among those without T2DM. Prevalent diagnoses among adults with T2MD were mood (21.22%), SUD (17.02%: tobacco 13.25%, alcohol 4.00%, drugs 4.22%), and anxiety diagnoses (13.98%). Among adults with T2DM, SUD was positively associated with mood, anxiety, personality, somatic, and schizophrenia diagnoses. ConclusionsWe examined a large diverse sample of individuals and found clinical evidence of SUD and psychiatric comorbidities among adults with T2DM. These results highlight the need to identify feasible collaborative care models for adults with T2DM and SUD related psychiatric comorbidities, particularly in primary care settings, that will improve behavioral health and reduce health risk.

Factors Associated with Incomplete Bladder Emptying in Older Women with Overactive Bladder Symptoms.

To identify factors associated with incomplete bladder emptying in older women with overactive bladder (OAB) symptoms. Retrospective medical records review. Urodynamic laboratory at a large academic health care system in the southeast United States. Medical records from women aged 65 and older who presented to clinic with OAB symptoms and who had completed outpatient urodynamic testing (N = 203). Three cutoff values of postvoid residual (PVR) volume (100, 150, 200 mL) were created a priori to classify incomplete bladder emptying. Age; vaginal parity; diabetes mellitus (DM); and a history of cerebrovascular accident, back surgery or injury, pelvic surgery, anti-urinary incontinence surgery, and pelvic organ prolapse were included in univariate and multivariate logistic regression models to determine factors associated with incomplete bladder emptying. When PVR volume of 200 mL or greater was used to define incomplete bladder emptying, history of back surgery or injury (odds ratio (OR) = 4.30, 95% confidence interval (CI) = 1.16-15.91) and pelvic surgery (OR = 4.42, 95% CI = 1.51-12.95) were significantly associated with incomplete bladder emptying in multivariate analysis. No factors were significantly associated with PVR volumes of 100 mL or greater or 150 mL or greater. PVR volumes should be assessed in older women with OAB symptoms and a history of back surgery or injury or pelvic surgery, especially before the administration of antimuscarinic medications. Caution in interpreting these findings is needed because PVR volumes obtained from pressure-flow studies may differ from those obtained from natural voiding. Lack of data on medications that can contribute to incomplete bladder emptying and the small number of women with cerebrovascular accident, back surgery or injury, and DM may have influenced the results.

Use of Nondisclosure Agreements in Medical Malpractice Settlements by a Large Academic Health Care System.

Honesty and transparency are essential aspects of health care, including in physicians' and hospitals' responses to medical error. Biases and habits associated with medical malpractice litigation, however, may work at cross-purposes with compassion in clinical care and with efforts to improve patient safety. To determine the frequency of nondisclosure agreements in medical malpractice settlements and the extent to which the restrictions in these agreements seem incompatible with good patient care. We performed a retrospective review of medical malpractice claim files, including settlement agreements, for claims closed before (fiscal year 2001-2002), during (fiscal year 2006-2007), and after (fiscal years 2009-2012) the implementation of tort reform in Texas. We studied The University of Texas System, which self-insures malpractice claims that involve 6000 physicians at 6 medical campuses in 5 cities. Nondisclosure provisions in medical malpractice settlements. During the 5 study years, The University of Texas System closed 715 malpractice claims and made 150 settlement payments. For the 124 cases that met our selection criteria, the median compensation paid by the university was $100,000 (range, $500-$1.25 million), and the mean compensation was $185,372. A total of 110 settlement agreements (88.7%) included nondisclosure provisions. All the nondisclosure clauses prohibited disclosure of the settlement terms and amount, 61 (55.5%) prohibited disclosure that the settlement had been reached, 51 (46.4%) prohibited disclosure of the facts of the claim, 29 (26.4%) prohibited reporting to regulatory agencies, and 10 (9.1%) prohibited disclosure by the settling physicians and hospitals, not only by the claimant. Three agreements (2.7%) included specific language that prohibited the claimant from disparaging the physicians or hospitals. The 50 settlement agreements signed after tort reform took full effect in Texas (2009-2012) had stricter nondisclosure provisions than the 60 signed in earlier years: settlements after tort reform were more likely to prohibit disclosure of the event of settlement (36 [72.0%] vs 25 [41.7%]; P < .001), to prohibit disclosure of the facts of the claims (31 [62.0%] vs 20 [33.3%]; P = .003), and to prohibit reporting to regulatory bodies (25 [50.0%] vs 4 [6.7%]; P < .001). An academic health system with a declared commitment to patient safety and transparency used nondisclosure clauses in most malpractice settlement agreements but with little standardization or consistency. The scope of nondisclosure was often broader than seemed needed to protect physicians and hospitals from disparagement by the plaintiff or to avoid publicizing settlement amounts that might attract other claimants. Some agreements prohibited reporting to regulatory agencies, a practice that the health system changed in response to our findings.

Marked hyperferritinemia does not predict for HLH in the adult population

AbstractHemophagocytic lymphohistiocytosis (HLH) is a rare syndrome of uncontrolled immune activation that has gained increasing attention during the last decade. The diagnosis of HLH is based on a constellation of clinical and laboratory abnormalities, including elevated serum ferritin levels. In the pediatric population, marked hyperferritinemia is specific for HLH. To determine what conditions are associated with profoundly elevated ferritin in the adult population, we performed a retrospective analysis in a large academic health care system. We identified 113 patients with serum ferritin levels higher than 50 000 µg/L. The most frequently observed conditions included renal failure, hepatocellular injury, infections, and hematologic malignancies. Our results suggest that marked hyperferritinemia can be seen in a variety of conditions and is not specific for HLH in adults.

Derivation and Prospective Validation of a Predictive Score for the Rapid Diagnosis of Thrombotic Thrombocytopenic Purpura: The Plasmic Score

Introduction: Thrombotic thrombocytopenic purpura (TTP) is a hematologic emergency that requires the immediate initiation of therapy without access to the results of definitive testing and often in the setting of limited prior clinical experience. In this context, a simple, validated, and rapidly deployable diagnostic scoring system could be a powerful tool.Methods: We studied 214 consecutive adults admitted to 3 major teaching hospitals between 2004 and 2014 with thrombocytopenia (range: 5 to 149,000/μL) and schistocytes suspicious for TTP (derivation cohort). Twenty-nine different laboratory and clinical parameters were evaluated to identify those most highly associated with an ADAMTS13 activity level of ≤10%. Cutoffs for continuous variables were determined by generating receiver operator curves and identifying the cutoff point associated with the maximum Youden Index. Those parameters most associated with a low ADAMTS13 level by univariate analysis (P≤0.1) were included in a logistic regression to create a parsimonious predictive model. The resulting model was validated through a prospectively collected internal cohort (n=40) and an external cohort (n=147) drawn from a different academic medical center.Results: Four individual laboratory parameters (platelet count <30,000/μL, MCV <90fL, creatinine <2.0mg/dL, and INR <1.5), a combined laboratory parameter (evidence for hemolysis based on either a reticulocyte count >2.5%, an indirect bilirubin >2.0 mg/dL, or absent haptoglobin), and two historical features (no active cancer, no history of bone marrow or solid organ transplantation) were predictive of an ADAMTS13 activity of ≤10% in our logistic regression model. Based on similar odds ratio estimates, each factor was assigned a value of 1 point, with scores of 0-4 considered low risk, 5-6 intermediate, and 7 high. The derivation cohort (n=214) contained 62 cases of idiopathic TTP with low ADAMTS13 activity level. When applied against this cohort, the score effectively sorted patients according to ADAMTS13 level (See Table) and demonstrated an AUC of 0.96 (95% CI 0.92-0.98).An internally-collected prospective validation cohort (n=40) contained 6 cases of idiopathic TTP with low ADAMTS13 activity, and the score demonstrated an AUC of 0.96 (95% CI 0.84-1.00). For the second independent validation, an external cohort from a large academic healthcare system in the southeastern United States was used (n=147) that contained 69 cases of idiopathic TTP with low ADAMTS13 activity. When tested against this dataset, cases with scores falling in the low, intermediate, and high risk groups were respectively 4.3%, 56.8%, and 96.2% likely to have an ADAMTS13 activity level ≤10% (see Table). This corresponded to an AUC of 0.90 (95% CI 0.84-0.95). Scores corresponded to certain diagnoses, with 21/27 cases of DIC and 23/31 cases of drug-associated thrombotic microangiopathy (TMA) sorting to the low risk category. Scores were associated with one year survivals of 46.9%, 69.2%, and 96.9% for the low, intermediate, and high risk groups respectively (see Figure, p < 0.0001 by log-rank).Conclusions: A simple, validated clinical scoring system can assist in the early diagnosis of TTP by accurately identifying cases with a high likelihood of having severe ADAMTS13 deficiency. The word PLASMIC can be used to recall the 7 features of the score (platelets, lysis, active cancer, stem cell or solid organ transplant, MCV, INR, and creatinine). This scoring system also has the potential to provide accurate diagnostic and prognostic information for other TMAs. The ability to rapidly diagnose TTP with this approach would facilitate the initiation of appropriate therapy, while in low-risk cases clinicians could pursue alternative diagnoses with a high degree confidence.Table 1Proportion ADAMTS13 ≤10%Derivation (%)ValidationInternal (%)External (%)Risk CategoryLow (0-4)0/84 (0)0/22 (0)2/47 (4.3)Intermediate (5-6)25/77 (32.4)3/15 (20)42/74 (56.8)High (7)37/39 (94.9)3/3 (100)25/26 (96.2) [Display omitted] DisclosuresMakar:Alexion Pharmaceutials: Research Funding.