NorthShore University HealthSystem Research Articles

Reduced Length of Stay Using Clinical Decision Support Tool (ASAP) for Empiric Antibiotic Selection

Background: Empiric antibiotic selection is challenging and requires knowledge of the local antibiogram, national guidelines and patient-specific factors, such as drug allergy and recent antibiotic exposure. Clinical decision support for empiric antibiotic selection has the potential to improve adherence to guidelines and improve patient outcomes. Methods: At NorthShore University HealthSystem, a 4-hospital, 789 bed system, an automated point-of-care decision support tool referred to as Antimicrobial Stewardship Assistance Program (ASAP) was created for empiric antibiotic selection for 4 infectious syndromes: pneumonia, skin and soft-tissue infections, urinary tract infection, and intra-abdominal infection. The tool input data from the electronic health record, which can be modified by any user. Using an algorithm created with electronic health record data, antibiogram data, and national guidelines, the tool produces an antibiotic recommendation that can be ordered via a link to order entry. If the tool identifies a patient with a high likelihood for a multidrug-resistant infection, a consultation by an infectious diseases specialist is recommended. Utilization of the tool and associated outcomes were evaluated from July 2018 to May 2019. Results: The ASAP tool was executed by 140 unique, noninfectious diseases providers 790 times. The tool was utilized most often for pneumonia (194 tool uses), followed by urinary tract infection (166 tool uses). The most common provider type to use the tool was an internal medicine hospitalist. The tool increased adherence to the recommended antibiotic regimen for each condition. Antibiotic appropriateness was assessed by an infectious diseases physician. Antibiotics were considered appropriate when they were similar to the antibiotic regimen recommended by the ASAP. Inappropriate antibiotics were classified as broad or narrow. When antibiotic coverage was appropriate, hospital length of stay was statistically significantly shorter (4.8 days vs 6.8 days for broad antibiotics vs 7.4 days for narrow antibiotics; P < .01). No significant differences were identified in mortality or readmission. Conclusions: A clinical decision support tool in the electronic health record can improve adherence to recommended empiric antibiotic therapy. Use of appropriate antibiotics recommended by such a tool can reduce hospital length of stay.Funding: NoneDisclosures: None

Reducing Antibiotic Prescribing in Primary Care for Respiratory Illness.

One-third of outpatient antibiotic prescriptions for pediatric acute respiratory tract infections (ARTIs) are inappropriate. We evaluated a distance learning program's effectiveness for reducing outpatient antibiotic prescribing for ARTI visits. In this stepped-wedge clinical trial run from November 2015 to June 2018, we randomly assigned 19 pediatric practices belonging to the Pediatric Research in Office Settings Network or the NorthShore University HealthSystem to 4 wedges. Visits for acute otitis media, bronchitis, pharyngitis, sinusitis, and upper respiratory infection for children 6 months to <11 years old without recent antibiotic use were included. Clinicians received the intervention as 3 program modules containing online tutorials and webinars on evidence-based communication strategies and antibioti c prescribing, booster video vignettes, and individualized antibiotic prescribing feedback reports over 11 months. The primary outcome was overall antibiotic prescribing rates for all ARTI visits. Mixed-effects logistic regression compared prescribing rates during each program module and a postintervention period to a baseline control period. Odds ratios were converted to adjusted rate ratios (aRRs) for interpretability. Among 72 723 ARTI visits by 29 762 patients, intention-to-treat analyses revealed a 7% decrease in the probability of antibiotic prescribing for ARTI overall between the baseline and postintervention periods (aRR 0.93; 95% confidence interval [CI], 0.90-0.96). Second-line antibiotic prescribing decreased for streptococcal pharyngitis (aRR 0.66; 95% CI, 0.50-0.87) and sinusitis (aRR 0.59; 95% CI, 0.44-0.77) but not for acute otitis media (aRR 0.93; 95% CI, 0.83-1.03). Any antibiotic prescribing decreased for viral ARTIs (aRR 0.60; 95% CI, 0.51-0.70). This program reduced antibiotic prescribing during outpatient ARTI visits; broader dissemination may be beneficial.

Clinical Analytics Prediction Engine (CAPE): Development, electronic health record integration and prospective validation of hospital mortality, 180-day mortality and 30-day readmission risk prediction models.

BackgroundNumerous predictive models in the literature stratify patients by risk of mortality and readmission. Few prediction models have been developed to optimize impact while sustaining sufficient performance.ObjectiveWe aimed to derive models for hospital mortality, 180-day mortality and 30-day readmission, implement these models within our electronic health record and prospectively validate these models for use across an entire health system.Materials & methodsWe developed, integrated into our electronic health record and prospectively validated three predictive models using logistic regression from data collected from patients 18 to 99 years old who had an inpatient or observation admission at NorthShore University HealthSystem, a four-hospital integrated system in the United States, from January 2012 to September 2018. We analyzed the area under the receiver operating characteristic curve (AUC) for model performance.ResultsModels were derived and validated at three time points: retrospective, prospective at discharge, and prospective at 4 hours after presentation. AUCs of hospital mortality were 0.91, 0.89 and 0.77, respectively. AUCs for 30-day readmission were 0.71, 0.71 and 0.69, respectively. 180-day mortality models were only retrospectively validated with an AUC of 0.85.DiscussionWe were able to retain good model performance while optimizing potential model impact by also valuing model derivation efficiency, usability, sensitivity, generalizability and ability to prescribe timely interventions to reduce underlying risk. Measuring model impact by tying prediction models to interventions that are then rapidly tested will establish a path for meaningful clinical improvement and implementation.

Internal and External Validation of a Machine Learning Risk Score for Acute Kidney Injury

Acute kidney injury (AKI) is associated with increased morbidity and mortality in hospitalized patients. Current methods to identify patients at high risk of AKI are limited, and few prediction models have been externally validated. To internally and externally validate a machine learning risk score to detect AKI in hospitalized patients. This diagnostic study included 495 971 adult hospital admissions at the University of Chicago (UC) from 2008 to 2016 (n = 48 463), at Loyola University Medical Center (LUMC) from 2007 to 2017 (n = 200 613), and at NorthShore University Health System (NUS) from 2006 to 2016 (n = 246 895) with serum creatinine (SCr) measurements. Patients with an SCr concentration at admission greater than 3.0 mg/dL, with a prior diagnostic code for chronic kidney disease stage 4 or higher, or who received kidney replacement therapy within 48 hours of admission were excluded. A simplified version of a previously published gradient boosted machine AKI prediction algorithm was used; it was validated internally among patients at UC and externally among patients at NUS and LUMC. Prediction of Kidney Disease Improving Global Outcomes SCr-defined stage 2 AKI within a 48-hour interval was the primary outcome. Discrimination was assessed by the area under the receiver operating characteristic curve (AUC). The study included 495 971 adult admissions (mean [SD] age, 63 [18] years; 87 689 [17.7%] African American; and 266 866 [53.8%] women) across 3 health systems. The development of stage 2 or higher AKI occurred in 15 664 of 48 463 patients (3.4%) in the UC cohort, 5711 of 200 613 (2.8%) in the LUMC cohort, and 3499 of 246 895 (1.4%) in the NUS cohort. In the UC cohort, 332 patients (0.7%) required kidney replacement therapy compared with 672 patients (0.3%) in the LUMC cohort and 440 patients (0.2%) in the NUS cohort. The AUCs for predicting at least stage 2 AKI in the next 48 hours were 0.86 (95% CI, 0.86-0.86) in the UC cohort, 0.85 (95% CI, 0.84-0.85) in the LUMC cohort, and 0.86 (95% CI, 0.86-0.86) in the NUS cohort. The AUCs for receipt of kidney replacement therapy within 48 hours were 0.96 (95% CI, 0.96-0.96) in the UC cohort, 0.95 (95% CI, 0.94-0.95) in the LUMC cohort, and 0.95 (95% CI, 0.94-0.95) in the NUS cohort. In time-to-event analysis, a probability cutoff of at least 0.057 predicted the onset of stage 2 AKI a median (IQR) of 27 (6.5-93) hours before the eventual doubling in SCr concentrations in the UC cohort, 34.5 (19-85) hours in the NUS cohort, and 39 (19-108) hours in the LUMC cohort. In this study, the machine learning algorithm demonstrated excellent discrimination in both internal and external validation, supporting its generalizability and potential as a clinical decision support tool to improve AKI detection and outcomes.

Economic Shocks From the Novel COVID-19 Pandemic for Anesthesiologists and Their Practices.

Economic Shocks From the Novel COVID-19 Pandemic for Anesthesiologists and Their Practices.

Helping Families Navigate Center Variability in Antenatal Counseling for Extremely Early Births.

* Abbreviation: AAP — : American Academy of Pediatrics ### Dr Dalia Feltman (Neonatologist) Institutional variability in rates of active treatment (delivery room resuscitation) affects survival rates of newborns born at 22 and 23 weeks’ gestation.1 Neonatology program directors responding to a 2015 survey reported one-quarter of their centers’ approaches to active treatment at 22 weeks’ gestation varied by individual neonatologists.2 Although deliveries at these ages are relatively uncommon, this variability raises questions ethicists, institutions, and parents are asking in academic meetings, in conference rooms, and on social media. The latest American Academy of Pediatrics (AAP) Committee on Fetus and Newborn guidance encourages institutions to develop counseling approaches for women facing delivery before 25 completed weeks’ gestation.3 More than half of the aforementioned survey respondents reported management changes occurring in the last 5 years for these earliest newborns.2 Most AAP Section on Neonatal Perinatal Medicine members responding to a 2017 survey desired guidance on options to offer parents facing extremely early delivery but did not agree at which level (eg, institutional, regional, or national) this guidance should occur.4 I asked 2 couples (unrelated to my institution) whose children were born extremely early to share their delivery stories and, using questions grounded by the framework of the AAP clinical report,3 offer advice to clinicians developing care approaches for families like theirs. ### Two Families’ Experiences Surrounding Extremely Early Delivery #### Carrie and Troy Tysdahl When we found out my water broke at only 22 weeks’ gestation, we were given only … Address correspondence to Dalia M. Feltman, MD, MA, NorthShore University HealthSystem, Evanston Hospital, Walgreen Building, Suite 1505, 2650 Ridge Ave, Evanston, IL 60201. E-mail: dfeltman{at}northshore.org

1158. Discontinuation of Urine Cultures by Infection Preventionists in Hospitalized Patients with Indwelling Urinary Catheters: Is It Safe?

BackgroundA majority of healthcare-associated urinary tract infections (UTIs) are caused by the use of urinary catheters (CAUTI). Finding of bacteriuria is common in catheterized patients and often leads to unnecessary antibiotic treatment, increased length of stay and additional healthcare costs. We implemented an innovative intervention to improve urine culture (UCx) orders and prevent overdiagnosis of CAUTIs.MethodsOrders for UCx in adult patients with short-term urinary catheters at NorthShore University HealthSystem, IL were reviewed daily for appropriateness based on the Infectious Diseases Society of America Guidelines. Appropriate urine testing was defined as: (1) presence of fever ( >38°C) within past 48 hours, (2) new urinary complaints: flank or suprapubic pain/tenderness or dysuria, frequency, urgency or incontinence within 48 hours after catheter removal, and (3) no other reasonable explanation for fever. If UCx was deemed inappropriate, ordering provider was contacted to cancel the order. Chart review was performed at least 30-days post-discharge to determine whether patients developed recurrent UTI, sepsis, were readmitted or expired.ResultsBetween 1 January to 31 March 2019, 65 UCx were submitted. Sixty-four patients (98%) did not meet criteria for testing. Most common reasons for not meeting criteria were absence of fever (60%) and no localizing UTI signs or symptoms (57%). 35 (54%) UCx were canceled after discussion with ordering providers. 21/35 patients (60%) were treated with antibiotics. All 35 patients were discharged, with a majority going to a skilled nursing facility (34%) or home (31%). 4/35 (11%) had a subsequent positive UCx. Two patients developed symptomatic UTI (sUTI) during the index admission. Two patients developed sUTI within 30-days post-discharge; one patient was transitioned to hospice after completion of therapy. All 4 patients were treated for sUTI.ConclusionWe were able to safely discontinue UCx in 89% of patients. A majority of patients were already started on empiric treatment and development of subsequent sUTI was infrequent (11% of patients). Our findings suggest that discontinuation of inappropriately ordered UCx is safe with low risk for sepsis or mortality.Disclosures All authors: No reported disclosures.

Increases in Rates of Suicide and Suicide Attempts Among Black Adolescents.

Over the last 10 years, the rate of adolescents ages 15 to 19 committing suicide in the United States increased drastically, with an overall rate of 6.7 suicides per 100 000 per year in 2007 and a rate of 11.8 per 100 000 per year in 2017.1 The rate increased for both boys and girls as well as among white, black, and Hispanic adolescents (Table 1). The group with the most dramatic increase was black girls: 1.2 suicides per 100 000 per year in 2007 and 4.0 per 100 000 per year in 2017. View this table: TABLE 1 Rates of Suicide by Sex and Race Among Adolescents Ages 15–19 in 2007 and 20171 Suicide, nonetheless, is an infrequent event. Identifying adolescents at increased danger is aided by the use of suicide risk factors, such as depression and bipolar disorder; substance intoxication … Address correspondence to Benjamin N. Shain, MD, PhD, NorthShore University HealthSystem, 49 S Waukegan Rd, Deerfield, IL 60015. E-mail: aias{at}ix.netcom.com

1222 Co-Prescription of Ondansetron With Other QT-Prolonging Drugs: Data From a Large Community-Based Hospital System

INTRODUCTION: The Food and Drug Administration (FDA) has warned about the co-prescription of ondansetron with other QT-prolonging medications. Electronic medical records (EMR) with built-in warning systems for drug-drug interactions (DDIs) are designed to improve patient safety. However, overriding of warnings are common. Response to warnings for potentially life-threatening DDIs such as QT-prolongation has not been fully investigated. We studied outpatient co-prescription of ondansetron with other QT-prolonging medications in a large community-based health practice. METHODS: Data for all outpatient prescriptions for oral ondansetron from January 1 to March 31st, 2016 at NorthShore University HealthSystem was reviewed. A de-identified dataset was developed that included all outpatient prescriptions for oral ondansetron in patients receiving another QT-prolonging drug. Duplicate orders were removed. Interaction warnings, including the severity of the interactions are generated at the time of order placement. Prescriber response to DDI warnings and further analysis of the co-prescribed QT-prolonging drugs was performed. Demographic information was also studied. RESULTS: 11,935 cases of inpatient and outpatient co-prescriptions of ondansetron with and one or more QT-prolonging drug occurred during the study period. After filtering, there were 884 outpatient ondansetron and QT-prolonging co-prescription orders in 723 patients. All warnings were “severe interactions”. Only 93 (10.5%) prescriptions in outpatients were canceled, while 791 (89.5%) of the prescription orders were overridden. The most common classes of QT-prolonging medications taken by patients with overridden orders for ondansetron were selective serotonin reuptake inhibitors (SSRIs) (33%), macrolide antibiotics (29%), fluoroquinolone antibiotics (13%), and serotonin antagonist reuptake inhibitors (SARIs) (11%) (See Table 1). The most commonly co-prescribed QT prolonging drugs are listed in Table 2. Twelve patients receiving QT-prolonging antiarrhythmic agents had completed orders for ondansetron. 617 (78%) of overridden prescriptions occurred in women. CONCLUSION: Warnings for ondansetron co-prescription with other QT-prolonging drugs are usually ignored. Co-prescription of ondansetron with other QT-prolonging drugs is more common in female patients. This is of particular concern because women are more likely to develop serious arrhythmias related to ondansetron-induced QT-prolongation.

221 Risk of Bone Density Loss in Microscopic Colitis

INTRODUCTION: Microscopic colitis (MC) is an inflammatory bowel disease of the colon that presents with chronic watery diarrhea. These patients have many risk factors for osteoporosis including female predominance, mean age in the 60s, lower body-mass index (BMI) and active smoking. Additionally the mainstay of therapy is a corticosteroid, budesonide. Bone density has not been rigorously studied in the population. Our aim was to evaluate for bone mineral density in patients with a diagnosis of MC with and without the use of corticosteroids. METHODS: We queried our prospective MC patient registry at NorthShore University HealthSystem, a clinical database of biopsy-confirmed patients with MC. Patients were included if they had a dual energy X-ray absorptiometry (DEXA) scan. DEXA measurements of the radius, femoral neck, and lumbar spine were used to evaluate bone mass. Baseline clinical factors were also collected, particularly tobacco use, BMI, and treatments for MC including corticosteroids. In this case-control study, age and gender-matched controls without diarrhea served as the control population. RESULTS: Of the 192 subjects in our registry, only 94 patients had undergone or had available bone density evaluations. The mean age at diagnosis was 69 years (range 42–91), with 91% female. No significant difference in rates of osteoporosis or osteopenia was observed in patients with MC (OR 1.58 [95% 0.75–3.30, P = 0.22]). Bone loss, defined as osteopenia or osteoporosis, was present in 82% of MC patients, compared to 75% in the controls. 76% of patients with MC received prolonged treatment with budesonide, and 24% received no steroids. A significant difference in bone mass was not appreciated between MC patients treated with a steroid, compared to controls (OR 1.15 [95% 0.33–4.03, P &lt; 0.83]). CONCLUSION: Patients with MC have increased rates of osteopenia or osteoporosis compared to controls. However, our control cohort had a significantly higher degree of bone loss than current US estimates of 45–55%, which likely affected our statistical significance. Thus, bone loss is an underappreciated but important problem in microscopic colitis. With increased awareness, DEXA scans should be performed routinely in MC patients, particularly in those who have known risk factors for osteoporosis.

1687-P: An Examination of Clinical Outcomes in People with T1D and Obesity

A lack of knowledge exists concerning how best to treat type 1 diabetes (T1D) and obesity simultaneously. This cross-sectional study examined whether people with T1D and obesity have differences in the prevalence of outcomes, treatment methods, and healthcare utilization. The data were pulled from the electronic medical record for 1546 patients with T1D (≥ 18 years) seen at NorthShore University HealthSystem during 2017. The study population was 53% female; mean age was 46.4 ± 18.0 years. About 32% of patients had a normal weight (BMI 20-24.99) and 28% of patients had obesity (BMI ≥ 30). The patients with obesity were compared to the patients with normal weight (NW).There was a significantly higher prevalence of hypertension (p&amp;lt;0.001) and depression (p=0.038) among patients with obesity compared to patients with NW. No significant difference in the prevalence of other comorbidities (heart failure, myocardial infarction, cancer, dementia, peripheral vascular disease, chronic kidney disease, stroke) were observed. Patients with obesity had a significantly higher rate of retinopathy, neuropathy, and microalbuminuria (p&amp;lt;0.0001, p=0.005, p=0.005, respectively) compared to patients with NW. Adjusted for age, sex, and years of T1D duration, patients with obesity were 57%, 92%, and 60% more likely to have neuropathy, retinopathy, and microalbuminuria, respectively, than NW patients. There were no difference in mean HbA1C or total cholesterol level between the groups. Adjusted for age, sex, and years of T1D duration, the use of an insulin pump was not associated with BMI. Patients with obesity attended more outpatient visits in 2017 than NW patients (p=0.001), but did not differ in ER or inpatient visits. The higher rate of diabetes-related outcomes, hypertension, depression and healthcare utilization among T1D patients with obesity, compared to those of NW provide the impetus for further investigation into how best to treat patients with T1D and obesity to prevent the disparities in outcomes elucidated by this study. Disclosure E.R. Stewart: None. A.M. Monie: None. L.K. Billings: Advisory Panel; Self; Novo Nordisk A/S, Sanofi-Aventis. Consultant; Self; Novo Nordisk A/S. Research Support; Self; Novo Nordisk A/S. Speaker's Bureau; Self; Novo Nordisk A/S.

2394. Different Clostridioides difficile Ribotypes Among Patients With Colonization, Initial Clinical Disease, and Recurrent Clinical Disease

Background C. difficile is the most common cause of hospital infections with a spectrum of presentation from asymptomatic carriage to severe recurrent diarrhea. Certain C. difficile ribotypes are associated with severe disease, but there are little data on ribotypes in asymptomatic carriers or severe recurrent disease. The aim of this study was to compare virulence potential of C.diff ribotypes with clinical disease severity.MethodsThis retrospective study included patients aged ≥18 years at NorthShore University HealthSystem (NUHS) from February 1, 2015 to May 30, 2017. Three groups of patients with positive PCR test for C. diff toxin gene were selected: (1) Asymptomatic patients positive for rectal carriage; (2) symptomatic outpatients with a single positive test (CDI); and (3) patients with recurrent CDI who underwent FMT. Clinical data were extracted from the Enterprise Database Warehouse. Isolates underwent fluorescent PCR ribotyping and were assigned to clades. Ribotypes with “high” (e.g., 027 and 078) and “low” (e.g., 106) virulence potential were defined as such. Virulence potential of cryptic ribotypes were considered “unknown.” We used Χ2 and independent samples median tests to compare categorical and continuous variables, respectively.Results129 C. diff isolates (asymptomatic, N = 66; CDI, N = 33; FMT, N = 30) were ribotyped with 60 types identified. Median age was higher in asymptomatic patients [80.5 (IQR 70.8–90) years] compared with both CDI and FMT [69 (58–81) and 69 (51–83.5) years, respectively, P = 0.004] Low virulence ribotypes were identified more frequently in asymptomatic carriers than those with CDI or FMT (22/66 vs. 8/33 vs. 1/30, respectively, P = 0.006). High virulence ribotypes were found in all groups, with highest frequency in the FMT group (23/30) vs. asymptomatic (25/67) or CDI (13/33), P = 0.001).ConclusionPatients with severe or recurrent CDI had ribotypes associated with high virulence potential. In addition, asymptomatic carriers were more likely to have ribotypes of C.diff historically associated with a low virulence potential. Molecular C.diff typing may have a role in evaluating asymptomatic C.diff colonization vs. clinical disease.Disclosures All authors: No reported disclosures.

972. Asymptomatic Carriage of Clostridiodes difficile and Risk of Subsequent Infection

Background C. difficile is one of the most common healthcare-associated infections in the United States. Studies of patients with asymptomatic carriage of toxigenic C. difficile have reported conflicting results on the risk of subsequent C. difficile infection (CDI). Older studies suggest that the risk was low and colonization may be protective. Subsequent studies indicate that asymptomatic carriers have a 6-fold greater risk of developing CDI. The aims of our study were to assess the burden of asymptomatic C. difficile carriage and risk of subsequent CDI.MethodsAdult inpatients at NorthShore University HealthSystem, Illinois hospitals between August 1, 2017 and February 28, 2018 were eligible for the study. Focused admission screening of patients at high risk of C. difficile carriage was performed: (1) history of CDI or colonization, (2) prior hospitalization past 2 months, or (3) admission from a long-term care facility. A rectal swab was collected and tested using the cobas® Cdif Test (Roche) real-time PCR. The development of hospital onset CDI (HO-CDI) in colonized patients was monitored prospectively for at least 2 months. HO-CDI testing of colonized patients was performed using the Cepheid GeneXpert RT-PCR. HO-CDI was defined as patients hospitalized for at least 72 hours with 3 or more episodes of diarrhea/24 hours, in the absence of other potential causes of diarrhea. Patient demographics were collected using a standardized form and data analyzed using VassarStats.ResultsThere were 6,104 patients enrolled in the study and 528 (8.7%) were positive on admission for toxigenic C. difficile carriage. The mean age of colonized patients was 75.5 years (range 24–103) and 56.4% (298 patients) were females. Of 528 colonized patients, 21 (4%) had a positive CDI test. A total of 7 patients (1.3%) developed HO-CDI. Mean time to positive HO-CDI was 46.1 days (range 5–120 days). Of 5,576 patients that were negative for C difficile carriage on admission, 14 (0.3%) patients developed HO-CDI. The relative risk of HO-CDI was 5.28 (95% CI: 2.14–13.03, P = 0.05).ConclusionWe found that 8.7% of at-risk admissions were asymptomatic toxigenic C. difficile carriers. While only 1.3% developed HO-CDI, asymptomatic carriers had a 5 times higher risk of subsequent CDI compared with non-carriers.Disclosures All authors: No reported disclosures.

2178. Detection of Key Potential Healthcare Pathogens Using Periodic Point Prevalence Surveillance

BackgroundSurveillance for asymptomatic carriage of multidrug-resistant (MDR) pathogens is useful to determine the burden of these organisms and help guide infection prevention strategy. We currently perform surveillance cultures for Gram-negative multidrug-resistant pathogens (GNMDR) in the ICUs on a monthly basis. We added a quarterly point prevalence survey to all hospital units for these and other key pathogens over one year to determine whether our program should expand beyond the ICU and include other organisms.MethodsRectal samples were collected quarterly for 1 year starting June 2016 at NorthShore University HealthSystem, a four-hospital, 789 bed system. All hospitalized patients present on the day of the point prevalence testing had a double-headed rectal swab collected. One swab was plated to VACC agar (Remel) for culture of GNMDR and VRE, and the second was plated to CCFAHT (Anaerobe Systems) for C. difficile (Cdif) culture. All samples were collected on a specified day at each of our 4 hospitals, one hospital per week, and sent to the central microbiology lab for processing. Testing for GNMDR included the following pathogens: Carbapenem-resistant Enterobacteriaceae (CRE), ESBLs, and Gram negative organisms susceptible to ≤2 drug classes.ResultsA total of 987 surveillance samples were collected. The number of patients with MDR in the ICU vs. non-ICU units is described in Table 1. There was an 11% greater difference in the percentage of patients colonized with GNMDR and Cdif in non-ICU patients compared with ICU patients (P = 0.006). An important discovery was three patients colonized with CRE outside the ICU that were previously unknown. The burden of ESBL, VRE and Cdif carriage was also greater outside the ICU.Table 1.Comparison of Patients in ICU vs. Non-ICU with Important Hospital PathogensNumber of Patients With:No. TestsImportant Pathogens (%)ESBLCREMDRVREToxigenic CdifNon-ICU833175 (21%)79356447ICU15417 (11%)102117ConclusionThe point prevalence surveillance uncovered a significant amount of MDRs in our non-ICU units, particularly three CREs that were previously unknown. These results suggest there is a large burden of MDR organisms outside the ICU.Disclosures All authors: No reported disclosures.

Post-Infectious Irritable Bowel Syndrome Persists After Fecal Microbiota Transplant for C. difficile Infection

Introduction: Post-infectious irritable bowel syndrome (IBS-PI) is an increasingly recognized subset of IBS where symptoms arise following an acute gastrointestinal infectious illness. After C. difficile infection (CDI), IBS-PI symptoms can mimic recurrent infection. While fecal microbiota transplant (FMT) has been shown to be effective in reducing recurrence of CDI, reports on FMT’s impact on IBS have been mixed. Little has been reported on FMT’s effects on IBS-PI after CDI. This study aims to describe the experience of IBS-PI patients at our center who have undergone FMT. Methods: All patients at NorthShore University Health System receiving FMT for recurrent CDI were prospectively enrolled in our FMT registry. Demographic and clinical information, including pre-existing gastrointestinal disease, were reviewed. Patients were prospectively followed and queried on their clinical symptoms in the office or by phone at baseline, 24 hours, 7 days, 4 weeks and 6 months after FMT. Results: Between May 2015 and November 2017, 169 unique patients underwent FMT for recurrent or severe CDI (Table 1). The mean age of subjects was 67, and 123 (73%) were female. In total, 79 (46.7%) carried a diagnosis of IBS. Of those with IBS, 47 (27.8%) had documented clinical suspicion of IBS-PI by their GI physician, while 32 (18.9%) had pre-existing IBS. 35/47 (74.4%) of IBS-PI patients had symptoms consistent with IBS-D, 3 (6.4%) IBS-C, 6 (12.7%) IBS-M, and 3 (6.4%) unknown. FMT was effective in preventing CDI relapse in 142/169 (84%) of all patients, and 40/47 (85.1%) patients with IBS-PI. The questionnaire response rate for eligible patients with IBS-PI was 39/47 (83%) at 4 weeks after FMT, and 34/47 (72.3%) at 6 months (Table 2). At 4 weeks, 25 (64.1%) reported semi-formed, mushy/soft or loose stools, and 8 (20.5%) reported abdominal pain. At 6 months, 22 (64.7%) reported diarrhea, while 13 (38.2%) reported abdominal pain. 17 (39.5%) were treated with probiotics, 9 (20.9%) on dicyclomine, 7 (16.2%) on loperamide, and 2 (4.65%) on laxatives. Conclusion: IBS-PI is common in patients presenting for FMT for recurrent CDI. Response to FMT is similar amongst IBS-PI and non-IBS patients. However, almost 2/3 of IBS-PI patients continue to experience diarrheal symptoms up to 6 months after FMT, suggesting the limited role of FMT in IBS-PI. These results are relevant when setting expectations regarding symptom relief with the procedure. Further studies regarding IBS-PI management after recurrent CDI are needed.166_A Figure 1 No Caption available.166_B Figure 2 No Caption available.

Overcoming suvivorship care planning implementation challenges through decentralization.

38 Background: To meet Commission on Cancer accreditation requirements, cancer programs must implement processes to monitor the dissemination of survivorship care plans (SCP) for patients with Stages I-III cancers who were treated with curative intent and completed active therapy. Challenges of SCP delivery across disease sites include lack of designated/trained staff, time burden, knowledge of current evidence-based guidelines, and sustainability. We describe the challenges NorthShore University HealthSystem Kellogg Cancer Center (NKCC) and their Living in the Future (LIFE) Cancer Survivorship Program faced in meeting this standard and how evolving the SCP delivery process has resulted in a sustainable model. Methods: LIFE implemented a technology-based SCP tool using a centralized consultative model led by a nurse practitioner (NP) with specialized survivorship training. Physicians referred eligible patients to the survivorship clinic for an education visit where they received a SCP from the NP. Since the centralized model was dependent on one person for delivery, a more sustainable model was needed. NKCC transitioned to a decentralized process, moving SCP creation and delivery responsibility to all oncology care providers (OCPs). Although not all OCPs had specialized survivorship training, care quality was supported by automated SCP creation based on evidence-based care recommendations embedded in the technology. Results: To date, 143 evidence-based SCPs have been delivered since tool implementation in April 2017. During the centralized model (April 25– June 30) 67 SCPs were created by the lead LIFE NP; 76 were created during the decentralized process while the lead NP was on leave (July 3 – Oct 1). By using a technology-based SCP, OCPs incorporated SCP delivery into their workflow and no longer had to refer patients to a separate clinic. Conclusions: This project demonstrates the feasibility of a sustainable, decentralized process using a technology-based SCP as an option for augmenting centralized SCP delivery. A comparable number of patients received a SCP during both processes with an equivalent number of SCPs being delivered via the decentralized model by OCPs supported by evidence-based technology.

Use of an Electronic Medical Record to Track Adherence to the Mediterranean Diet in a US Neurology Clinical Practice.

ObjectiveWe describe our experience with routinely capturing and analyzing Mediterranean diet data via structured clinical documentation support tools built into the electronic medical record and describe adherence to the Mediterranean diet in patients at risk for either stroke or dementia in a US neurology clinical practice.Patients and MethodsThe Mediterranean diet is associated with a reduced risk of stroke and dementia. The Department of Neurology at NorthShore University HealthSystem routinely evaluates patients at initial and annual outpatient visits using structured clinical documentation support (SCDS) tools built into the electronic medical record (EMR). For patient evaluations in our Vascular Neurology and Brain Health subspecialty clinics, SCDS tools in the EMR include the validated 14-item questionnaire for Mediterranean diet adherence (PREvención con DIeta MEDiterránea [PREDIMED]) that autoscores, auto-interprets, writes to the progress note, and electronically captures data. Our study population includes patients seen at these clinics from July 1, 2015, through November 29, 2017.ResultsAt their initial office visit, 25.5% (95/373) of Brain Health patients scored 10 or more points (“strongly adherent”) on the PREDIMED (median, 8; range, 0-14) whereas 6.7% (55/829) of Vascular Neurology patients achieved a score of 10 or more points (median, 6; range, 0-12). By contrast, 34.7% (2586/7447) of individuals in the original PREDIMED cohort were strongly adherent to the Mediterranean diet.ConclusionPREDIMED scores can be electronically captured to tailor nutrition interventions by assessing baseline adherence at the time of their initial neurology clinic visit. Patients in our Midwestern US clinics were weakly adherent to the Mediterranean diet. This suggests a major opportunity for nutrition intervention and education in US neurology clinical practices, toward preserving and improving brain health.

Outpatient Co-prescription of Ondansetron With Other QT-Prolonging Drugs in a Large Community-Based Health System: A Warning Ignored

Introduction: Ondansetron is a 5HT-3 serotonin receptor antagonist that is commonly used to treat nausea and vomiting. The FDA has warned of QT-prolonging effects of ondansetron, especially when combined with other QT-prolonging drugs, or in patients with known cardiac disease. Electronic medical records with warning systems for drug-drug interactions have been promoted for patient safety. We present a review of de-identified data from a large community-based health system to determine how providers responded to electronic warnings when prescribing ondansetron combined with other QT-prolonging drugs. Methods: All prescriptions for ondansetron in combination with QT-prolonging drugs at NorthShore University HealthSystem between January and March 2016 were included in the initial database as part of a quality improvement project. Inpatient orders and IV ondansetron orders were removed. The name and class of co-prescribed QT-prolonging drugs, warning severity, response to the warning by prescribing practitioner (over-ride warning and prescribe, or cancel order), reason for the warning over-ride, patient age and gender were evaluated. Results: There were a total of 9203 episodes of co-prescription of ondansetron and another QT-prolonging drug in 2739 adult patients during the study time period. After filtering, there were 770 outpatient co-prescriptions in 500 patients. Mean patient age was 46.2 years (range 18 to 90 years). Orders were for 592 females (76.8%) and 178 males (33.2%). All warnings were considered “major interactions”. Most orders, 651 (84.5%) prescriptions were administered, while only 119 (15.5%) orders were canceled. The most common classes of co-prescribed QT-prolonging drugs were macrolide antibiotics (40.0%), flouroquinolone antibiotics (17.0%), SARI antidepressants (13.4%), and SSRI antidepressants (9.9%), (Table 1). There were nine (1.4%) cases of co-prescriptions with antiarrhythmic drugs (amioderone and flecainide). The most commonly co-prescribed QT-prolonging medications were azithromycin, levofloxacin, and trazadone (Table 2). Reasons for warning overrides were “benefit outweighs risk” in 266 (40.8%) and “not clinically significant” in 116 cases (17.7%).Table: Table. Classes of Drugs Co-Prescribed with OndansentronTable: Table. List of Drugs Co-Prescribed with OndansetronConclusion: After viewing warnings, practitioners still prescribed ondansetron with other QT-prolonging drugs 85% of the time. This form of co-prescription with ondansetron neglects FDA recommendations and could put patients at risk for cardiac arrhythmias.

Less Is More: Surgical Procedure Time and Risk of Infections, Length-of-Stay, and Readmission across Three Distinct Surgeries

Background Longer surgical total procedure times (TPT) have been associated with increased postoperative complications. It is unclear what the effect of TPT is on length-of-stay (LOS) or 30-day readmission rate (RAR).Methods We performed a retrospective study of patients undergoing knee arthroplasty (KA), colectomy, and craniectomy at NorthShore University HealthSystem from 1/2007 to 12/2013. Clinical data were extracted from the Data Warehouse and charts were reviewed. We standardized surgery times for each procedure and categorized into two groups: times <75‰ (short procedures) vs. times >75‰ (long procedures). We used χ 2 and t-test to compare categorical and continuous variables. We performed multivariate logistic regression for predictors of surgical site infection (SSI).ResultsIn univariate analyses, long procedures were associated with higher incidence of fevers, SSI, longer LOS, and 30-day RAR (Table 1). TPT was not associated with other postoperative complications. TPT remained an independent predictor of SSI in multivariate (MV) analysis (Table 2).ConclusionHigh TPT was associated with increased SSI, LOS, and 30-day RAR. Understanding variation in TPT may help decrease SSI and healthcare utilization.Table 1:Variables Stratified by (TPT) Characteristic N (%) Short N = 5920 Long N = 1980 P-value Fever ≥ 100.4 1042 (17.6%)450 (22.7%)* Body mass index ≥ 35 1143 (19.7%)519 (26.6%)* Female 3867 (65.3%)1117 (56.4%)* Surgery KA 4788 (80.9%)1604 (81.0%)0.99 Colectomy 467 (7.9%)155 (7.8%) Craniectomy 665 (11.2%)221 (11.2%) Complication 420 (7.1%)154 (7.8%)0.31 Urinary tract infection 76 (1.3%)26 (1.3%)0.92 Pneumonia 75 (1.3%)28 (1.4%)0.62 SSI 67 (1.1%)41 (2.1%)* Venous thromboembolism 174 (2.9%)61 (3.1%)0.75 SSI w/in 30 day of discharge 84 (1.4%)48 (2.4%)* Mean LOS, days (SD) 3.69 (2.89)4.05 (3.38)* 30d RAR 277 (4.7%)142 (7.2%)**P-value < 0.01Table 2:MV logistic regression€ of SSI within 30-day of discharge OR (95% CI) P-value Age 0.99 (0.98–1.01)0.97 Female 1.21 (0.84–1.75)0.31 Surgery KA Ref. * Colectomy 70.2 (36.6–134.6) Craniectomy 20.3 (10.0–41.0) TPT > 75 th %ile 1.59 (1.08–2.33)0.02 Fever > 100.4 4.10 (2.85–5.94)* Diabetes 1.92 (1.27–2.91)* €Including only variables with P-value <0.2 in UV analysis. * P-value <0.01.Disclosures All authors: No reported disclosures.

Using Group-Based Trajectory Temperature Modeling to Predict Postoperative Infections after Total Knee Arthroplasty

BackgroundFever is common in the postoperative setting and frequently physiologic. Despite this, roughly half of febrile patients undergo testing for infectious complications, of which only a few reveal infection. We analyzed whether temperature trajectories could help optimize postoperative (post-op) risk assessment in total knee arthroplasty (TKA) patients.MethodsWe included adult patients who underwent primary TKA between January 1, 2007–December 31, 2013 within NorthShore University HealthSystem. Patients were excluded if infection was suspected before/during surgery. Patient data were extracted from the Database Warehouse. A physician verified post-op complications by chart review. We performed group-based trajectory modeling (GBTM) with covariates: age, BMI, gender, co-morbid conditions and procedure time (STATA). We compared complications per group by χ2 test and evaluated associations with any post-op complication by multivariable (MV) logistic regression (SPSS).ResultsWe identified 5495 independent patients, following three distinct temperature trajectories (Figure 1) – low (group 1), medium (group 2), high (group 3). Noninfectious complications were more likely than infectious complications, and complications were 5x more common in group 3 vs. group 1 (Table 1). In MV logistic regression, membership in group 3 was independently associated with developing a post-op complication, adjusting for age, presence of renal failure and presence of a cardiac arrhythmia (OR 4.4, 95% CI 3.2–6.0, P < 0.01).ConclusionGBTM may help identify TKA patients at increased risk of a post-op complication in real-time, thus helping clinicians avoid unnecessary testing and antibiotics in the post-op setting.Figure 1Results of GBTM showing mean maximum temperature, in 4-hour increments, in 3 trajectories. Shading indicates 95% confidence intervalsTable 1Post-op Complications by Trajectory. Complications N (%) Group 1 N = 2610 Group 2 N = 2472 Group 3 N = 412 Overall 76 (2.9)140 (5.7)64 (15.5) Venous Thromboembolism 35 (1.3)78 (3.2)36 (8.7) Urinary Tract Infection 19 (0.7)27 (1.1)12 (2.9) Pneumonia 13 (0.5)17 (0.7)13 (3.2) Skin/Soft-tissue infection 1 (<0.01)3 (0.1)3 (0.7)All p-values <0.01*Includes surgical site infectionDisclosures All authors: No reported disclosures.