SAEM Annual Meeting Chicago

Presentation Number:539

Background: Acute non-variceal upper gastrointestinal (GI) bleeding is a common indication for hospital admission. To appropriately risk-stratify such patients, endoscopy is recommended within 24 hours. Given the possibility to safely manage patients as outpatients after endoscopy, risk stratification as part of an emergency department (ED) observation unit (OU) protocol is proposed.
Objectives: Our objective was to determine the ability of an OU upper GI bleeding protocol to identify a low-risk population, and to expeditiously obtain endoscopy and disposition patients. We also identified rates of outcomes including changes in hemoglobin, abnormal endoscopy findings, admission, and revisits.
Methods: University, tertiary care ED with yearly census of 70,000 and a 20 bed OU that manages 500 patients/month on over 30 observation protocols.
All OU patients placed on the GI bleeding protocol from 3/1/09 to 1/31/11. Chart review was used to obtain demographics, presenting symptoms, medical history, OU course, lab data, endoscopy results, and return to health system. Descriptive analyses included proportions, means and medians with 95% CI.
Results: There were 98 eligible patients. Mean age was 41 years, 47% were male, and 34% African-American. Symptoms included hemetemesis (n=43), melena (32), hematochezia (25), and abdominal pain (66). No patients had cirrhosis. Charlson comorbidity score was <2 in 84 patients (86%). Initial hemoglobin was ≥10 g/dL in 92 patients (94%). Median hemoglobin change was 1.6 g/dL, with 14 patients (14%) dropping ≥3 g/dL. Median time from ED triage to OU placement was 4.7 hours and from OU placement to endoscopy was 11.9 hours. Median OU length of stay was 15.1 hours. Endoscopic findings included peptic ulcer in 13 (13%); gastritis, duodenitis, or esophagitis in 35 (36%); and active bleeding in 1(1%). Eighteen patients (18%) were admitted. Eleven patients (11%) returned to the ED within 7 days, with 7 admitted. ED return within 30 days occurred in 25 patients (25%) with 9 admitted.
Conclusion: A low risk cohort presenting with non-variceal upper GI bleeding can be selected for early endoscopy risk stratification and ED OU management. A majority of such patients can be safely discharged after endoscopy and do not return to the ED.

Presentation Number:540

Background: Acute uncomplicated pyelonephritis (pyelo) requires no imaging but a CT flank pain protocol (CTFPP) may be ordered to determine if patients with pyelo and flank pain also have an obstructing stone. The prevalence of kidney stone and the characteristics predictive of kidney stone in pyelo patients is unknown.
Objectives: To determine elements on presentation that predict ureteral stone, as well as prevalence of stone and interventions in patients undergoing CT for pyelo.
Methods: Retrospective study of patients at an academic ED who received a CTFPP scan between 8/05 and 4/09. 5497 CTFPPs were identified and 1899 randomly selected for review. Pyelo was defined as: positive urine dip for infection and >5 WBC/HPF on formal urinalysis in addition to: flank pain/CVA tenderness, chills, fever, nausea, or vomiting. Patients were excluded for: <18 y.o., renal disease, pregnancy, urological anomaly, or recent trauma. Clinical data (~178 elements) were gathered blinded to CT findings; CT results were abstracted separately and blinded to clinical elements. CT findings of hydronephrosis and hyrdroureter (hydro) were used as a proxy for hydro that could be determined by ultrasound prior to CT. Patients were categorized into three groups: ureteral stone, no significant findings, and intervention or follow-up required. Classification and Regression Tree analysis was used to determine which variables could identify ureteral stone in this population of pyelo patients.
Results: Out of the 1899 patients, 105 (7.0%) met criteria for pyelo; subjects had a mean age of 39 +/- 15.9 and 82% (n=87) were female. CT revealed 31 (29%, 95% C.I.= .22-.39) symptomatic stones, and 72 (68%, 95% C.I.= .59-.76) exams with no significant findings. Two patients needed intervention/ follow-up (1%, 95% C.I.= .0052-.0667), one for perinephric hemorrhage and the other for pancreatitis. Hydro was predictive for ureteral stone with an OR=18.4 (95% C.I.= 6.4-52, p<.0001). Eleven (35%) ureteral stone patients were admitted and 9 (8%) of them had procedures. Of these patients 100% had CT signs of obstruction, 8 (88%) had hydronephrosis and 1 (11%) had hydroureter.
Conclusion: Hydronephrosis was predictive of ureteral stone and in-house procedures. Prospective study is needed to determine whether CT scan is warranted in patients with pyelonephritis but without hydronephrosis or hydroureter.

Presentation Number:541

Background: Colorectal cancer (CRC) is the second leading cause of cancer death in the United States for men and women combined, and can present emergently with symptoms such as abdominal pain, bleeding, and obstruction. Emergency presentation as the first indication of colorectal cancer is generally thought to be associated with advanced disease and poor outcome.
Objectives: The specific aim of this analysis was to describe characteristics of patients presenting to the Emergency Department (ED) at their index diagnosis, and to determine whether emergency presentation precludes treatment with curative intent.
Methods: We performed a retrospective cohort analysis on a prospectively maintained institutional tumor registry to identify patients diagnosed with CRC from 2008-2010. EMRs were reviewed to identify which patients presented to the ED with acute symptoms of CRC as the initial sign of their illness. The primary outcome variable was treatment plan (curative vs. palliative). Secondary outcome variables included demographics, tumor type and location. Descriptive statistics were conducted for major variables. Χ² and Fisher’s exact tests were used to detect the association between categorical variables. Two-sample t-test was used to identify the association between continuous and categorical variables.
Results: Between Jan 1 2008 and Dec 31 2010, 376 patients were identified at our institution with CRC. 214(57%) were male and 162(43%) were female, with mean age 60.6; SD: 13.3. 33 (8.8%) patients initially presented to the ED, of which 5 (15.5%) received palliation. Of 339 patients who initially presented elsewhere, 69 (20.5%) received palliation. Acute ED presentation with CRC symptoms did not preclude treatment with curative intent (p = 0.47). Patients who presented emergently were more likely to be female (64% vs male 41%; p=0.01), and older (65 vs. 60; p=0.02). There was no statistically significant relationship between age, gender, tumor location or type and treatment approach.
Conclusion: Patients with CRC may present to the ED with acute symptoms, which ultimately leads to the diagnosis. Emergent presentation of CRC does not preclude patients from receiving therapy with curative intent.

Presentation Number:542

Background: There is a growing recognition of cannabinoid hyperemesis (CH) in the Emergency Medicine literature. The impact of this cannabis induced cyclic vomiting syndrome to Emergency Department (ED) practice and resource utilization remains unknown.
Objectives: To identify patients with suspected CH in the ED and quantify their resource utilization at three institutions.
Methods: This multicenter, ambispective cohort study identified patients over a two year period who presented with recurrent vomiting to a community and two academic teaching hospitals (annual ED volume 52,000, 93,000 and 53,000 patients respectively). Inclusion criteria were age ≥ 18 and ≥ 4 episodes over 12 months of recurrent, persistent vomiting. Patients were excluded if there was no history of cannabis use or if an alternative organic etiology for their recurrent vomiting was identified. The primary outcome was resource utilization among these patients, including imaging, ED visits and hospitalizations. Data analysis was descriptive.
Results: 20 patients were identified over a two year period with suspected CH. The mean age was 30 ± 10 years and 50% were male. The average duration of preceding illness was 4.8 ± 2.6 years. On average, patients smoked 4.3 marijuana joints per day, and 70% of the cohort reported compulsive warm bathing, a finding common in published CH patients. The mean number of abdominal computerized tomography (CT) scans, abdominal/pelvic ultrasounds and abdominal radiographs per patient were 5.3 ± 4.1, 3.8 ± 3.6 and 5.5 ± 6 respectively. The mean number of ED visits and hospital admissions per patient for symptoms of CH was 17.3 ± 13.6 and 6.8 ± 9.4 respectively. For one participating ED with 7 patients, the median charge for ED visits and hospital admissions over the course of a patient’s illness was $95,023 (range $62,420 to $268,110).
Conclusion: Patients with CH are uncommon but utilize substantial ED and hospital resources. Whether greater awareness of CH, early identification, and early counseling can limit such utilization is unknown.

Presentation Number:543

Background: Increased scrutiny is occurring from regulatory agencies including the Department of Public Health and The Joint Commission about the use of non-sterile enema preparations in the ED for constipation. This includes the “off-label” use of milk and molasses enemas as there is no reported data in the medical literature to determine safety and efficacy. Other preparations that are at risk to undergo scrutiny include tap water and soapsuds enemas.
Objectives: To evaluate the safety and efficacy of milk and molasses (M&M) enemas in the ED.
Methods: Structured retrospective by a trained data abstracter to review of ED records at two emergency departments (one urban teaching site, one community) between July 15, 2002 -July 15, 2010. Success and complications were defined a priori. Primary success was defined as the patient having a bowel movement. Secondary measures of success included improved pain score by 2 or more points or lowering of a heart rate initially over 100 bpm by 20 or more bpm. Complications were defined as: hemodynamic compromise, increased pain as reported on the nursing pain scale, electrolyte disturbances, bacteremia, bowel perforation, rectal pain or bleeding, cardiac dysrhythmias, anaphylaxis, dizziness or syncope or hospital admission for issues surrounding enema. Descriptive statistics were performed.
Results: Over the 8-year study period, there were 2013 enemas given, of which 261 were M&M enemas. 214 were given alone. 34 were given after other types of enemas or treatments. 13 were given prior to a secondary enema. Success rates (bowel movement) for M&M enemas were 188/214 (85.4%) and after other treatments 28/34. 5 additional patients improved with the secondary measures (87.3% overall success). There were 8/261 complications (3.1%) of which 4 had an increased HR, 2 had decreased BP, one had an increased pain score, and one had a fever to 101.5.
Conclusion: M&M enemas have a high success rate with a low complication rate when used in the ED.

Presentation Number:544

Background: In 2005, the AAMC released recommendations for clinical and effective communication skills expected of American medical school graduates. With ever-expanding medical school class sizes, have the clinical and communication skills deteriorated in the past 5 years?
Objectives: To perform a needs assessment of program directors regarding their expectations for clinical and communication skills of incoming PGY-1 residents independent of chosen medical specialty.
Methods: An internally created and expert-panel review survey was created and distributed to program directors for all ACGME accredited programs in our institution. The survey included content designed to assess (1) demographic information on each training program, (2) program director assessment of level of preparation for medical school graduates regarding clinical/ communication skills to date (10-point Likert scale with 10 indicating very well prepared), and (3) program director indications of importance for various clinical/ communication skills expected for incoming PGY-1 residents. The survey was distributed electronically using Zoomerang to all program directors with an endorsement of the Designated Institutional Official (DIO). Survey responses were collected and analyzed by the study authors.
Results: 27 of 71 (38%) of eligible program directors completed the survey. 17 (63%) represented medical specialties and the remaining 10 (37%) represented surgical specialties. The majority (63%) of programs were relatively small representing less than 20 total residents. Program director respondents predominantly had less than 5 years of experience (56%). Program directors indicate that incoming residents are less prepared in clinical skills recently as compared to 5 years prior (4.96 versus 6.0, 95% CI 4.19 - 6.72). A similar decreased level of preparation for communication skills was also found (5.52 versus 6.04, 95% CI 4.76 - 6.79). The clinical/ communication skills indicated as ‘most lacking’ in incoming PGY-1 residents were basic history and physical exam skills.
Conclusion: Medical students are less prepared than they were 5 years prior in the area of clinical and communication skills. The undergraduate medical education and GME communities may take this into account with future curricular policies and designs.

Presentation Number:545

Background: Patients seen in the emergency department (ED) may be dispositioned to an observation unit for up to 24 hours of additional evaluation and treatment. A certain proportion of these patients fails observation status and requires hospital admission. Patient factors predictive of admission have not been well studied. Identifying such factors would improve the accuracy of disposition decision-making in the ED.
Objectives: The primary objective of this study was to determine the relationship between advanced age and need for admission from an emergency department (ED) observation unit. The secondary objective was to determine if initial ED vital signs are predictive of admission.
Methods: We conducted a prospective, observational cohort study of ED patients placed in the ED-based observation unit of a tertiary care hospital. Data collected included age, vital signs, gender, medical history, laboratory values, observation protocol, and disposition. The primary outcome was admission versus discharge from the unit. Multivariable penalized maximum likelihood logistic regression was used to identify independent predictors of need for hospital admission. Age was examined continuously and at a cutoff of ≥65 years. Vital signs were examined continuously and at commonly accepted cutoffs. The analysis was additionally controlled for demographics, co-morbid conditions, laboratory values, and observation protocol.
Results: Three hundred patients were enrolled, 12% (n=35) ≥65 years old and 11% (n=33) requiring admission. In multivariable analysis, age was not associated with admission (odds ratio [OR] 0.30, 95% confidence interval [CI] 0.05-1.67). Significant predictors of admission included: initial systolic blood pressure ≥180 mmHg (OR 4.19, 95% CI 1.08-16.30), log Charlson co-morbidity score (OR 2.93, 95% CI 1.57-5.46), and white blood cell (WBC) count ≥14,000/mm3 (OR 11.35, 95% CI 3.42-37.72).
Conclusion: Age ≥65 years is not associated with need for admission from an ED observation unit. Older adults can successfully be cared for in these units. Initial temperature, respiratory rate, and pulse were not predictive of admission, but extremely elevated blood pressure was predictive. Other relevant predictor variables included comorbidities and elevated WBC count. Advanced age should not be a disqualifying criterion for disposition to an ED observation unit.

Presentation Number:546

Background: Approximately 1/3 of community-dwelling older adults experience a fall each year, and 2.2 million are treated in U.S. Emergency Departments (ED) annually. The ED offers a potential location for identification of high risk individuals and initiation of fall-prevention services that may decrease both fall rates and resource utilization.
Objectives: The goal of this study was to: 1) validate an approach to identifying older adults presenting with falls to the ED using administrative data; and 2) characterize the older adult who falls and presents to the ED and determine the rate of repeat ED visits, both fall-related and all visits, after an index fall-related visit.
Methods: We identified all older adults presenting to either of the 2 hospitals serving Durham County residents during a six month period. Manual chart review was completed for all encounters with ICD9 codes that may be fall-related. Charts were reviewed 12 months prior and 12 months post index visit. Descriptive statistics were used to describe the cohort.
Results: A total of 4452 older adults were evaluated in the ED during this time period, 1714 (55.7%) had an ICD9 code for a potentially fall-related injury. Of these, record review identified 534 (12%) with a fall from standing height or less. Of the fallers, 65.9% of the patients were discharged, 31% were admitted and 3% were admitted under observation. Of those who fell, 38.2% had an ED visit within the previous year. Approximately 1/3 (33.3%) of these were fall related. Over half (53.4%) of the patients who fell returned to the ED within one year of their index visit. A large proportion (44.4%) of the return visits was fall-related. Follow up with a primary care provider or specialist was recommended in 46% of the patients who were discharged. Overall mortality rate for fallers over the year following the index visit was 18%.
Conclusion: Greater than fifty percent of fallers will return to the ED after an index fall, with a large proportion of the visits related to a fall. A large number of these fallers are discharged home with less than fifty percent having recommended follow up. The ED represents an important location to identify high-risk older adults to prevent subsequent injuries and resource utilization.

Presentation Number:547

Background: Elderly patients frequently present to the emergency department after falling.
Objectives: We studied whether falls from a standing position resulted in an increased risk for intracranial or cervical injury verses falling from a seated or lying position.
Methods: This is a prospective observational study of patients over the age of 65 who presented with a chief complaint of fall to a tertiary care teaching facility. Patients were eligible for the study if they were over age 65, were considered to be at baseline mental status, and were not triaged to the trauma bay. At presentation, a questionnaire was filled out by the treating physician regarding mechanism and position of fall, with responses chosen from a closed list of possibilities. Radiographic imaging was obtained at the discretion of the treating physician. Charts of enrolled patients were subsequently reviewed to determine imaging results, repeat studies done, or recurrent visits. All patients were called in follow-up at 30 days to assess for delayed complications related to the fall. Data were entered into a standardized collection sheet by trained abstractors. Data were analyzed with Fisher Exact and descriptive statistics. This study was reviewed and approved by the institutional review board.
Results: Two-hundred sixty two patients were enrolled during the study period. One-hundred ninety eight of these had fallen from standing and 64 fell from either sitting or lying positions. The mean age for patients was 84 (SD 7.9) for those who fell from standing and 84 (SD 8.4) for those who fell from sitting or lying. There were 6 patients with injuries who fell from standing: three with subdural hematomas, one with a cerebral contusion, one with an osteophyte fracture at C6, and one with an occipital condyle fracture with a chip fracture of C1. There were 2 patients with injuries who fell from a seated or lying position: one with a traumatic subarachnoid hemorrhage and one with a type II dens fracture. The overall rate of traumatic intracranial or cervical injury in elders who fell was 3%. No patients required surgical intervention. There was no difference in rate of injury between elders who fell from standing versus those who fell from sitting or lying (p=1).
Conclusion: Elderly patients presenting with fall rarely have intracranial or cervical spine injury. Patient position at time of fall does not predict presence of injury.

Presentation Number:548

Background: Rapid identification of geriatric emergency department (ED) at high-risk for recidivism, functional decline, or death could focus preventative resources on susceptible populations.
Objectives: To compare the prognostic test performance for the Identification of Seniors at Risk (ISAR) and the Triage Risk Screening Tool (TRST) in community-dwelling geriatric ED patients stratified by Short-Blessed Test-defined dementia and lower health literacy as defined by the Rapid Assessment of Health Literacy in Medicine (REALM) for the composite outcome of 1- and 3-month ED recidivism, hospitalization, institutionalization, or death.
Methods: This prospective consecutive patient trial was conducted at one urban medical center ED. Eligible subjects were consenting English-speaking patients over age 65-years. A research assistant obtained the baseline 14-component Older American Resources and Services Activities of Daily Living (OARS ADL) score on each subject along with their ISAR and TRST scores. Another investigator blinded to the baseline results conducted structured telephone follow-up interviews at 1- and 3-months to quantify the composite outcome of reported ED recidivism, hospitalization, OARS ADL functional decline, and interval death.
Results: Among 225 enrolled subjects, 159 had successful 1-month follow-up. In those with follow-up, 45% were male with mean age 75-years, 54% were African-American, 45% had cognitive dysfunction, and 12% had an elementary school reading level. The TRST and ISAR labeled 65% and 82% of patients as high-risk, respectively. At three-months, 51% reported diminished function, 35% another ED evaluation and hospitalization, 2% had been institutionalized, and 70% had the composite outcome. Neither TRST nor ISAR predicted 1-month or 3-month composite outcomes in either general geriatric patients or those with cognitive impairment or lower health literacy (TABLE).
Conclusion: Both instruments identify the majority of patients as high-risk which will not be helpful in allocating scarce resources. Neither the ISAR nor the TRST can distinguish geriatric ED patients at high or low risk for 1- or 3-month adverse outcomes. These prognostic instruments are not more accurate in dementia or lower literacy subsets. Future instruments will need to incorporate different domains related to short-term adverse outcomes.

Presentation Number:549

Background: For older adults, both inpatient and outpatient care involves not only the patient and physician, but often a family member or informal caregiver. They can assist in medical decision making and in performing the patient’s activities of daily living. To date, multiple outpatient studies have examined the positive roles family members play during the physician visit. However, there is very limited information on the involvement of the caregiver in the ED and their relationship with the health outcomes of the patient.
Objectives: To assess whether the presence of a caregiver influences the overall satisfaction, disposition and outpatient follow-up of elderly patients.
Methods: We performed a three step inquiry of patients over 65 years old who arrived to the UPenn ED. Patients and care partners were initially given a questionnaire to understand basic demographic data. At the end of the ED stay, patients were given a satisfaction survey and followed through 30 days to assess: time to disposition, whether the patient was admitted or discharged, outpatient follow-up and ED revisit rates. Chi-squared and t-tests were used to examine the strength of differences in the elderly patients’ sociodemographics, self-rated health, receiving aid with their Instrumental Activities of Daily Living and number of health problems by accompaniment status. Multivariate regression models were constructed to examine whether the presence or absence of caregivers affected satisfaction, disposition and follow-up.
Results: Overall satisfaction was higher among patients who had caregivers (2.4 points), among patients who felt they were respected by their physician (3.8 points) and had lower lengths of stay (2 hours). Patients with caregivers were also more likely to be discharged home (OR 2.4) and to follow-up with their regular physician (OR 2.1). There was no evidence to suggest caregivers impacted the overall rates of revisits back to an ED.
Conclusion: For older adults, medical care involves not only the patient and physician, but often a family member or an informal care companion. These results demonstrate the positive influence of caregivers on the patients they accompany, and emergency physicians should define ways to engage these caregivers during their ED stay. This will also allow caregivers to participate when needed and can help to facilitate transitions across care settings.

Presentation Number:550

Background: Shared decision making has been shown to improve patient satisfaction and clinical outcomes for chronic disease management. Given the presence of individual variations in the effectiveness and side effects of commonly used analgesics in older adults, shared decision making might also improve clinical outcomes in this setting.
Objectives: We sought to characterize shared decision making regarding the selection of an outpatient analgesic for older ED patients with acute musculoskeletal pain and to examine associations with outcomes.
Methods: We conducted a prospective observational study with consecutive enrollment of patients age 65 or older discharged from the ED following evaluation for moderate or severe musculoskeletal pain. Two essential components of shared decision making, 1) information provided to the patient and 2) patient participation in the decision, were assessed via patient interview at one week using four-level Likert scales.
Results: Of 233 eligible patients, 110 were reached by phone and 87 completed the survey. Only 25% (21/87) of patients reported receiving ‘a lot’ of information about the analgesic, and only 21% (18/87) reported participating ‘a lot’ in the selection of the analgesic. There were trends towards white patients (p=.06) and patients with higher educational attainment (p=.07) reporting more participation in the decision. After adjusting for gender, race, education, and initial pain severity, patients who reported receiving ‘a lot’ of information were more likely to report optimal satisfaction with the analgesic than those receiving less information (78% vs. 47%, p<.05). After the same adjustments, patients who reported participating ‘a lot’ in the decision were also more likely to report optimal satisfaction with the analgesic (82% vs. 47%, p<.05) and greater reductions in pain scores (mean reduction in pain 4.6 vs. 2.7, p<.05) at one week than those who participated less.
Conclusion: Patient perceptions of shared decision making assessed at one week were associated with greater satisfaction with the medication and decreased pain scores. Prospective assessment of decision making at the time of the ED evaluation is needed to better understand the value of shared decision making in the selection of analgesics for older adults with acute musculoskeletal pain.

Presentation Number:551

Background: Older ED patients often present with muscle weakness and impaired mobility-related activities of daily living (mrADLs). These findings have been associated with inflammation and low serum 25-hydroxyvitamin D (25(OH)D) levels in older patients with chronic illnesses.
Objectives: We hypothesized that muscle weakness, reduced fat free mass (FFM), higher levels of inflammatory cytokines and lower 25(OH)D would be associated with mrADL impairment in the ED and mrADL decline at 30 days in older ED patients.
Methods: We performed a pilot, prospective, cohort study in an urban teaching hospital ED with 74,000 visits. We included patients ≥ 65 years. We excluded those with new onset chronic diseases that would affect mrADLs, from an extended care facility, or unable to complete the study procedures. We recorded demographics and mrADLs. Grip strength was measured using a hydraulic handheld dynamometer and was dichotomized by gender and body mass index. FFM was measured with the Heitman equation. Serum inflammatory cytokines (IL1β, IL6, IL10, TNF-α) were measured using a Luminex 200 multiplex analyzer. C-reactive protein (CRP) was measured using rate nephelometry. Serum 25(OH)D was measured using liquid chromatography - tandem mass spectrometry. Patients were called 30 days later to assess mrADLs. Non-normally distributed data were log transformed. Data are presented as means, proportions, and differences between means and proportions with 95% CI. No preset α error rates were determined for this pilot study.
Results: We evaluated 312 patients, 168 were excluded, 80 refused, 64 were enrolled. The mean age was 77 years; 53% were female, 97% were white, and 61% were admitted. Weakness was present in 31% (20/64, 95% CI 20-44%) and mrADL impairment in 28% (18/64, 95% CI 18-41%). Patients with impaired mrADLs in the ED were weaker, had a lower FFM, and lower 25(OH)D levels (Table 1); no differences in inflammatory cytokines were found. Follow-up was obtained on 92% (59/64). Those with mrADL decline at 30 days were weaker, had a lower FFM, higher log CRP, log IL6, and log IL10 (Table 2); no differences in 25(OH)D or other cytokines were found.
Conclusion: Weakness and lower FFM are associated with mrADL impairment in the ED and decline at 30 days. Lower 25(OH)D levels are associated with mrADL dependence in the ED. IL6, IL10, and CRP are associated with mrADL decline at 30 days.

Presentation Number:552

Background: Older adults frequently experience difficulty accessing prompt outpatient acute care and consequently must visit the ED for evaluation of minor illnesses. Telemedicine is a novel approach currently being evaluated to provide older adults in-home care for acute illnesses.
Objectives: To understand 1) experiences of patients and caregivers receiving care via telemedicine; 2) experiences of providers and support staff delivering such care; and 3) barriers and facilitators to delivering care via telemedicine to older adults.
Methods: We conducted a qualitative evaluation of a program providing senior living community residents with in-home telemedicine care for acute illness. A researcher accompanied the Certified Telemedicine Assistant (CTA) to observe and collect field notes. Afterwards, the researcher interviewed all stakeholders using a semi-structured interview guide to elicit opinions regarding their experiences. Demographic and clinical information was collected for all patients. Discrete statements from interviews and field notes were coded and arranged into themes. Concordance or discordance between stakeholders was grouped by visit to allow for triangulation.
Results: After 35 interviews from 10 telemedicine visits, redundancy was achieved. Statements were classified into 3 domains (clinical care, staff training, and technology) and 10 main themes. Patients and caregivers reported high satisfaction with their telemedicine-enhanced in-home care experience, remarking particularly on its convenience and promptness. Providers agreed that the quality of telemedicine-enhanced care was superior to telephone consultation, but disagreed about its efficiency (the time spent per visit). CTAs desired additional technical and geriatrics-specific training. Technological challenges were often cited as a barrier to program implementation. All stakeholders believed telemedicine helped patients avoid ED visits.
Conclusion: Given the positive views by stakeholders, additional refinement and evaluation are warranted. Concerns regarding efficiency and technological maturity warrant further attention but should be surmountable. Cost-effectiveness of telemedicine-enhanced acute care requires evaluation.

Presentation Number:553

Background: Fractures from falls is frequent complaint in the ED especially in females≥65 as a result of osteoporosis. The purpose of this study was to identify barriers to osteoporosis testing in elderly female patients presenting to the ED.
Objectives: To identify barriers to osteoporosis testing in elderly female patients presenting to the ED that services a large immigrant and non-English speaking population.
Methods: A prospective, cross-sectional, survey based study was conducted at Urban Level 1 Trauma center with annual ED visits of 120,000/year. Trained research assistants interviewed a convenience sample of patients over 36 months with a 3-page survey recording demographics, knowledge and obtaining DEXA scan testing for osteoporosis screening, and other current female preventive health recommendations from the Agency for Healthcare Research and Quality. Chi square test was used for categorical data as appropriate. Logistic regression was performed for the significant factors.
Results: A total of 796 females ≥65 yrs were interviewed over the study period with a median age of 79 yrs (IQR 73-85), 277(35%) immigrants with 692(87%) English speaking, 742(93%) had Private medical doctor (PMD), 45(5.65%) and 762(96%) insured. Overall 604(76%) had knowledge about osteoporosis screening and 27(3%) did not answer the question. Total 544(68%) had osteoporosis screening in the past and 27(3%) did not answer the question. Total 30 patients who didn’t answer the questions were excluded. There was significant difference in having the knowledge and receiving the osteoporosis test among females who had PMD and who didn’t have PMD on chi square test, with a p value of <0.02 and 0.01, respectively. Logistic regression with osteoporosis testing as outcome and adjusted with osteoporosis screening knowledge, age, immigration , smoking, having PMD and insurance status showed that it was associated with knowledge about osteoporosis screening with OR 48(CI 28-84), p<0.001 and was not associated with immigration status or having PMD.
Conclusion: Patients presenting to ED seem to be willing for osteoporosis screening once they are educated about it and is not associated with cultural back ground or having PMD. Every elderly patient in the ED could be easily taught about having osteoporosis screening to prevent risk of fractures in the future.

Presentation Number:554

Background: Quality of life (QOL) measurements have become increasingly important in outcomes-based research and cost-utility analyses. Dementia is a prevalent, often unrecognized, geriatric syndrome that may limit the accuracy of patient self report in a subset of patients. The relationship between caregiver and geriatric patient QOL in the emergency department (ED) is not well understood.
Objectives: To qualify the relationship between caregiver and geriatric patient QOL ratings in ED patients with and without cognitive dysfunction.
Methods: This was a prospective, consecutive patient, cross-sectional study over two-months at one urban academic medical center. Trained research assistants screened for cognitive dysfunction using the Short Blessed Test and evaluated health impairment using the Quality of Life-Alzheimer’s Disease (QOL-AD) Test. When available in the ED caregivers were asked to independently complete the QOL-AD. Consenting subjects were non-critically ill, English-speaking, community-dwelling adults over 65 years of age. Responses were compared using Wilcoxon Signed Ranks test to assess the relationships between patient and caregiver scores from the QOL-AD stratified by normal or abnormal cognitive screening results. Significance was defined by p<0.05.
Results: Patient QOL ratings were obtained from 108 patient-caregiver pairs. Patients were 51% female, 52% African-American with a mean age of 76-years, and 58% had abnormal cognitive screening tests. Compared with caregivers, cognitively normal patients had no significant QOL assessment differences except for questions of energy level and overall mood. On the other hand, cognitively impaired patients differed significantly on questions of energy level and ability to perform household chores with a trend towards significant differences for living setting (p=0.097) and financial situation (p=0.057). In each category, the differences reflected a caregiver underestimation of quality compared with the patient’s self-rating.
Conclusion: Discrepancies between QOL domains and total scores for patients with cognitive dysfunction and their caregivers highlights the importance of identifying cognitive dysfunction in ED-based outcomes research and cost-utility analyses. Further research is needed to quantify the clinical importance of the patient- and caregiver-assessed quality of life.

Presentation Number:555

Background: Age is often a predictor for increased morbidity and mortality. However, it is unclear whether old age is a predictor of adverse outcome in syncope.
Objectives: To determine whether old age is an independent predictor of adverse outcome in patients presenting to the emergency department following a syncopal episode.
Methods: A prospective observational study was conducted from June 2003 to July 2006 enrolling consecutive adult ED patients (>18 years) presenting with syncope. Syncope was defined as an episode of transient loss of consciousness. Adverse outcome or critical intervention were defined as gastrointestinal bleeding or other hemorrhage, myocardial infarction/percutaneous coronary intervention, dysrhythmia, alteration in antidysrhythmics, pacemaker/defibrillator placement, sepsis, stroke, death, pulmonary embolus or carotid stenosis. Outcomes were identified by chart review and 30 day follow-up phone calls.
Results: Of 575 patients who met inclusion criteria, an adverse event occurred in 24% of patients. Overall, 35% of patients with risk factors had adverse outcomes compared to 1.6% of patients with no risk factors. In particular, 28/127 (22%; 95% CI 16-30%) of patients <65 with risk factors had adverse outcomes, while 85/196 (43%; 95% CI 36-50%) of the elderly with risk factors had adverse outcomes. In contrast, among young people 2/196 (1%; 95% CI 0.04-3.8%) of patients without risk factors had adverse outcomes while 2/56 (3.6%; 95% CI 0.28-13%) of patients >65 without risk factors had adverse outcomes.
Conclusion: Although the elderly are at greater risk for adverse outcomes in syncope, age >65 or older alone does not appear to be a predictor of adverse outcome following a syncopal event. Based on this data, it should be safe to discharge home from the ED patients with syncope, but without risk factors, regardless of age.

Presentation Number:556

Background: Warfarin is the most commonly used long-term anticoagulant in the United States and certain antibiotics have been observed to potentiate its effects.
Objectives: Review data concerning patients on chronic warfarin therapy with a supratherapeutic INR when seen in the Emergency Department (ED) and correlate INR findings with antibiotic use both in the ED and as outpatients.
Methods: Five year retrospective chart review from a community ED with 40,000 annual visits. Inclusion criteria included age ≥65, on warfarin, INR ≥ 4.0, and no trauma. Data included 4 weeks medications, laboratory values, chief complaint and disposition. Descriptive statistics were analyzed and group comparisons made with Chi-squared and Student t-tests.
Results: 169 patients (42.6% male) met inclusion/exclusion criteria and had mean age 0f 80.7 years (SD: 7.5; Range: 65-97). 42 (24.8%) were on both warfarin and ASA, 7 (4.1%) on both warfarin and clopidogrel and 5 (3.0%) on all three: warfarin, clopidogrel, and ASA. 34 (20.1%) patients had been on an antibiotic in the last 4 weeks and 68 (40.2%) were given antibiotics in the ED despite an elevated INR. The mean initial ED INR was 5.55 (SD: 1.45) with 75% between 4.0 and 6.3. Of the 68 patients given an antibiotic in the ED, regardless of prior antibiotic use, 15 (22.1%) had an increase of their INR within 24 hours following ED treatment, significantly higher than the 7.9% INR increases for those patient not receiving antibiotics in the ED (RR=2.76; 95%CI: 1.25, 6.21; P=.009). 22 (64.7%) of the outpatient antibiotics and 39 (57.4%) of the antibiotics given in the ED are known to react with warfarin including quinolones (35.3%) and macrolides (11.8%). However there was no significant difference found in the mean INR of patients depending on prior antibiotic usage or antibiotic type.
Conclusion: In this study, 20.1% of elderly patients on warfarin seen with supratherapeutic INR in the ED had been on antibiotics during the preceding four weeks. Patient with supratherapeutic INR given any antibiotic in the ED will be three time more likely to increase their INR value within 24 hours. Emergency physician should be very cautious prescribing antibiotics to patients in the ED with supratherapeutic INR.

Presentation Number:557

Background: Nursing home patients, many of whom have complicated medical histories and cognitive impairment limiting their ability to transfer their medical information to medical providers, comprise a large number of Emergency department patients.
Objectives: We sought to determine if the absence of adequate history of present illness (HPI) as determined by emergency providers (EPs) affected emergency department length of stay (LOS) or likelihood of admission for nursing home patients.
Methods: This study was performed at a level 1 trauma center with an annual ED census of 75,000 patients. A convenience sample of 100 patients was enrolled prospectively by EPs between July and November of 2011. EPs were queried as to the adequacy of the HPI obtained from the patient as well as from the nursing home paperwork transferred with the patient. Charts were then retrospectively reviewed for total ED LOS and whether the patient was admitted or discharged. All data were entered into a standardized Excel spreadsheet by trained data abstractors. Differences in length of stay were analyzed with students T test. Differences in admission rates were compared with Fisher Exact. The study protocol was reviewed by the IRB and found to be exempt.
Results: The average age for enrolled patients was 83.5 (SD +/- 13). There was no difference in ED length of stay between patients in whom EPs deemed the nursing home paperwork as having an adequate HPI (177 minutes) versus those in whom they deemed the paperwork inadequate (163 minutes, p=0.46). The LOS for patients who could provide adequate history on their own was not significantly different from those who could not (162 minutes compared to 175 minutes, p = 0.44). 62 patients were discharged from the ED, and 38 were admitted. There was no difference in admission rates among patients in whom adequate HPI was obtained from nursing home paperwork or from the patient himself as compared to those in whom HPI could not be gathered by these means (p=0.36). Presence or absence of advanced directive paperwork also did not have any relationship to patient’s admission or discharge status (p = 0.30).
Conclusion: The absence of adequate HPI available to ED providers does not adversely effect patient LOS in the emergency department. Inadequate HPI does not appear to influence EPs disposition decisions.

Presentation Number:558

Background: Most ED HIV testing programs remain limited by the unsustainable cost of exogenous staff and penetration rates <10%. Kiosks have increased registration efficiency in clinical settings and have shown promising results for public health screening.
Objectives: To evaluate the feasibility and perceived acceptability of kiosks for offering HIV testing and gathering relevant information from ED patients.
Methods: An oral fluid HIV testing program for patients aged 18-64 years was instituted in 2006 with non-targeted screening utilizing a hybrid staffing model in an adult ED with an annual census of 60000. A 1-month usability study assessing integration of kiosk/tablet using a registration tablet (intended to integrate with future ED kiosk registration system) and a risk assessment kiosk to streamline HIV offering and risk assessment was conducted in July, 2011 and had extremely favorable results. The system was employed in a 24 day feasibility study for 16 hours/day during weekdays. Retrospective data analysis was performed to compare the numbers of patients offered and tested, and confirmed new HIV positive cases before and after implementation of the kiosk. We used 24 weekdays from August 2011 to September 1 as reference. Patient’s perceived acceptability was assessed by a 5 point Likert scale survey which assessed ease of use and preference. Chi square test was performed.
Results: Among 2738 acuity level 3-5 patients during the reference phase, 851, including 646 eligible patients, were offered testing. 508 (79%) patients were tested and 1 (0.2%) was confirmed positive. Among 2605 patients during the kiosk phase, 1030 (kiosk: 40% vs reference: 31%, p<0.05), including 783 eligible patients (kiosk: 30% vs reference: 24%, p<0.05), were offered testing via the registration tablet. 525 (67%) patients agreed to test and 481 (61%) completed the risk assessment kiosk and were tested. 1 (0.2%) tested was reactive. 88% of patients thought the kiosk was easy to use, and 71% preferred using the kiosk to an “in person” interaction to assess for interest in HIV testing and risk assessment.
Conclusion: A kiosk system at ED registration offered HIV testing to more patients than the hybrid staffing model, and over time could potentially increase the number of patient tested. Patients favorably perceived this novel approach which might offer a sustainable mechanism for ED HIV testing.

Presentation Number:559

Background: The CDC recommends targeted HIV screening for all patients seeking treatment for sexually transmitted infections (STI). Little is known regarding 1) ED compliance with this recommendation and 2) the prevalence of HIV in this subpopulation.
Objectives: The purpose of this study was to determine the HIV screening rate and prevalence of HIV among ED patients tested for STIs.
Methods: In our ED, a hybrid HIV testing program, combining both rapid HIV screening and physician-initiated rapid HIV testing, has been in place since 2009. During the time of this study, STI testing was at the discretion of the treating physician and there was no protocol linking HIV and STI testing. We retrospectively analyzed the laboratory and electronic medical records of all ED visits between January 1 and December 31, 2010 that had STI tests performed. STI tests included gonorrhea/chlamydia (GC/CT) and/or syphilis (RPR) tests. We determined the proportion of all STI-tested patients who also underwent HIV testing (defined as targeted HIV screening) and the prevalence of HIV among this group. Predictors of targeted HIV screening were identified using multivariable logistic regression.
Results: STI tests were performed in 4,531 (5.4%) of the 83,988 ED visits and 367 (8.0%) were positive for either GC/CT (n=332) and/or syphilis (n=35). The mean age of STI-tested patients was 33 (SD=15) years, and most were female (64%), Black (46%), English speaking (75%), and unmarried (83%). Targeted HIV screening was completed in 2,028 (44.8%) of the STI-tested patients, of which 17 (0.8%) were confirmed positive. In multivariable analysis, patients were less likely to be screened for HIV if they were: treated in the Fast-Track area (OR 0.59; 95% CI 0.45-0.77), non-English/non-Spanish speaking (OR 0.57; 95% CI 0.37-0.88), and tested for GC/CT only (OR 0.11; 95% CI 0.09-0.13). Patients were more likely to be screened for HIV if they were: admitted (OR 1.22; 95% CI 1.03-1.37) or tested GC/CT reactive (OR 1.32; 95% CI 1.05-1.42).
Conclusion: Despite the presence of an ED HIV testing program, many patients tested for STI were not screened for HIV. Our findings confirm that this is a high-risk population and that missed opportunities for the diagnosis of HIV likely occurred.

Presentation Number:560

Background: Adherence to national guidelines for HIV and syphilis screening in EDs is not routine. In our ED, HIV and syphilis screening rates among patients tested for gonorrhea and chlamydia (GC/CT) have been reported to be 45% and 30%, respectively.
Objectives: To determine the impact of a sexually transmitted infection (STI) laboratory order set on HIV and syphilis screening among ED patients tested for GC/CT. We hypothesized that a STI order set would increase screening rates by at least 30%.
Methods: A 6 month, quasi-experimental study in an urban ED comparing HIV and syphilis screening rates of GC/CT-tested patients before (control phase) and after the implementation of a STI laboratory order set (intervention phase). The order set linked blood-based rapid HIV and syphilis screening with GC/CT testing. Consecutive patients completing GC/CT testing were included. The primary outcome was the absolute difference in HIV and syphilis screening rates among GC/CT-tested patients between phases. We estimated that 550 subjects per phase were needed to provide 90% power (p-value of ≤0.05) to detect an absolute difference in screening rates of 10%, assuming a baseline HIV screening rate of 45%.
Results: The ED census was 42,461. Characteristics of patients tested for GC/CT were similar between phases: the mean age was 33 years (SD=12) and most were female (65%), Black (49%), Hispanic (30%), and unmarried (84%). GC/CT testing rates (5.6% control vs. 5.8% intervention, p=.57) and rates of reactive GC/CT tests (10.7% control vs. 9.1% intervention, p=.18) were similar between phases. The HIV screening rate among GC/CT-tested patients during the control phase was 41% (516/1,263) vs. 78% (969/1,241) during the intervention phase (absolute difference 37%, 95% CI 34% to 41%). The syphilis screening rate among GC/CT-tested patients during the control phase was 41% (521/1,263) vs. 72% (888/1,241) during the intervention phase (absolute difference 31%, 95% CI 27% to 34%). The intervention led to more syphilis diagnoses (control n=6 vs. intervention n=13) but not HIV (control n=7 vs. intervention n=5).
Conclusion: Laboratory order sets that automatically link HIV and syphilis tests to GC/CT tests significantly increase targeted screening rates. Larger studies are necessary to formally evaluate the yield associated with this increased screening.

Presentation Number:561

Background: HIV prevalence in Switzerland is 0.4% and 30% of HIV patients are still diagnosed with CD4 counts <200 cells/microliter. In March 2010, the Swiss Federal Office of Public Health published updated guidelines recommending Physician-initiated counseling and testing (PICT) for HIV testing if acute HIV infection is suspected or if HIV is among the differential diagnose with criteria laid out for each of these categories. Whist HIV testing in the emergency departments (ED) is recommended, only 1% of patients are currently screened. Lack of awareness among physicians has been cited in the literature as one barrier to guideline implementation.
Objectives: To test if physicians working in ED of 5 large teaching hospitals in western Switzerland, admitting 175,000 patients / year, are aware of updated national guidelines.
Methods: A survey was delivered to 167 ED physicians in the summer of 2011. The survey consisted of 26 case histories designed to test whether physicians would request an HIV test according to the new guidelines and if they knew when the PICT strategy was allowed. Finally, physicians were asked to estimate of the number of HIV tests they had ordered in the previous 4 weeks, and if they were aware of the new HIV guidelines. Results were presented as mean and standard deviation or median and interquartile range (IQR) for skewed data, or as proportions; Fisher’s exact test was used to compare proportions; a p value <0.05 indicated significant differences.
Results: 143 physicians returned the survey (86%); mean age was 32 +/- 8 years, with a median postgraduate experience of 6 years (IQR 3-12); 52% were male and 17% were attendings. 57% of respondents answered correctly ,with no difference between attendings and residents (p=0.09); 2 of the 3 questions with the lowest score (<30% answered correctly) were failure to recognize situations in which HIV testing was indicated, and the third one a failure to recognize acute HIV infection. 82% of physicians were not aware of the new guidelines. The median number of ordered test was 1 (IQR 0-2, range 1-10).
Conclusion: Swiss ED physicians rarely perform HIV tests and are not aware of current HIV screening guidelines. An education intervention is required if ED physicians are expected to play a significant role in the reduction of undiagnosed HIV patients.

Presentation Number:562

Background: In an effort to scale up HIV testing, the Centers for Disease Control and Prevention (CDC) released revised HIV testing recommendations in September 2006 calling for opt-out consent, elimination of separate written consent and curtailed counseling. There is little empirical data on patients’ response to these changes.
Objectives: Explore ED patient’s knowledge, beliefs, attitudes and acceptability towards revised CDC HIV testing recommendations.
Methods: Participants were recruited in proportion to the racial, gender and ethnic makeup of the study ED (in proportion stratified sampling). Trained qualitative Interviewers presented participants with a stimulus followed by questions about opt-out consent, elimination of separate written consent and curtailed counseling. Three investigators trained in qualitative methods coded all transcripts using an iterative coding strategy until theoretical saturation was achieved.
Results: 34 in-depth interviews were conducted, five with patients ages 13-17 and five with Spanish speaking patients. Nineteen (56%) participants were women. The mean age was 31. Most were Hispanic (38%) or African American/Black (44%). Only 1 (2.9%) participant knew about the revised testing recommendations. Qualitative accounts indicated that participants believed opt-out consent would result in increased testing but this was confounded by misunderstanding of the consent process, “so the opt-out is, you basically don’t have a choice”. Participants thought eliminating separate written consent was a positive change “it’s better to have everything in one form” but could result in people being tested without their knowledge. Attitudes diverged over curtailed counseling with some strongly in favor of keeping in-person counseling while others believed it was no longer necessary. Despite these differences participants felt patients, “should have options” for counseling because, “everybody ain’t the same”.
Conclusion: ED patients were unaware of the CDC’s revised HIV testing recommendations. Most believed opt-out consent and elimination of separate written consent were positive changes but could result in patient’s being tested without their knowledge. Attitudes toward curtailed counseling were polarized but participants agreed on the need to accommodate personal preferences.

Presentation Number:563

Background: The Centers for Medicare and Medicaid Services have recommended the use of immunization programs against influenza disease within hospitals since the 1980s. The Emergency Department (ED) being the “safety net” for most non-insured people is an ideal setting to intervene and provide primary prevention from influenza.
Objectives: The purpose of this study is to assess whether a pharmacist based influenza immunization program is feasible in the ED, and successful in increasing the percentage of adult patients receiving the influenza vaccine.
Methods: Implementation of pharmacist based immunization program was developed in coordination with ED physicians and nursing staff in 2010. The nursing staff using an embedded electronic questionnaire within their triage activity screened patients for eligibility for the influenza vaccine. The pharmacist using an electronic alert system within the electronic medical record identified patient who we deemed eligible and if agreed the pharmacist vaccinated the patient. Patient who refused to be vaccinated were surveyed to ascertain their perception concerning immunization offered by a pharmacist in the ED. Feasibility and safety data for vaccinating patient in the ED was recorded.
Results: 149 patients were approached and enrolled into the study. Of the 149, 41% agreed to receive the influenza vaccine from a pharmacist in the ED. The median screening time was 5 minutes and median vaccination time was 3 minutes for a total of 8 minutes from screening time to vaccination time. 74% were willing to receive the influenza vaccine from a pharmacist, and 78% were willing to receive the vaccine in the ED. The main reason given for refusing to receive the influenza vaccine was “patient does not feel at risk of getting the disease”; only 14.6% stated they were vaccinated recently.
Conclusion: A pharmacist based influenza immunization program is feasible in the ED, and has the potential to successfully increase the percentage of adult patients receiving the vaccine.

Presentation Number:564

Background: HIV screening in EDs is advocated to achieve the goal of comprehensive population screening. Yet, HIV testing in the ED is sometimes thwarted by a patient’s condition (e.g. intoxication) or environmental factors (e.g. other care activities). Whether it is possible to test these patients at a later time is unknown.
Objectives: We aimed to determine if ED patients who were initially unable to receive an HIV testing offer might be tested in the ED at a later time. We hypothesized that factors preventing testing are transient and that there are subsequent opportunities to repeat testing offers.
Methods: We reviewed medical records for patients presenting to an urban, academic ED who were approached consecutively to offer HIV testing during randomly selected periods from January 2008 to January 2009. Patients for whom the initial attempted offer could not be completed were reviewed in detail with standardized abstraction forms, duplicate abstraction, and third party discrepancy adjudication. Primary outcomes included repeat HIV testing offers during that ED visit, and whether a testing offer might eventually have been possible either during the initial visit or at a later visit within 6 months. Outcomes are described as proportions with confidence intervals.
Results: Of 824 patients approached, initial testing offers could not be completed for 120 (15%). These 120 were 62% male, 52% white, and had a median age of 41 (18-64). A repeat offer of testing during the initial visit would have been possible for 99/120 (83%), and 52/99 (53%) were actually offered testing on repeat approach. Of the 21 for whom a testing offer would not have been possible on the initial visit, 14 (67%) had at least one additional visit within 6 months, and 11/14 (79%) could have been offered testing on at least one visit. Overall, a repeat testing offer would have been possible for 110/120 (93%, 95% CI 85-96%).
Conclusion: Factors preventing an initial offer of HIV testing in the ED are generally transient. Opportunities for repeat approach during initial or later ED encounters suggest that, given sufficient resources, the ED could succeed in comprehensively screening the population presenting for care. ED screening personnel who are initially unable to offer testing should repeat their attempt.

Presentation Number:565

Background: The Center for Disease Control and Prevention recommends that emergency departments adopt an “opt-out” rapid HIV screening model in order to identify HIV infected patients. Previous studies nationwide have shown acceptance rates for HIV screening of 20-90% in emergency departments. However, it is unknown how acceptance rates will vary in a culturally and ethnically diverse urban emergency department.
Objectives: To determine the characteristics of patients that accept or refuse “opt-out” HIV screening in an urban emergency department.
Methods: A self-administered, anonymous, survey is administered to ED patients who are 18 to 64 years of age. The questionnaire is administered in English, Russian, Mandarin, and Spanish. Questions include demographic characteristics, HIV risk factors, perception of HIV risk, and acceptance of rapid HIV screening in the emergency department.
Results: To date 145 patients responded to our survey. Of the 145, 102 (70.3%) did not accept an HIV test (group 1) in their current ED visit and 43(29.7%) accepted an HIV test (group 2). The major two reasons given for opting out (i.e., group 1) was “I do not feel that I am at risk” (59.8%) and “I have been tested for HIV before” (25.5%). There was no difference between the groups in regards to gender (P=.737), age (P=.351), religious affiliation (.750), marital status (P=.331), language spoken at home (P=.211), and whether they had been HIV tested before (73.2% in group 1 and 59.4% in group 2; P=.123). However, there was a statistically significant difference with regards to educational level and income. More patients in group 1 (69.0%) and 46.1% in group 2 had less than a college level education (P<.05). Similarly, more patients in group 1 (58.3%) and only 34.8% in group 2 had an annual household income of ≤ $25,000.00 (P<.05).
Conclusion: In a culturally and ethnically diverse urban emergency department, patients with a lower socioeconomic status and educational level tend to opt out of HIV screening test offered in the ED. No significant difference in acceptance of ED HIV testing was found to date based on primary language spoken at home or religious affiliation

Presentation Number:566

Background: EDs are well recognized to be a frequent source of care for known HIV-infected individuals in the U.S. Understanding healthcare utilization patterns of HIV-infected ED patients may provide useful information for improving provision of services for this patient population. However, few studies have investigated this issue at the national level.
Objectives: To describe and compare the demographic characteristics and ED resource utilization patterns of known HIV-infected adult ED patients in the United States in 2009 with those of the general ED patient population.
Methods: We identified demographics, HIV infection status, and ED utilization pattern in 2009 from a weighted sample of U.S. ED visits using the National Hospital Ambulatory Medical Care Survey, a nationally representative survey. Data on visits by patients aged ≥ 18 years were analyzed using procedures for multiple-stage survey data.
Results: In 2009, there was an estimated 102,899,094 visits to hospital EDs by adults in the U.S. Among them, 698,313 visits or 7-in-1000 ED visits were documented as HIV-infected. Using U.S. Census estimate in adult population (n=232,637,362) and CDC’s estimate (known positive status: n=941,950), the estimated ED visit rate was 0.74 per year for known infected persons while the rate was 0.44 per year for non-HIV infected individuals (p<0.05). As compared to general ED patients, these visits were more likely be male (61% vs. 43%, p<0.05), and Black (51% vs. 23%, p<0.05) but there were no differences by age and geographical region. While no difference was recorded in reported urgency for evaluation between the 2 groups, the total number of diagnostic/screening services ordered or provided was significantly higher in visits for HIV-infected patients (diagnostic: 4.7±0.4 vs. 3.4±0.1, p<0.05). Furthermore, HIV-infected patients also had a significantly higher number of medications administered in the ED (1.9±0.2 vs. 1.4±0.1, p<0.05). ED visits by HIV-infected patients also had longer lengths of ED stay (317±26.0 minutes vs. 222.5±5.6 minutes, p<0.05) and were more likely to be admitted (29% vs. 15%, p<0.05), than their non-HIV infected counterparts.
Conclusion: Although ED visits by HIV-infected individuals in the U.S. are relatively infrequent, they occur at rates higher than the general population, and consume significantly more ED resources than the general population.

Presentation Number:567

Background: The influence of wound age on the risk of infection in simple lacerations repaired in the emergency department (ED) has not been well studied. It has traditionally been taught that there is a “golden period” beyond which lacerations are at higher risk of infection and therefore should not be closed primarily. The proposed cutoff for this golden period has been highly variable (3-24 hours in surgical textbooks).
Objectives: To answer the following research question: are wounds closed via primary repair after the golden period at increased risk for infection?
Methods: We searched MEDLINE, EMBASE, and other databases as well as bibliographies of relevant articles. We included studies that enrolled ED patients with lacerations repaired by primary closure. Exclusion: 1. intentional delayed primary repair or secondary closure, 2. wounds requiring intra-operative repair, skin graft, drains, or extensive debridement, and 3. grossly contaminated or infected at presentation.
We compared the outcome of wound infection in two groups of early versus delayed presentations (based on the cut-offs selected by the original articles). We used “Grading of Recommendations Assessment, Development and Evaluation” (GRADE) criteria to assess the quality of the included trials. Frequencies are presented as percentages with 95% confidence intervals. Relative risk (RR) of infection is reported when clinically significant.
Results: 418 studies were identified. Four trials enrolling 3724 patients in aggregate met our inclusion/exclusion criteria. Two studies used a 6-hour cut-off and the other two used a 12-hour cut-off for defining delayed wounds. The overall quality of evidence was low. The infection rate in the wounds that presented with delay ranged from 1.4% to 32%. One study with the smallest sample size (Morgan et al), which only enrolled lacerations to hand and forearm, showed higher rate of infection in patients with delayed wounds (Table). The infection rate in delayed wound groups in the remaining three studies were not significantly different from the early wounds.

Conclusion: The evidence does not support the existence of a golden period nor does it support the role of wound age on infection rate in simple lacerations.

Presentation Number:568

Background: Although clinical studies in children have shown that temperature elevation is an independent and significant predictor of bacteremia in children, the relationship in adults is largely unknown or equivocal.
Objectives: Review the incidence of positive blood cultures on critically ill adult septic patients presenting to an Emergency Department (ED) and determine the association of initial temperature with Bacteremia.
Methods: July 2008 to July 2010 retrospective chart review on all patients admitted from the ED to an urban community hospital with sepsis and subsequently expiring within 4 days of admission. Fever was defined as a temperature ≥38°C. SIRS criteria were defined as: 1) temperature ≥ 38°C or ≤ 36°C, 2) heart rate ≥ 90 beats/minute, 3) respiratory rate ≥ 20 or mechanical ventilation, 4) WBC ≥ 12,000/mm3 or < 4,000 or bands ≥ 10%. Data abstracted includes age, gender, chief complaint, admitting diagnosis, initial temperature, days to death, WBC, lactate, blood and urine culture results.
Results: 117 cases met inclusion/exclusion criteria. 99 (84.6%) had two or more SIRS criteria. Fever was present in 49 (41.9%) cases but 17 (14.5%) were hypothermic. There were 23 patients (19.7%) with positive blood cultures. Gram negative organisms grew in 16 (69.6%) patients, with Escherichia Coli and Streptococcus Pneumoniae being the most prevalent gram negative and gram positive bacteria. Patients with fever had 28.6% positive blood cultures versus 15.7% for euthermic patients and 5.9% for hypothermic patients. Only fever was found to be an independent predictor of a positive blood culture in this group of critical sepsis patients. The adjusted odd ratio was 2.86 (95% CI: 1.02, 8.04; P=.046).
Conclusion: In this study of critically ill adult septic patients, only fever was found to be a significant predictor for positive blood culture with an adjusted odd ratio of 2.86.

Presentation Number:569

Background: Blood cultures are a commonly ordered test. However, the general indications for blood cultures are poorly defined and culture yields are remarkably low. The true positive rate is usually low (2-8%) with few clinical usual results. There are often equal false positive to true positive rates. The low yield and false positives have significant financial costs.
Objectives: The purpose of this study is to undertake a detailed study of the use of blood cultures in the ED and to identify predictors of bactermia. We hypothesize that most cultures are unnecessary to patient care and can be safely eliminated.
Methods: This is a retrospective chart review taking place at an teaching hospital with an annual census of 40-45,000. All patients for whom a blood culture was performed in 2009-2010 were eligible for inclusion. A detailed chart review was performed on consecutive patients from 10/1-12/31/2010. Demographic, clinical and laboratory data were abstracted for each patient. All variables were then used to derive a clinical prediction model for culture positivity.
Results: : There were 2872 blood cultures done in 2009 (7% of patients), 3375 done in 2010 (8%). Detailed information was obtained on 600 patients. There were 90 positive culture results (15%). Of these 28 were contaminants. 361 patients had a focal infection and a local source culture was done in 257 (of which 151 were positive). 44 (50%) patients with positive blood cultures had local infections and of these 30 had focal cultures with positive results. 21 were concordant, but 9 were discordant and of these the blood culture was a contaminant in 7. Variables contained in the predictor model were temperature, total WBC, neutrophil %, platelet count, and presence of a focal infection. Using a probability of >2.5%, the sensitivity was 94%, while reducing the number of cultures by 25%.
Conclusion: Using an evidence-based approach to guide the use of blood cultures can lead to significant cost savings. We have developed a prediction rule for predicting a true positive blood culture. Having a focal infection is an independent and important predictor of bacteremia. However, results of blood culture results are concordant with focal culture result and when discordant overwhelmingly contamination. Even in those patients with a >2.5% probability of bacteremia, by our model, blood cultures could be forgone in those with focal infections.

Presentation Number:570

Background: The International Classification of Diseases, 9th Edition (ICD9) conflates "cellulitis and abscess" under a single heading. This has prevented epidemiologic studies based on national datasets from determining the relative importance of abscess versus other skin and soft-tissue infections (SSTI) during the emergence of community-associated methicillin-resistant S. aureus (CA-MRSA).
Objectives: We sought to determine the increase in abscesses as distinct from non-abscess SSTIs during the emergence of CA-MRSA. We hypothesized that abscesses increased while non-abscess SSTIs did not.
Methods: We analyzed diagnostic and billing codes from all visits, 1997-2007 in an urban emergency department, in an area now endemic for CA-MRSA. We identified SSTI visits via ICD codes 681-682 (with subcodes), then used billing codes for drainage procedures to identify abscesses. We hand-reviewed 3% of all SSTI records to quantify misclassification by billing data. We use logistic regression to model the odds of abscess vs. any ED diagnosis by year, reporting the annual increase in the odds ratio with its 95% confidence interval.
Results: From 1997-2007, 1.25% of all visits were for SSTI, and 22% had procedures indicating abscess. The odds of a visit being for abscess (versus any other diagnosis) increased 11% per year (95%CI 9-13%), or 3.1-fold over the 11-year study period. Non-abscess SSTI increased only 2% per year (95%CI 1-3%). Among SSTI visits, the odds of a visit being for abscess increased 8% per year (95%CI 6-10%), or 2.4-fold over the study period. Billing codes misclassified 40% of abscesses as non-abscesses, but only 1.5% of non-abscesses as abscesses, likely rendering these results conservative.
Conclusion: We found a dramatic increase in abscesses and minimal increase in other SSTIs. This provides indirect support for current treatment guidelines, which do not recommend that CA-MRSA be targeted for most non-abscess SSTIs.

Presentation Number:571

Background: Antimicrobial resistance is a problem that affects all emergency departments.
Objectives: Our goal was to examine all urinary pathogens and their resistance patterns from urine cultures collected in the Emergency Department (ED).
Methods: This study was performed at an urban/suburban community-teaching hospital with an annual volume of 40,000 visits. Using electronic records, all cases of urine cultures received in 2009 were reviewed for data including type of bacteria, antibiotic resistance, and healthcare exposure (HCX). HCX was defined as no prior hospitalization within the previous six months, hospitalization within the previous three months, hospitalization within the previous six months, nursing home resident (NH), and presence of an indwelling urinary catheter (UC). An investigator abstracted all data with a second re-abstracting a random 5% for Kappa statistics between 0.697 and 1.00. Group comparisons were made using Pearson Chi-square or Fisher test as appropriate with simple logistic regression. Significance was set at 0.05.
Results: 1935 positive urine cultures were obtained. Seventy-nine cultures were excluded as contaminants and one culture was excluded due to missing data, leaving a total of 1855. 52.4% (972/1855) of patients had no HCX within the previous six months, 27.9% (517/1855) were NH residents, 19.7% (366/1855) had HCX within the previous six months, and 15.6% (279/1855) had HCX within the previous three months. 28.4% (527/1855) were from patients that had a UC. Pathogens obtained included: Escherichia coli (56.9%), Klebsiella (10.4%), Enterococcus (10.1%), Proteus (8.9%), Pseudomonas (4.4%), Staphococcus aureus (2.5%), Citrobacter (2.3%), and Enterobacter (1.9%). The resistance pattern for Escherichia coli was significantly affected more by HCX (P-values 0.001 to 0.008) than UC (P-values 0.033 to 0.554). Proteus resistance were similar to Escherichia coli but Klebsiella was the opposite.
Conclusion: Our study results maintain that E. coli is the most common pathogen seen in all of our collected ED urine cultures sent regardless of healthcare exposure. It further suggests that resistance is affected by patient healthcare exposure and presence of an indwelling urine catheter.

Presentation Number:573

Background: Coagulase-negative staphylococci are the most common cause of catheter and implanted device infection. They are also the most common cause of false positive blood cultures. Thus, patients from whose blood these organisms are recovered often face mandatory hospitalization and broad spectrum antibiotics until the clinical significance of the culture can be determined (usually days). Improved means of discriminating pathogenic from contaminating organisms are greatly needed.
Objectives: We examined the utility of limited genetic sequencing of bacterial isolates using multilocus sequence typing (MLST) to discriminate between known pathogenic blood culture isolates of S. epidermidis and isolates recovered from skin.
Methods: Ten blood culture isolates from patients meeting the Centers for Disease Control and Prevention (CDC) criteria for clinically significant S. epidermidis bacteremia and ten isolates from the skin of healthy volunteers were studied. MLST was performed by sequencing ~ 400 bp regions of 7 genes (arc, aroE, gtR, mutS, pyr, tpiA, and yqiL). Genetic variability at these sites was compared to an international database (www.sepidermidis.mlst.net) and each strain was then categorized into a genotype on the basis of known genetic variation. The ability of the gene sequences to correctly classify strains was quantified using the support vector machine function in the statistical package R. 1,000 bootstrap resamples were performed to generate confidence bounds around the accuracy estimates.
Results: Between strain variability was considerable, with yqiL being most variable (6 alleles) and tpiA being least (1 allele). The mutS gene, responsible for DNA repair in S. epidermidis, showed almost complete separation between pathogenic and commensal strains. When the 7 genes were used in a joint model, they correctly predicted bacterial strain type with 90% accuracy (IQR 85, 95%).
Conclusion: Multilocus sequence typing shows excellent early promise as a means of distinguishing contaminant versus truly pathogenic isolates of S. epidermidis from clinical samples. Near-term future goals will involve developing more rapid means of sequencing and enrolling a larger cohort to verify assay performance.

Presentation Number:574

Background: Emergency departments (ED) are an essential part of healthcare system response to large-scale disease outbreaks such as influenza. Already stretched to near capacity, the expected increase in patient volume in the event of pandemic influenza requires the implementation of interventions to reduce ED crowding. Determining the most effective interventions to manage patient flow during specific types of outbreaks may improve care and enhance safety.
Objectives: To identify and categorize ED and hospital-based interventions to mitigate the impact of ED crowding from four types of respiratory disease outbreaks using an expert panel: 1) low-severity, low-volume (i.e. seasonal flu) 2) low-severity, high-volume (i.e. 2009 H1N1), 3) high-severity, low-volume (SARS), and 4) high- severity, high-volume (i.e. 1918 H5N1).
Methods: Study team members who were subject experts reviewed the existing academic literature for potential ED and hospital-based interventions studied to manage ED crowding from a respiratory disease outbreak. After identifying an initial list of potential interventions, a survey was distributed to 34 additional influenza and emergency care experts, who rated each intervention on their perception of both its ease of implementation and importance on a 1-5 Likert scale. Each intervention was rated for use in each of the 4 separate influenza outbreak scenarios. Survey results were presented to survey participants at an in-person conference in November 2011. After discussion, each participant categorized interventions into “Low Priority,” “Medium Priority,” or “High Priority,” for each scenario using the nominal group technique. Data were then aggregated via a weighted scoring system to generate a final categorization of each intervention.
Results: The pre-conference survey achieved a 100% response rate of all 34 expert participants. The highest-performing interventions as identified during the conference are presented by influenza scenario in Table 1 and Table 2.
Conclusion: We identified several potentially useful interventions to mitigate ED crowding from an influenza outbreak. Recommended priority levels for interventions varied by the type of influenza outbreak scenario. Prioritization of these high-performing interventions during a future respiratory outbreak may enhance response effectiveness and maximize scarce resources.

Presentation Number:575

Background: While cough is a prominent symptom and mode of transmission for severe respiratory infections, e.g. influenza, tuberculosis, and pertussis, the acoustic characteristics of cough have not been assessed for their value as a disease-screening tool. Real-time screening of respiratory infections would significantly improve appropriate triage, evaluation and isolation of patients who present to emergency departments (EDs).
Objectives: We conducted an exploratory study using acoustic analysis to characterize coughs associated with respiratory illnesses.
Methods: We collected cough recordings from patients who presented to an urban ED of a tertiary care academic medical center from May 2010 to May 2011 with influenza (flu) or flu-like illness (ILI), pneumonia, other respiratory infections, or non-infectious respiratory conditions, as well as healthy controls. We used acoustic analysis software to characterize cough acoustic features. Acoustic features distinguishing between healthy controls and patients with respiratory conditions were identified. We then tested the ability of each acoustic feature to distinguish specific respiratory illnesses. Operating characteristics of sensitivity vs. false-positives were plotted for all possible thresholds of each feature. Pairwise comparisons were made using the Kolmogorov-Smirnov test.
Results: We enrolled 247 patients and 136 controls. We identified six potentially relevant cough acoustic features: 1) Onset Slope, 2) Variability about the Local Median (VLM), 3) Middle Slope, 4) Final Shape, 5) Train Length and 6) Amplitude Ratio. Each feature distinguished some comparisons among respiratory illnesses. Results for several features were highly statistically significant (p<<0.001), e.g. VLM distinguished individuals with both ILI and confirmed flu from controls, from pneumonia, from all other respiratory conditions (all p<<0.001) and from bronchitis (p=0.02). Onset and Train Length each distinguished flu/ILI from controls (p<<0.001). The features had low cross-correlations.
Conclusion: Automatic algorithmic processing of patient cough recordings shows great promise as a method for real-time, rapid screening in the ED. Incorporating multiple cough features can be expected to improve performance. These features need to be tested in a larger sample size to determine if they have clinical utility.

Presentation Number:576

Background: Antiviral medications are recommended for patients with influenza who are hospitalized or at high risk for complications. However, timely diagnosis of influenza in the ED remains challenging. Influenza rapid antigen tests have short turn-around times, making them potentially useful in the ED setting, but their sensitivities may be too low to assist with treatment decisions.
Objectives: To evaluate the test characteristics of the BinaxNow Influenza A&B rapid antigen test (RAT) in ED patients.
Methods: We prospectively enrolled a systematic sample of patients of all ages presenting to two EDs with acute respiratory symptoms or fever during three consecutive influenza seasons (2008-2011). Research personnel collected nasal and throat swabs, which were combined and tested for influenza with RT-PCR using CDC-provided primers and probes. ED clinicians independently decided whether to obtain a RAT during clinical care. RATs were performed in the clinical laboratory using the BinaxNow Influenza A&B test on nasal swabs collected by ED staff. The study cohort included subjects who underwent both a research PCR and clinical RAT. RAT test characteristics were evaluated using PCR as the criterion standard with stratified sub-analyses for age group and influenza subtype (pandemic H1N1 (pH1N1), non-pandemic influenza A, influenza B).
Results: 561 subjects were enrolled; 131 subjects were PCR positive for influenza (76 pH1N1, 20 non-pandemic influenza A, and 35 influenza B). For all age groups, RAT sensitivities were low and specificities were high (Table). RAT sensitivities were low for each influenza subtype: pH1N1: 0.276 (95% CI: 0.183-0.393); non-pandemic influenza A: 0.150 (95% CI: 0.040-0.389); influenza B: 0.229 (95% CI: 0.110-0.406).
Conclusion: The BinaxNow Influenza A&B RAT demonstrated low sensitivity across all age groups and influenza subtypes in a three-year sample of ED patients. A negative BinaxNow RAT should not be used as a criterion to withhold antiviral medications.

Presentation Number:577

Background: Healthcare-associated infections (HAI) are a leading cause of preventable morbidity, mortality and costs. Little is known about infection prevention practices in the ED setting, though patient variety, high throughput, and crowding make the ED an important site for transmission of HAI.
Objectives: To compare ED vs. Intensive Care Unit (ICU) implementation of evidence-based practices to prevent central line-associated bloodstream infections (CLABSI) and catheter-associated urinary tract infections (CAUTI), and improve hand hygiene (HH). We hypothesized lower adoption in EDs than ICUs.
Methods: We developed one survey for ED directors focused on ED practices, and another for infection preventionists (IPs) focused on hospital practices. Sites without an ICU reported ward data (4.0%). We first conducted a 22-site pilot, revising the instruments based on interviews with respondents and pilot site providers. We then surveyed a random sample (n=412) of all non-specialty, non-federal US EDs from the National ED Inventory (N=4824), with disproportionate stratified sampling for teaching hospitals (n=119) and high-volume EDs (annual census >50,000; n=147). We mailed surveys to ED directors and hospital IPs up to 3 times and phoned and emailed non-respondents. We compare proportions with chi-square tests. Proportions, 95% CIs and statistics are weighted to produce nationally representative estimates.
Results: 73% of EDs, 84% of IPs, and 62% of both responded. 52% (CI 45-59%) of EDs have implemented the central line insertion checklist vs. 92% (88-96%) of hospital ICUs (P<.001). 33% (27-39%) of EDs require documentation of an appropriate indication for Foley catheter placement, vs. 51% (44-58%) of ICUs (P=.001). Almost all EDs and ICUs monitor HH by direct observation (99% vs. 98%, ns). HH compliance was ≥80% for the most recent reporting period at 48% (40-56%) of EDs, vs. 72% (64-76%) of ICUs/wards (P<.001). Fewer EDs than ICUs reported participating in a project to reduce CLABSI or improve HH (Table 1).
Use of the central line checklist and appropriate HH are less frequent in EDs than in ICUs (Table 2).
Conclusion: Evidence-based infection prevention practices are less prevalent in US EDs compared to hospital ICUs. Research is needed to develop effective implementation strategies for EDs

Presentation Number:578

Background: Staphylococcus aureus has been reported to cause community acquired pneumonia (CAP), but the prevalence and clinical features of staphylococcal CAP are not well described.
Objectives: To describe the prevalence and clinical characteristics of culture-positive S. aureus CAP among patients hospitalized with CAP in Middle Tennessee.
Methods: As part of the ongoing multi-site CDC EPIC study, we prospectively enrolled children (<18 years old) and adults hospitalized with CAP at three Nashville hospitals, 1/3/2010 - 9/24/2011. Inclusion criteria included clinical evidence of acute respiratory infection and radiographic evidence of pneumonia. Exclusion criteria included: recent hospitalization; tracheostomy; cancer with neutropenia; HIV infection with CD4 < 200; and among nursing home residents, inability to independently perform activities of daily living. Sources for bacterial cultures included blood, sputum (adults only), bronchoalveolar lavage (BAL), tracheal aspirate and pleural fluid. Only sputum specimens with a Bartlett score ≥ 1+ were considered adequate for culturing.
Results: Among 461 children enrolled, 7 (2%) had S. aureus cultured from ≥ 1 specimen, including 5 with methicillin-resistant S. aureus (MRSA) and 2 with methicillin-susceptible S. aureus (MSSA). Specimens positive for S. aureus included 3 pleural fluid, 2 blood, 2 tracheal aspirates, and 1 BAL. Two children with S. aureus had evidence of co-infection: 1 influenza A, and 1 Streptococcus pneumoniae. Among 673 adults enrolled, 17 (3%) grew S. aureus from ≥ 1 specimen, including 9 with MRSA and 8 with MSSA. Specimens positive for S. aureus included 5 blood, 11 sputum, and 3 BAL. Five adults with S. aureus had evidence of co-infections: 2 coronavirus, 1 respiratory syncytial virus, 1 S. pneumoniae, and 1 Pseudomonas aeruginosa. Presenting clinical characteristics and outcomes of subjects with staphylococcal CAP are summarized in Tables 1-2.
Conclusion: These preliminary findings suggest S. aureus is an uncommon cause of CAP. Although the small number of staphylococcal cases limits conclusions that can be drawn, in our analysis staphylococcal CAP appears to be associated with co-infections, pleural effusions and severe disease. Future work will focus on continued enrollment and developing clinical prediction models to aid in diagnosing staphylococcal CAP in the ED.

Presentation Number:579

Background: Resuscitation of a critically-ill child requires knowledge of a child's weight. Since delaying resuscitations to obtain a weight measurement is impractical, it is important to have other methods for rapid weight estimation. Accurate methods for weight estimation of children are readily available in developed countries; however, their utility in developing countries with higher rates of malnutrition and infectious disease is unknown.
Objectives: The goal of this study was to determine the accuracy of a height-based weight estimation (i.e. the Broselow tape) compared to age-based estimations among pediatric patients in Western Kenya.
Methods: A two-month prospective cross-sectional study of all sick children presenting to a government referral hospital in Eldoret, Kenya was performed. 967 children were divided into weight-based groups of <10kg (n=267), 10-18 kg (n=390), and >18 kg (n=310) for analysis. Measured weight was compared to predicted weights according to the Broselow tape and commonly used Advanced Pediatric Life Support (APLS) and Nelson's age-based formulas. The method for weight prediction was determined to be equivalent to the actual weight if the 95% confidence interval for the mean percentage difference between the actual and predicted weight was less than 10%. Accuracy was also defined by agreement on the Broselow color zone.
Results: The overall mean percentage difference between the actual weight and Broselow predicted weight was -2.2% while APLS and Nelson's predictions were -5.2% and -10.4% respectively. The Broselow tape closely predicted the weight of children in Western Kenya, overestimating weight by only 0.45 kg. The mean percentage difference between the actual and Broselow predicted weight by weight groups was +3.1%, -3.1%, and -5.6%. The overall agreement between Broselow color zone and actual weight was 65.5% with a tendency to overestimate actual weight by one color zone.
Conclusion: The Broselow tape accurately predicts the weights of children in Western Kenya and outperforms age-based formulas.

Presentation Number:580

Background: Conducting surveys to assess the impact of humanitarian assistance in complex emergencies is particularly difficult due to restricted access and insecurity. Lot quality assurance sampling (LQAS) is one method that can be used in such a setting.
Objectives: This investigation evaluated the primary health and water/sanitation services of a non-governmental organization (NGO), Medair, in conflict-affected West Darfur, Sudan, to examine the feasibility of LQAS in assessing humanitarian aid in a complex emergency.
Methods: Respondents were chosen using a stratified, three-stage probability sampling method. The catchment area was divided into seven supervisory areas (SAs); 19 individuals were identified from each SA. Respondents were women between 15-49 years who had a live birth within the last two years. A questionnaire primarily based on UNICEF’s Multiple Indicator Cluster Survey 4 was developed for the assessment.
Results: 82% of pregnant women had at least two antenatal care visits, but 46% received clean delivery kits. There was low coverage of the polio vaccine (29%) among children 12-23 months. Only 29% of children with fever received proper treatment for malaria, and 29% of children with diarrhea received increased fluid intake. Of households surveyed, 78% had access to an improved water source. Only 19% of households had hand-washing facilities.
Conclusion: LQAS is a rapid and statistically valid assessment method yielding data representative of a population. It can be applied in complex humanitarian emergencies. The results of this survey showed that the provision of primary healthcare in West Darfur has room for improvement. LQAS helped Medair manage their resources to improve low-performing aspects of their programs in low-performing SAs. LQAS has multiple advantages over the more widely used cluster sampling method. LQAS requires fewer samples and less time on the ground in each community to conduct surveys, a crucial factor as it decreases the risk to surveyors in insecure settings. Periodic LQAS surveys over defined time intervals are a practical management tool for NGOs to compare the performance of programs by multiple indicators and by SAs. LQAS is a useful tool to help NGOs understand how best to allocate their limited resources to meet performance targets.

Presentation Number:581

Background: Emergency Care has been a neglected public health challenge in Sub-Saharan Africa (SSA). The goal of Global Emergency Care Collaborative (GECC) is to develop a sustainable model for emergency care delivery in low resource settings. GECC is developing a training program for Emergency Care Practitioners (ECPs).
Objectives: To analyze the first 500 patient visits at Karoli Lwanga "Nyakibale" Hospital ED in rural Uganda to determine the knowledge and skills needed in training ECPs.
Methods: A descriptive cross-sectional analysis of the first 500 consecutive patient visits in the ED’s patient care log was reviewed by an unblinded abstractor. Data on demographics, procedures, laboratory testing, bedside ultrasounds (US) performed, radiographs (XRs) ordered, and diagnoses were collated. All authors discussed uncertainties and formed a consensus. Descriptive statistics were performed.
Results: Of the first 500 patient visits, procedures were performed in 367 (73.4%) patients, including 244 (48.8%) who had IVs placed, 47 (9.4%) who received wound care and 42 (8.4%) who received sutures. Complex procedures, such as procedural sedations, lumbar punctures, orthopedic reductions, nerve blocks, and tube thoracostomies, occurred in 49 (9.8%) patients. Laboratory testing, XRs and USs were performed in 188(37.6%), 99 (19.8%) and 45 (7%) patients, respectively.
Infectious diseases were diagnosed in 217 (43.4%) patients; 78 (15.6 %) with malaria and 57 (11.4%) with pneumonia. Traumatic injuries were present in 140 (28%) patients; 77 (15.4%) needing wound care and 31 (6.2%) with fractures. Gastrointestinal and neurological diagnoses affected 58 (11.6%) and 27 (5.4%) patients, respectively.
Conclusion: ECPs providing emergency care in SSA will be required to treat a wide variety of patient complaints and effectively use laboratory testing, XRs and USs. This demands training in a broad range of clinical, diagnostic and procedural skills, specifically in infectious disease and trauma, the two most prevalent conditions seen in this rural SSA ED.

Presentation Number:582

Background: Few studies assess which ultrasound applications are utilized in Tanzania. No studies correlate ultrasound diagnosis to clinical outcomes. Point-of-care ultrasound in Tanzania is likely limited secondary to lack of formal education in ultrasound.
Objectives: To assess which ultrasound applications are most utilized in Tanzania, then determine accuracy of ultrasound diagnoses compared to surgical findings. Additionally, assess the ultrasound knowledge of Tanzanian healthcare providers after a one-day workshop.
Methods: A retrospective chart review was done of patients who received an ultrasound examination in Shirati, Tanzania from January 2009 to May 2010. Pre-operative ultrasound diagnoses of surgical patients were compared to operative pathology.
Emergency ultrasound fellowship trained physicians provided didactic and hands-on training for one-day to Tanzanian healthcare providers. A pre- and post-test was administered to assess current knowledge and immediate recall.
Results: 469 ultrasound exams were performed: 159 obstetric, 100 gynecologic, 72 abdominal, 16 renal and 11 miscellaneous. We excluded 111 exams due to insufficient data. Ninety-two percent of patients were female, average age 30 years (range 2-80). Surgical data on 949 patients during the corresponding time frame was obtained. 39 patients received a pre-operative ultrasound: 12 obstetrical, 21 gynecologic, and 6 abdominal exams. Sensitivity and specificity of bedside ultrasound to detect surgical pathology was 80% (95% CI 0.60-0.91) and 11% (95% CI 0.00-0.49), respectively.
Thirty-five healthcare providers attended the ultrasound workshop. We excluded 8 participants due to incomplete test results. Paired t-test showed a mean pre- and post-test scores of 20.4%, (SD 0.12) and 53.7% (SD 0.15), P<0.001, respectively. Mean pre- and post-test scores were hepatobiliary 12.6% and 37.8%, physics 25.3% and 71%, OB/Gyn 29.7% and 68.5% and FAST 31% and 56.7%, respectively. There was a statistically significant improvement of all post-test scores except for the FAST exam.
Conclusion: Ultrasound is primarily used in Tanzania for obstetric applications. Healthcare providers can adequately use bedside ultrasound, however, specificity is poor. After an ultrasound course, there was a statistically significant increase in immediate recall of knowledge across all topics, except for FAST.

Presentation Number:583

Background: Current Chinese EMS is faced with many challenges due to a lack of systematic planning, national standards in training, and standardized protocols for pre-hospital patient evaluation and management.
Objectives: To estimate the frequency with which pre-hospital care providers perform critical actions for selected chief complaints in a county-level EMS system in Hunan Province, China.
Methods: In collaboration with Xiangya Hospital (XYH), Central South University in Hunan, China, we collected data pertaining to pre-hospital evaluation of patients on EMS dispatches from a “120” call center over a 2-month period. This call center services an area of just under 5000 km² with a total population of 1.36 million. Each EMS team consists of a driver, a nurse, and a physician. This was a cross-sectional study where a single trained observer accompanied EMS teams on transports of patients with a chief complaint of chest pain, dyspnea, trauma, or altered mental status. In this convenience sample, data was collected daily between 8 AM and 6 PM. Critical actions were pre-determined by a panel of emergency medicine faculty from XYH and the University of Maryland School of Medicine. Simple statistical analysis was performed to determine the frequency of critical actions performed by EMS providers.
Results: During the study period, 1170 patients were transported, 452 of whom met the inclusion criteria. 218 (48.2%) evaluations were observed directly for critical actions. Table 1 shows the frequency of critical actions performed by chief complaint. None of the patients with chest pain received an ECG even though the equipment was available. Rapid glucose was checked in only 2.1% of patients presenting with altered mental status. A lung exam was performed in 22.7% of patients with dyspnea, and the respiratory rate was measured in 9.1%. Among patients transported for trauma, blood pressure and heart rate were only measured in 1% and 4.1%, respectively.
Conclusion: In this observation study of pre-hospital patient assessments in a county-level EMS system, critical actions were performed infrequently for the chief complaints of interest. Performance frequencies for critical actions ranged from 0 to 22.7%, depending on the chief complaint. Standardized pre-hospital patient care protocols should be established in China and further training is needed to optimize patient assessment.

Presentation Number:584

Background: The utilization of non-invasive positive pressure ventilation (NIPPV) has become widespread in emergency departments (ED). Current clinical guidelines recommend its use in acute cardiogenic pulmonary edema (CHF) and chronic obstructive pulmonary disease (COPD). It remains unclear whether NIPPV is safe and effective in the management of other presentations of respiratory failure (RF).
Objectives: The objective of this study is to assess the utilization of NIPPV in ED RF.
Methods: This is a retrospective analysis in a tertiary academic receiving facility from 2005-2011. Inclusion criteria included all patients >18 years with a presentation of RF requiring emergency mechanical ventilation (MV). The charts were reviewed for pertinent demographic and admission data, including co-morbidities, type of MV, LOS and outcome. Variables were analyzed using student T test and chi square where appropriate.
Results: Over the 5 year period a total of 1173 patients met criteria, of which 277 (23.61%) had NIPPV, and 896(76.39%) were intubated (ETI). An increase in utilization of NIPPV was observed over the study period: 12 cases in 2005, 43 in 2009, 125 in 2010. Within the NIPPV group, there were fewer deaths 41(14.8%) vs 228(25.45%) p=0.0001, lower hospital length of stay (LOS) 10.44 days 95%CI(9.16,11.73) vs 17.34 95%CI(16,18.7) p<0.0001 and ICU LOS 4.0 days (3.35,4.65) vs 9.91(9.01, 10.81) p<0.0001. While most patients (43%) initiated on NIPPV had CHF 61(22%) or COPD 58(21%), other conditions included pneumonia 45(16%), sepsis 39(14%), aspiration 15(5%), and trauma 4(1%). NIPPV was utilized for palliation in 13 (4.7%) cases. There were 54(19.5%) failures of NIPPV requiring intubation in the ED 8(2.9%) and inpatient 46(16.6%). NIPPV failures were represented by diagnoses of multi-trauma, COPD, pneumonia, overdose, asthma, and musculoskeletal disorder.
Conclusion: NIPPV in this study was successful in the management of diverse presentations of acute respiratory failure with few observed failures in the ED. Prospective studies are needed to propose the broader application of this modality and to investigate factors predicting failure.

Presentation Number:585

Background: Approximately 12-20% of patients treated with epinephrine for anaphylaxis receive a second dose but the risk factors associated with repeat epinephrine use remain poorly defined.
Objectives: To determine whether obesity is a risk factor for requiring 2+ epinephrine doses for patients who present to the emergency department (ED) with anaphylaxis due to food allergy or stinging insect hypersensitivity.
Methods: We performed a retrospective chart review at four tertiary care hospitals that care for adults and children in New England between the following time periods: Massachusetts General Hospital (1/1/01-12/31/06), Brigham and Women’s Hospital (1/1/01-12/31/06), Children's Hospital Boston (1/1/01-12/31/06), Hasbro Children's Hospital (1/1/04-12/31/09). We reviewed the medical records of all patients presenting to the ED for food allergy or stinging insect hypersensitivity using ICD9CM codes. We focused on anthropomorphic data and number of epinephrine treatments given before and during the ED visit. Among children, calculated BMIs were classified according to CDC growth indicators as underweight, healthy, overweight or obese. All patients who presented on of after their 18th birthday were considered adults.
Results: The study population was comprised of 310 ED patients (250 children and 60 adults) who received epinephrine for anaphylaxis and had data for both height and weight; 260 (84%) received 1 dose of epinephrine while 50 (16%) received 2+ doses. Patients had a median age of 8.6 (IQR 3.0-16.7) years and 46% were male. Overall, 23 (7%) were underweight, 162 (52%) healthy weight, 56 (18%) overweight, and 69 (22%) obese. The distribution of BMI categories was similar in those receiving 1 vs. 2+ doses (P=0.79). Among those receiving 1 dose, 8% were underweight, 52% healthy, 18% overweight and 22% obese. Among those receiving 2+ doses, 4% were underweight, 54% healthy, 18% overweight and 24% obese. There was no difference in the need for additional doses of epinephrine between adults and children (P=0.90). Finally, when adjusting for age and sex, there was no association between obesity (vs. non-obesity) and receiving 2+ doses (odds ratio 1.10; 95% confidence interval 0.68-1.78; P=0.71).
Conclusion: Among ED patients with anaphylaxis treated with epinephrine, obesity was not associated with 2+ doses.

Presentation Number:586

Background: Little is known about the comparative effectiveness of noninvasive ventilation (NIV) versus invasive mechanical ventilation (IMV) in chronic obstructive pulmonary disease (COPD) patients with acute respiratory failure.
Objectives: To characterize the use of NIV and IMV in COPD patients presenting to the emergency department (ED) with acute respiratory failure and to compare the effectiveness of NIV vs. IMV.
Methods: We analyzed the 2006-2008 Nationwide Emergency Department Sample (NEDS), the largest, all-payer, US ED and inpatient database. ED visits for COPD with acute respiratory failure were identified with a combination of COPD exacerbation and respiratory failure ICD-9-CM codes. Patients were divided into 3 treatment groups: NIV use, IMV use, and combined use of NIV and IMV. The outcome measures were inpatient mortality, hospital length of stay (LOS), hospital charges, and complications. Propensity score analysis was performed using 42 patient and hospital characteristics and selected interaction terms.
Results: There were an estimated 101,000 visits annually for COPD exacerbation and respiratory failure from approximately 4,700 EDs. Ninety-six percent were admitted to the hospital. Of these, NIV use increased slightly from 14% in 2006 to 16% in 2008 (P=0.049), while IMV use decreased from 28% in 2006 to 19% in 2008 (P<0.001); the combined use remained stable (~4%). Inpatient mortality decreased from 10% in 2006 to 7% in 2008 (P<0.001). NIV use varied widely between hospitals, ranging from 0% to 100% with median of 11%. In a propensity score analysis, NIV use - compared to IMV - significantly reduced inpatient mortality (risk ratio, 0.57; 95% confidence interval [CI], 0.48-0.56), shortened hospital LOS (difference, -3 days; 95%CI, -4 to -3), and reduced hospital charges (difference, -$36,044; 95%CI, -$38,234 to -$33,855). NIV use was associated with a lower rate of iatrogenic pneumothorax compared with IMV use (0.04% vs. 0.6%, P<0.001). An instrumental analysis confirmed the benefits of NIV use, with a 5% reduction in inpatient mortality in the NIV-preferring hospitals.
Conclusion: NIV use is increasing in US hospitals for COPD with acute respiratory failure; however, its adoption remains low and varies widely between hospitals. NIV appears to be more effective and safer than IMV in the real-world setting.

Presentation Number:587

Background: Dyspnea is a common ED complaint with a broad differential diagnosis and disease-specific treatment. Bronchospasm alters capnographic waveforms, but the affect of other causes of dyspnea on waveform morphology is unclear.
Objectives: We evaluated the utility of capnographic waveforms in distinguishing dyspnea caused by reactive airway disease (RAD) from non-RAD in adult ED patients.
Methods: This was a prospective, observational, pilot study of a convenience sample of adult patients presenting to the ED with dyspnea. Waveforms, demographics, past medical history, and visit data were collected. Waveforms were independently interpreted by two blinded reviewers. When the interpreters disagreed, the waveform was re-reviewed by both reviewers and an agreement was reached. Treating physician diagnosis was considered the criterion standard. Descriptive statistics were used to characterize the study population. Diagnostic test characteristics and inter-rater reliability are given.
Results: Fifty subjects were enrolled. Median age was 52 years (range 21-82), 50% were female, 34% were Caucasian. 29/50 (58%) had a history of asthma or chronic obstructive pulmonary disease. RAD was diagnosed by the treating physician in 19/50 (38%) and 32/50 (64%) had received treatment for dyspnea prior to waveform acquisition. The interpreters agreed on waveform analysis in 47/50 (94%) cases (Kappa = 0.88). Test characteristics for presence of acute RAD, including 95%CI, were: overall accuracy 70% (55.2%-81.7), sensitivity 69% (43.5%-86.4%), specificity 71% (51.8%-85.1%), positive predictive value 59% (36.7%-78.5%), negative predictive value 79% (58.5%-91.0%), positive likelihood ratio 2.25 (1.36-3.72), negative likelihood ratio 0.42 (0.23-0.74).
Conclusion: Inter-rater agreement is high for capnographic waveform interpretation, and shows promise for helping to distinguish between dyspnea caused by RAD and dyspnea from other causes in the ED. Treatments received prior to waveform acquisition may affect agreement between waveform interpretation and physician diagnosis, impacting the observed test characteristics.

Presentation Number:588

Background: Asthma and chronic obstructive pulmonary disease (COPD) patients who present to the Emergency Department (ED) usually lack adequate ambulatory disease control. While evidence-based care in the ED is now well defined, there is limited information regarding the pharmacologic or non-pharmacologic needs of these patients at discharge.
Objectives: This study evaluated patients’ needs with regard to the ambulatory management of their respiratory conditions after ED treatment and discharge.
Methods: Over 6 months, 94 adult patients with acute asthma or COPD, presenting to a tertiary care Alberta Hospital ED and discharged after being treated for exacerbations were enrolled. Using results from standardized in-person questionnaires, charts were reviewed by respiratory researchers to identify care gaps.
Results: Overall, 58 asthmatic and 36 COPD patients were enrolled. More patients with asthma required education on spacer devices (52% vs 31%). Few asthma (9%) and no COPD patients had written action plans; asthma patients were more likely to need adherence counseling (53% vs 36%) for preventer medications. More patients with asthma required influenza vaccination (72% vs 39%; p = 0.003); pneumococcal immunization was low (36%) in COPD patients. Only 22% of asthmatics reported ever being referred to an asthma education program and 19% of the COPD patients reported ever being referred to pulmonary rehabilitation. At ED presentation, 28% of the asthmatics required the addition of inhaled corticosteroids (ICS) and 16% required the addition of ICS/long acting beta-agonist (ICS/LABA) combination agents. On the other hand, 36% of COPD patients required the addition of long acting anticholinergics (LAAC) while most (83%) were receiving preventer medications. Finally, 31% of COPD and 29% of asthma patients who smoked required smoking cessation counseling.
Conclusion: Overall, we identified various care gaps for patients presenting to the ED with asthma and COPD. There is an urgent need for high-quality research on interventions to reduce these gaps.

Presentation Number:589

Background: Non-invasive positive pressure ventilation (NPPV) improves survival and decreases length of stay in patients with respiratory failure secondary to acute exacerbation of chronic obstructive pulmonary disease (AECOPD). Application of NPPV is common in most Emergency Departments (EDs); however, it is unclear whether patients are being screened appropriately for respiratory failure and if those meeting criteria for NPPV receive this intervention.
Objectives: We measured the adherence to eligibility screening using blood gas (BG) values, as well as NPPV treatment for eligible AECOPD patients.
Methods: In this retrospective study involving three 400+ bed tertiary care hospitals in Calgary, Alberta, data from the Regional Emergency Department Informatics System (i.e., tracking and admissions data) and Sunrise Clinical Manager (i.e., laboratory data and physician order entry) databases were linked for descriptive analysis. All patients with a diagnosis of AECOPD requiring hospitalization between May1, 2010 and March 31, 2011 and who received systemic corticosteroids (both as markers for severity) were included. Outcomes were: 1) the proportion of eligible patients who had either arterial or venous BG measurement; 2) the proportion of these patients with hypercapnic respiratory failure meeting eligibility criteria for NPPV (pCO2 > 45 and pH < 7.30); and 3) the proportion of NPPV-eligible patients receiving a trial of NPPV treatment.
Results: 912 eligible AECOPD patients were included in the analysis. Overall, 67.4% (95% CI: 64.3%, 70.5%) underwent BG determination with 19.2% (95% CI: 16.2%, 22.6%) meeting eligibility criteria for NPPV. Only 59.3% (95% CI: 49.9%, 68.2%) of these patients with respiratory failure had NPPV ordered, while 4.4% (95% CI = 2.9%, 6.7%) of all patients whose BG analysis did not demonstrate respiratory failure criteria had NPPV administered.
Conclusion: We found a significant practice gap related to screening for NPPV and use of NPPV in AECOPD. To our knowledge, no administrative metric has been proposed to assess the quality of AECOPD care in the ED. Further prospective validation of this quality improvement metric is warranted.

Presentation Number:590

Background: Serum levels of soluble ST2, a member of the interleukin-1 receptor family, predict mortality in ED patients with dyspnea secondary to acute heart failure and acute coronary syndrome. High levels of ST2 have also been described in pulmonary disease, but it is unknown if these are associated with adverse outcomes.
Objectives: We hypothesized that elevated ST2 levels would be associated with 180-day mortality and 180-day return ED visits or hospital readmission in patients presenting to the ED with non-cardiac causes of dyspnea.
Methods: This prospective observational study enrolled a convenience sample of patients presenting to a single academic tertiary care ED with a chief complaint of dyspnea. Exclusion criteria included new onset heart failure, prior heart failure with current BNP>500 pg/mL, ischemic chest pain, ECG changes indicative of myocardial infarction or ischemia, heart transplant, pericardial effusion, upper airway obstruction or chest wall trauma. Blood samples were collected at enrollment, and ST2 levels were compared to index ED/hospital outcomes and 180-day return ED visits, hospital readmissions and death. Staff were blinded to ST2 levels. Differences in medians were assessed using the Mann-Whitney U test.
Results: Of 82 patients enrolled, 37 (45%) were male, 33 (40%) were white, and 34 (42%) were hospitalized. The most frequent ED/hospital diagnosis was COPD/asthma in 29 (35%) patients. One patient died during index admission. During 180-day follow-up, 36/81 (44%) had a return ED visit, 21/81 (26%) were readmitted and 5/82 (6%) were deceased. Median ST2 level (ng/mL) was 227 (range 38-318) in patients who died vs. 32 (range 10-400) in those who did not (difference 195, 95% CI 48 to 343, p=0.006). Median ST2 level was 59 (range 13-400) in readmitted patients vs. 31 (range 10-353) in non-readmitted patients (difference 28, 95% CI -3 to 60, p=0.036). Median ST2 levels were 41 (range 11-400) in patients with a return ED visit vs. 31 (range 10-353) in those without a return visit (difference 10, 95% CI -10 to 20, p=0.23).
Conclusion: Patients with non-cardiac dyspnea who died or required readmission to the hospital within 180 days had higher levels of ST2 compared to those who did not. Soluble ST2 may have prognostic value in diseases with a primary pulmonary etiology.

Presentation Number:591

Background: The NAEPP Expert Panel Report 3 suggests that repeated lung function measures (FEV1 or PEF) 1 hour after initiation of treatment is the strongest single predictor of hospitalization. It is also stated that signs and symptoms scores may improve the ability to predict subsequent hospitalization.
Objectives: To compare repeated FEV1 to a repeated Modified Borg Dyspnea Scale (DS) as a predictor of subsequent hospitalization.
Methods: This is an interim, sub-analysis of an interventional, double-blinded study performed in an academic urban-based adult ED. Subjects with acute exacerbation of asthma with FEV1<50% predicted within 30 minutes following initiation of “standard care” (including a minimum of 5 mg nebulized albuterol; 0.5 mg nebulized ipratropium; and 50 mg corticosteroid) who consented to be in a trial were included. All treatment was administered by emergency physicians unaware of the study objectives. Patients were randomly assigned to treatment with placebo or an intravenous beta agonist. All subjects had FEV1 and DS obtained at baseline, 1, 2, and 3 hours after treatment. FEV1 was measured using a bedside Nspire spirometer, and DS was calculated using a Modified Borg Dyspnea score.
Results: 38 patients were included for analysis. Spearman’s Rho test (Rho) was used to measure correlations between FEV1 and DS at 1, 2, and 3 hours post study entry and subsequent hospitalization. Rho is negative for FEV1 (higher FEV1 correlates to lower rate of hospitalization) and positive for DS (higher DS correlates to higher rate of hospitalization). At each time point, DS were more highly correlated to hospitalization than were FEV1 (Table 1).
Conclusion: Dyspnea score at 1, 2, and 3 hours were significantly correlated with hospital admission, whereas FEV1 was not. In this set of subjects with moderate to severe asthma exacerbations, a standardized subjective tool was superior to FEV1 for predicting subsequent hospitalization.

Presentation Number:592

Background: We have previously described increased serum lactate concentrations in subjects with acute asthma exacerbation. It is not clear if this is due to increased work of breathing or possibly a side effect of treatment.
Objectives: 1) Determine if there is a significant correlation between increased treatment lactate or Δ lactate and serum albuterol concentrations during treatment of an acute asthma exacerbation after adjusting for dyspnea score (DS). 2) Determine if elevated treatment lactate concentrations or Δ lactate concentrations are associated with increased hospital admissions.
Methods: This is an interim, subgroup analysis of a prospective, interventional, double-blind study performed in an academic urban ED. Subjects that were consented for this trial presented with acute asthma exacerbations with FEV1 < 50% predicted within 30 minutes following initiation of “standard care” (includes a minimum of 2.5 mg nebulized albuterol; 0.5 mg nebulized ipratropium; and 50 mg of a corticosteroid). ED physicians who were unaware of the study objectives administered all treatments. Subjects were randomized in a 1:1 ratio to either placebo or investigational intravenous beta agonist arms. Blood was obtained at 1 and 1.25 hours after the start of the hour long infusion. Blood was centrifuged and serum stored at -80 degrees C, and then shipped on dry ice for albuterol and lactate measurements at a central lab. The treatment lactate and Δ lactate were correlated with 1 hr serum albuterol concentrations and hospital admission using partial Pearson correlations to adjust for DS.
Results: 38 subjects were enrolled to date, 20 with complete data. The mean baseline serum lactate level was 18.1 mg/dL (SD±8.6). This increased to 32.7 mg/dL (SD±15.0) at 1.25 hrs. The mean 1 hr DS was 3.85 (SD±2.0). The correlation between treatment lactate, Δ lactate, 1 hr serum albuterol concentrations (R, S and total) and admission to hospital are shown (Table 1). Both treatment and Δ lactate were highly conrrelated with total serum albuterol, R albuterol, and S albuterol. There was no correlation between treatment lactate or Δ lactate and hospital admission.
Conclusion: Lactate and Δ lactate concentrations correlate with albuterol concentrations in patients presenting with acute asthma exacerbations after adjusting for dyspnea score, but do not correlate with hospital admission.

Presentation Number:593

Background: According to a recent CDC report, the prevalence of asthma in the United States increased 14% from 2001 to 2009. However, wider use of treatments such as corticosteroids appears to be reducing asthma exacerbation rates. It is uncertain which of these opposing factors would predominate in affecting the number of emergency department (ED) asthma visits.
Objectives: Our objective was to investigate this by examining recent trends in the number of ED asthma visits in a large database.
Methods: Design: Retrospective cohort. Setting: Consecutive patients seen by ED physicians in four suburban hospitals in New Jersey from 1/1/1996 to 12/31/2010. Protocol: We identified asthma patients by ICD9 codes. Data Analysis: We compared the ratio of yearly asthma visits to total ED visits for all patients and for specific gender and age groups. We used the Student t test, calculated 95% confidence intervals (CIs) and performed regression analyses. Alpha was set at 0.05.
Results: Of the 2,860,096 ED visits, 58,371visits (2.4%) had asthma. Fifty one percent were < 21 years old and 54% were female. We found a decline of 27% (95% CI: 23% - 30%, p<0.0001, R²=0.73, p<0.0001) in the overall yearly asthma visits to total ED visits from 1996 to 2010. When we analyzed gender and age groups separately, we found no statistically significant changes for females or for males < 21 years old (R² < 0.016, p > 0.65). For females and males > 21 years old, yearly asthma visits to total ED visits from 1996 to 2010 decreased 39% (95% CI: 33% - 43%, p<0.0001, R² = 0.90, p < 0.0001) and 20% (95% CI: 14% - 26%, p<0.0001, R² = 0.80, p < 0.0001), respectively.
Conclusion: We found an overall decrease in yearly asthma visits to total ED visits from 1996 to 2010. We speculate that this decrease is due to greater corticosteroid use despite the increasing prevalence of asthma. It is unclear why this decrease was seen in adults and not in children and why it was greater for adult females than males.

Presentation Number:594

Background: Electronic medical records (EMR) have been adopted by many U.S. healthcare systems. This provides new opportunities for research data collection.
Objectives: Our objectives were to describe the use of a unique data collection system that leveraged EMR technology and to compare its data entry error rate to traditional paper data collection.
Methods: This is a retrospective review of data collection methods during the first 12 months of a multicenter study of ED, anti-coagulated, head injury patients. On-shift ED physicians at five centers enrolled eligible patients and prospectively completed a data form. Enrolling ED physicians had the option of completing a one-page paper data form or an electronic “dotphrase” (DP) data form. Our hospital system uses an Epic®-based EMR. A feature of this system is the ability to use DPs to assist in medical information entry. A DP is a preset template that may be inserted into the EMR when the physician types a period followed by a code phrase (in this case “.ichstudy”). Once the study DP was inserted at the bottom of the electronic ED note, it prompted enrolling physicians to answer study questions. Investigators then extracted data directly from the EMR. Our primary outcomes of interest were the prevalence of DP data form use and rates of data entry errors.
Results: From 7/2009 through 8/2010, 883 patients were enrolled. DP data forms were used in 288 (32.6%; 95% CI 29.5, 35.7%) cases and paper data forms in 595 (67.4%; 95% CI 64.3, 70.5%). The prevalence of DP data form use at the respective study centers was 11%, 16%, 18%, 31% and 85%. Sixty-six (43.7 %; 95% CI 35.8, 51.6%) of 151 physicians enrolling patients utilized DP data entry at least once. Using multivariate analysis, we found no significant association between physician age, gender, or tenure and DP use. Data entry errors were more likely on paper forms (234/595, 39.3%; 95% CI 35.4, 43.3%) than DP data forms (19/288, 6.6%; 95% CI 3.7, 9.5%), difference in error rates 32.7% (95% CI 27.9, 37.6%, p<0.001).
Conclusion: DP data collection is a feasible means of study data collection. DP data forms maintain all study data within the secure EMR environment obviating the need to maintain and collect paper data forms. This innovation was embraced by many of our emergency physicians. We found lower data entry error rates with DP data forms compared to paper forms.

Presentation Number:595

Background: Inadequate randomization, allocation concealment and blinding can inflate effect sizes in both human and animal studies. These methodological limitations might in part explain some of the discrepancy between promising results in animal models and non-significant results in human trials. Whereas blinding is not always possible, in clinical or animal studies, true randomization with allocation concealment is always possible, and may be as important in minimizing bias.
Objectives: To determine the frequency with which published emergency medicine (EM) animal research studies report randomization, specific randomization methods, allocation concealment and blinding of interventions and measurements, and to estimate whether these have changed over time.
Methods: All EM animal research publications from 1/2000 through 12/2009 in Ann Emerg Med and Acad Emerg Med were reviewed by 2 trained investigators for a statement regarding randomization, and specific descriptions of randomization methods, allocation concealment, blinding of intervention and blinding of measurements, when possible. Raw initial agreement was calculated and differences were settled by consensus. The first (period 1=2000-2004) and second (period 2=2005-2009) 5-year periods were compared with 95% confidence intervals.
Results: Of 117 EM animal research studies, 109 were appropriate for review because they involved intervention in at least 2 groups. Blinding of interventions and measurements were not considered possible in 37% and 3%, respectively. Significant differences between period 1 and 2 were absent, although there was a trend towards less blinding of interventions and more blinding of measurements. Raw agreement was 91%.
Conclusion: Although randomization is mentioned in the majority of studies, allocation concealment and blinding remain underutilized in EM animal research. We did not compare outcomes between blinded and non-blinded, randomized and non-randomized studies, because of small sample size. This review fails to demonstrate significant improvement over time in these methodological limitations in EM animal research publications. Journals might consider requiring authors to explicitly describe their randomization, allocation and blinding methods.

Presentation Number:596

Background: The Injury Severity Score (ISS) is a useful measure of anatomic injury severity associated with mortality, surgical intervention and need for intensive care. However, ISS requires manual chart abstraction and is generally unavailable for injured patients cared for in non-trauma centers.
Objectives: To evaluate and validate a STATA mapping function that derives ISS from International Classification of Diseases, ninth revision (ICD9) diagnosis codes.
Methods: This was a multi-site validation study of injured patients transported by EMS in 7 regions and captured in 11 trauma registries representing 31 trauma centers (Level I-IV) in the Western United States from 2006-2008. ISS was calculated for each patient by a trauma registrar at each trauma center, which is considered the gold standard for ISS calculation. We used probabilistic linkage to match trauma registry records to administrative data files (emergency department and state discharge data). We then used the ICDPIC 3.0 module for STATA to derive ISS from ICD9 diagnosis codes for each patient. We compared registry-coded ISS to ICDPIC-derived ISS using descriptive statistics, Bland-Altman plots and Cohen’s Kappa measure of agreement. We also assessed agreement after collapsing ISS into six previously defined severity categories.
Results: 25,943 injured patients had matched registry-ISS and ICDPIC-ISS available for comparison. The median ISS values for each technique were: registry-ISS 9 (IQR 4-14) and ICDPIC-ISS 6 (IQR 4-13). There was a mean ISS difference of 1.8 (95% CI 1.8-1.9) with Bland-Altman 95% agreement limits -10.9 to 14.6, suggesting that the ICD9 mapping procedure slightly under-estimates ISS coded by trauma registrars. The two methods were in exact agreement for 58.2% (95% CI 57.6-58.8%) of patients; Cohen’s kappa was 0.55. When collapsed across six categories, the two methods generated a same-category ISS in 76.0% of patients and same or adjacent category in 95.2% of patients. Kappa between categories was 0.64.
Conclusion: When compared to ISS coded by chart review, ISS derived from ICD9 codes showed moderate to substantial agreement in our sample, though tended to slightly under-estimate injury severity. Such scores may be useful in research when hand-coded scores are unavailable.

Presentation Number:597

Background: Emergency Medical Services is an important element of acute stroke care. However, evaluation of prehospital stroke care is limited by lack of exchange of patient outcome data between hospitals and emergency medical services (EMS) agencies.
Objectives: In this study, we describe and demonstrate the feasibility of linking county wide patient level ambulance data with emergency department (EDD) and patient discharge data (PDD) using a probabilistic matching algorithm.
Methods: Probabilistic linkage was used to match county-wide ambulance data from 2005-2007 to hospital (EDD and PDD) records with a final ICD -9 diagnosis of stroke (430-436). The linkage model was based on the patient’s transport/admission date, date of birth, race, sex, county of residence, and destination hospital. Probabilistic linkage was performed using LinkSolv version 8.29746 which calculates the probability that a pair of records is a true match based on agreement/disagreement patterns of the linkage variables. Pairs of records with a match probability of 0.8 or higher were considered true matches. All other pairs were false matches and rejected.
Results: During 2005 - 2007 there were 310,731 patients transported to a facility in county and 34,785 hospital records with a diagnosis of stroke. Using the linkage algorithm we identified 11,473 (33%) matches with EMS records. Linkage rates increased each year with 30%, 34%, and 36% of hospital patients matching EMS record for 2005, 2006, and 2007 respectively. The median match probability was 0.993 and the IQR was 0.974 to 0.9996. By taking the compliment of the match probability we estimate our linked sample to include 255 (2%) false matches. Date of treatment/admission and the patient’s sex were observed to be the most reliable, disagreeing on less than one percent (1%) of all matched pairs. Patient’s zip code was the least reliable, disagreeing on one third of matched pairs.
Conclusion: Probabilistic matching can be used to create a comprehensive patient care record and evaluate prehospital care of stroke patients and also perform outcomes-based research.

Presentation Number:599

Background: The Bayesian approach to disease diagnosis in the ED is facilitated by the use of likelihood ratios (LR’s) to evaluate diagnostic tests. However, an extreme LR may seem overly optimistic if only a few patients have the corresponding test result, and application to serial non-independent tests is problematic.
Objectives: An extension of the LR, the average absolute likelihood ratio (AALR), was developed to assess the average change in the odds of disease that can be expected from a test, or series of tests, and an example of its use to diagnose wide QRS complex tachycardia (WCT) is provided.
Methods: Results from two retrospective multicenter case series were used to assess the utility of QRS duration and axis to assess for ventricular tachycardia (VT) in patients with undifferentiated regular sustained WCT. Serial patients with heart rate (HR) >120 beats per minute and QRS duration > 120 milliseconds (msec) were included. The final tachydysrhythmia diagnosis was determined by a number of methods independent of the ECG.
The AALR is defined as:
AALR = 1/N_Total[Σ (N_i *LR_i) (for LR>1) + Σ (N_k / LR_k) (for LR<1)]
where LR_i and LR_k are the interval LR’s, and N_i and N_k are the number of patients with test results within the corresponding intervals. ROC curves were constructed, and interval LR’s and AALR’s were calculated for the QRS duration and axis tests individually, and when applied together. Confidence intervals were bootstrapped with 10,000 replications using the R boot package.
Results: 187 patients were included: 95 with supraventricular tachycardia (SVT) and 92 with VT. Optimal QRS intervals (msec) for distinguishing VT from SVT were: QRS ≤ 130, 130 < QRS < 160, and QRS ≥ 160. QRS axis results were dichotomized to upward right axis (181-270 degrees) or not (-89 to 180 degrees). Results are listed in the table.
Conclusion: Application of the QRS interval and axis tests together for patients with wide QRS complex tachycardia changes the odds of ventricular tachycardia, on average, by a factor of 3.5 (95% CI 2.4 to 6.2), and this is mildly improved over the QRS duration test alone. Both a strength and weakness of the AALR is its dependence on the pretest probability of disease. The AALR may be helpful for clinicians and researchers to evaluate and compare diagnostic testing approaches, particularly when strategies with serial non-independent tests are considered.

Presentation Number:600

Background: Despite their wide use, randomized control trials (RCTs) continue to experience delays and problems with recruitment of clinicians and patients. Reduced sample size reduces the statistical power of the study, one of the main reasons trials are abandoned. There have been many studies on barriers to enrollment for a range of situations, but there is little to no data on barriers to enrolling ED patients with advanced illness in clinical trials.
Objectives: To identify barriers to the enrollment of ED patients with advanced illness who meet inclusion criteria for a clinical trial.
Methods: We prospectively tracked factors that affected patient accrual into an RCT of palliative care consultation for adults with metastatic solid tumors at an urban, academic ED located within a tertiary care referral center. Field notes were grouped into barrier categories and then quantified when possible. Patient demographics for those who did and did not enroll were extracted from the medical record and quantified. Patients that did not meet inclusion criteria for the study (e.g., cognitive impairment) were excluded from the analysis.
Results: Attempts were made to enroll 42 eligible patients in the study, and 23 were successfully enrolled (55% enrollment rate). Barriers to enrollment were deduced from the field notes and placed into the following categories from most to least common: patient refusal (6); diagnostic uncertainty regarding cancer stage (4); severity of symptoms preclude participation (4); patient unaware of illness or stage (3); and family refusal (2).
Conclusion: Patients, families, and diagnostic uncertainty are barriers to enrolling ED patients with advanced illness in clinical trials. It is unclear whether these barriers are generalizable to other study sites and disease processes other than cancer.

Presentation Number:601

Background: The delivery of palliative care, such as establishing goals of care, is not standard of care in most EDs. Preliminary data suggest that ED-initiated palliative care consultation can decrease hospital length of stay and costs for select patients with advanced illness.
Objectives: To test the feasibility of enrolling and randomizing ED patients with solid metastatic tumors to ED-initiated palliative care consultation versus care as usual in a crowded, urban ED.
Methods: DESIGN - Single-blind randomized controlled trial of ED-initiated palliative care consultation for patients with solid metastatic tumors versus usual care. SETTING - Urban, academic ED at a tertiary care referral center. PARTICIPANTS - Adult patients with solid metastatic tumors who were able to pass a cognitive screen, had never been seen by the palliative care service, spoke English or Spanish, and presented to the ED beginning in June 2011 between 9am and 5pm Monday-Friday met eligibility criteria; eligible patients were approached and enrolled between 9am and 5pm Monday-Friday in the ED and randomized via balanced block randomization to the intervention or control group. INTERVENTIONS/OBSERVATIONS - Intervention patients received a comprehensive palliative care consultation by the inpatient team, including an assessment of symptoms, spiritual/social needs, and goals of care. Outcomes include hospital length of stay, direct costs, and ED revisits and re-hospitalization at 12 weeks.
Results: 23 patients were enrolled and randomized. 28 patients were approached but excluded, and 19 additional patients met inclusion criteria but chose not to participate. With the exception of one patient who enrolled and whose family subsequently declined participation, all patients randomized to the intervention group received palliative care consultation. No patients have been lost to follow-up. Median hospital length of stay and direct costs of hospitalization for the intervention and control group were 8 versus 5 days, and $7782 versus $4029, respectively. At 12 weeks, the intervention and control group made 0.42 versus 0.55 ED revisits, and were re-hospitalized 0.50 versus 0.73 times.
Conclusion: Enrolling and randomizing ED patients with advanced cancer to early palliative care consultation versus care as usual is feasible in a crowded, urban, academic ED.

Presentation Number:602

Background: Successful communication between physicians and nurses can minimize errors and improve patient safety in providing patient care, however little opportunity is present in the professional practice curricula to engage in teamwork and collaboration. In this study, Indiana University (IU) nursing students, medical students, and EM residents participated in high-fidelity, mannequin-based interprofessional simulations.

Objectives: The purpose of this study is to determine whether interprofessional simulation improves self-confidence in interprofessional communication, attitudes toward reporting medical errors, and barriers to nurse-physician communication.

Methods: Senior IU EM residents (R3) and senior IU medical students (MS4) are required to participate in separate simulation sessions throughout the academic year. Senior Bachelors of Science nursing students (BSN) were recruited on a volunteer basis to participate. A pre and post survey was administered to all participants. The survey was a collaborative effort by MD and RN simulation faculty experts to include 3 sections: self-confidence in communication, error reporting, and barriers to physician-nurse communication.
Results: There were 168 simulation surveys completed (12 R3, 119 MS4, 37 BSN). The testing tool was found to be reliable with Chronbach-alphas from 0.69-0.90. Following the sessions, overall mean confidence in interprofessional communication improved in all 3 groups: R3 (p=0.078), MS4 (p=0.262), and BSN (p=0.026). Both medical students and nursing students had a statistically significant increase in attitude towards error reporting (p<0.001) while residents showed an increase (p=0.455). Barriers to communication showed a variable and not statistically significant difference between pre and post surveys in all groups.
Conclusion: Self-confidence in interprofessional communication and error reporting was improved through interprofessional simulation sessions, with statistically significant improvement in the senior medical student and senior nursing student groups. Given the divergence of nursing and physician undergraduate medical education over the past few decades, interprofessional simulation may provide an opportunity for both groups to improve communication prior to clinical practice.

Presentation Number:603

Background: In the past decade, there has been increased recognition of the importance of teaching end of life and palliative care skills to Emergency Physicians. Efforts to improve palliative care training in medical trainees with a classroom-based educational intervention yielded no significant effect on residents’ knowledge.
Objectives: The purpose of this study was to evaluate the use of a high fidelity mannequin bedside simulation scenario followed by a debriefing session as a tool to improve medical student knowledge of palliative care techniques.
Methods: Third year medical students participating in a 12-week simulation curriculum during a Surgery/Emergency Medicine/Anesthesia clerkship were eligible for the study. All students were administered a pretest to evaluate their baseline knowledge of palliative care and randomized to a control or intervention group. During week 3 or 4, students in the intervention group participated in and observed two end of life scenarios. Following the scenarios, a faculty debriefer trained in palliative care addressed critical actions in each scenario. During week 10, all students received a post-test to evaluate for improvement in knowledge. The pretest and post-test consisted of 12 questions addressing: prognostication, symptom control and the Medicare Hospice Benefit. Students were de-identified and pre- and post-tests were graded by a blinded scorer.
Results: From Jan-Dec 2011, 70 students were included in the study and 5 were excluded due to incomplete data. The mean score on the pretest for the intervention group was 3.16, and for the control group was 3.45 (p=0.90). The mean scores on the post-test were 4.54 for the intervention group and 4.3 for the control group (p=0.15) with no significant difference between the two groups. Both groups demonstrated a significant improvement in post-test scores (p<0.01). Scores on the 12-item test reached a mean of 37.8% correct, and a high score of 66.7% correct.
Conclusion: Simulation/debriefing alone does not account for the improvement in students’ scores from the pretest to post-test. Other experiences while on rotation may have been more influential. Further studies to determine whether the intervention affected student comfort level and feeling of preparedness dealing with end of life issues would be beneficial. Limitations of the study include difficulty level of the tests.

Presentation Number:604

Background: Simulation has become accepted as a valuable tool to the emergency medicine educator. Despite its recognized role, its use remains heterogeneous.
Objectives: We sought to determine if there is any consensus on simulation scenarios that belong in emergency medicine simulation curricula.
Methods: An online survey was distributed to members of the College of Residency Directors mailing list. Information regarding individual’s academic title, involvement in resident education, and use of simulation in resident education was obtained. Additionally, respondents were asked to provide their opinion on the most important simulation scenarios in free text. Results were then analyzed and scenario themes were extracted from free text response.
Results: 101 survey responses were analyzed. 95% of respondents noted that high fidelity mannequin based simulation is being used in resident education. Individuals responding identified themselves as Program Directors (23%), Associate Program Directors (8%), Assistant Program Directors (9%), Simulation Directors (10%), and other faculty. Of the respondents, 75% reported that their program has a formal simulation curriculum. 43% of individuals reported that their simulation curriculum was developed based on an individual’s decision, while 38% reported that it was the result of program consensus. Respondents reported that the most important simulation scenarios were those that incorporate protocols of Advanced Cardiac Life Support, Pediatric Advanced Cardiac Life Support, and Advanced Traumatic Life Support. Additionally, Teamwork/Leadership, Acute Coronary Syndrome, Sepsis, Respiratory Failure, and OB/pregnancy related disorders were identified as having high priority.
Conclusion: The results indicate that educators identify the most important scenarios as protocol based simulations. Respondents also suggested that scenarios of very common Emergency Department presentations bear a great deal of importance. Emergency Medicine educators assign priority to simulations involving professionalism and communication. Finally, many respondents noted that they use simulation to teach the presentation and management of rare or less frequent, but important disease processes. The identification of these scenarios would suggest that educators find simulation useful for filling in “gaps” in resident education.

Presentation Number:076

Background: Surrogate consent for treatment and research options on behalf of mentally incapacitated patients with acute stroke is currently the standard of practice in emergency departments across the nation. Many studies have, however, indicated the failure of surrogates to accurately predict patient preferences in a variety of other clinical settings.
Objectives: This study is designed to investigate the hypothesis that such a failure extends to the acute stroke setting as well--that significant discrepancies exist between surrogate decisions and the preferences of the patients they represent.
Methods: We performed a cross-sectional verbal survey of 200 patients in the University of Michigan ED without stroke, and the family member, friend, or significant other who had accompanied the patient, resulting in a total enrollment of 400. The patient was presented with 5 scenarios for treatment decisions in the event of an acute stroke or cardiac arrest. After each scenario, he or she indicated the likelihood of consenting to the treatment/protocol for each scenario. The same procedure was then performed separately on the surrogate.
Results: Overall, surrogates predicted patients’ treatment preferences with 80.2% accuracy. Patient/surrogate agreement for scenarios 1, 2 and 5 was 96%, 87% and 95% respectively. Scenarios 3 and 4--regarding a standard pharmacotherapy RCT and an adaptive RCT--gave rise to the vast majority of disagreement between patients and surrogates. In scenario 3, 69.5% of pairs refused the trial while 4.5% of pairs consented to the trial, resulting in an agreement rate of 74%. The adaptive clinical trial (scenario 4) represented that lowest rate of agreement at 49%.
Conclusion: The accuracy with which surrogates were able to predict patient preferences was highly dependent on the type of treatment being offered. Our study found substantially more agreement for standard treatments (scenarios 1 and 5) than for experimental/research protocols (scenarios 2, 3 and 4).The adaptive clinical trial was more acceptable than standard RCT, although there was substantially more patient/surrogate disagreement regarding participation in the adaptive trial, potentially due to the increased complexity of the design. Further research is needed into optimal consent approaches in time sensitive, high stakes diseases such as stroke and other acute neurological conditions.