The Proceedings of Medical Quality 2014.

539459

research-article2014

AJMXXX10.1177/1062860614539459American Journal of Medical QualityThe Proceedings of Medical Quality 2014

Article American Journal of Medical Quality 2014, Vol. 29(3S) 5S–27S © 2014 by the American College of Medical Quality Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1062860614539459 ajmq.sagepub.com

The Proceedings of Medical Quality 2014

The American College of Medical Quality met for the annual meeting from March 27 to 30, 2014, with a program including presentation of the abstracts published in this supplement. The College is pleased that the annual meeting could be the forum for these presentations, representing current research on a wide range of areas in quality and patient safety. Special thanks to all the researchers who submitted and presented their work and for SAGE Publications for making this supplement possible. Henry Johnson, MD, MPH Chair, Medical Quality 2014 Scientific Program Committee Vice President and Medical Director Midas+ Solutions, a Xerox Company

Oral Presentations Friday, March 28, 2014 (12-1 pm) Moderated by David Nash, MD, FACMQ Editor-in-Chief, American Journal of Medical Quality

acquired included demographic, perioperative, hemodynamic, and outcome data. Multiple outcome measures including Pneumonia (PNEU) and Ventilator > 48 hours (V48) were obtained and entered into a central database from which a risk-adjusted Semiannual Report (SAR) was generated. Specific interventions included enforcement of protocols, early extubation, nutrition, adherence to sedation holidays, and VAP bundles. Data are reported as observed to expected odds ratio (O:E) with 1 being performance as expected. Results: Three SARs were generated: July2011, January2012, and July2012. In July2011, the O:E for PNEU was 1.8 (5/368, 1.36%) while V48 was 2.54 (7/368, 1.9%). By January2012, the O:E for PNEU was 1.48 (13/926, 1.4%) while V48 declined to 1.11 (10/925, 1.08%). By July2012 the O:E for PNEU was 1.25 (13/1081, 1.2%; P < .05) and V48 was 1.04 (12/1080, 1.11%; P < .05). Conclusion: Progressive improvements in PNEU and V48 were observed. Given an estimated annual volume of 10 000 cases per year with a cost of $22 097 per episode of PNEU and $27 654 per V48, a projected 65 avoided episodes of PNEU and 105 avoided episodes of V48 could be realized with potential savings exceeding $4 000 000.

Abstract 1 Using National Surgical Quality Improvement Project (Nsqip) Data to Decrease Ventilator Days and Pneumonia in a Surgical ICU John McNelis1, Roseann Grandelli2, Maureen Keegan2 1

Jacobi Medical Center, Bronx, NY; 2Winthrop University Hospital, Mineola, NY

Introduction: National Surgical Quality Improvement Project (NSQIP) is a risk-adjusted database tracking surgical outcomes. NSQIP has been demonstrated to decrease complications, expenses, and mortality. In the study institution, a high rate of nosocomial pneumonias was observed in surgical patients. The authors utilized NSQIP to track effectiveness of interventions made. Methods: NSQIP collected preoperative through 30-day postoperative data on surgical patients. Data

July 2011 SAR

Jan 2012 SAR

July 2012 SAR

Estimaged Savings

1.80 2.54

1.48 1.11

1.25 1.04

$1 436 305 $2 903 655

PNEU O:E V48 O:E

Abstract 2 Patient Throughput and Operational Outcomes: Understanding the Relationship in the Context of a Collaborative Improvement Initiative Aline Holmes, Mary Ditri New Jersey Hospital Association, Princeton, NJ

Purpose: Inefficiencies resulting in misuse of acute care beds, reduced productivity, and excessive waste of resources complicate hospitals’ abilities to serve the growing number of patients in need of care.1 Leaders

Downloaded from ajm.sagepub.com at The University of Hong Kong Libraries on May 8, 2015

6S

American Journal of Medical Quality 29(3S)

face having to staff appropriately regardless of census, but variability in admissions, operating room scheduling, and treatment times can make this a difficult task. That same unpredictability creates work environments that risk unsafe practices and increased medical errors. Absent subsidies to fund local improvement models, hospital leaders are faced with identifying efficiencies and system improvements that will act as the foundation for improved performance.2 Research suggests that the way hospitals manage patient flow has been linked to a rise in nursing stress, overcrowding, medical errors, and increased mortality,3 as well as delays, cancellations, and avoidable readmissions.4 The purpose of this explanatory case study is 2-fold. It examines how a 15-month collaborative to smooth patient flow influenced patient throughput in 16 New Jersey hospitals. It then examines how improving patient throughput influences quality, safety, satisfaction, and operating margins. Methodology: This explanatory case study design examines how a collaborative approach to smoothing patient throughput in 16 New Jersey hospitals influences quality, safety, satisfaction, and operating margins. The study has time-specific boundaries and is not based on random sampling. Results: Individual hospital results are significant and include decreased mortality rates, reduced PACU boarding, decreased ALOS, decreased discharge wait times, improved staff satisfaction, and increased revenue based on improved capacity. Predictive modeling suggests that collectively the hospitals will achieve significant outcomes including 4% to 31% decreased risk in mortality for select patient populations, 3% to 47% decreased ALOS for select patient populations, 21% to 85% decrease in emergency department boarding time for medical admissions, 11 800 to 17 300 annual increase in patients treated without the addition of operating rooms or inpatient beds, 20 350 to 21 200 annual increase in emergency department visits, and $50 to $70+ million return on investment projected annual contribution margin improvement. Conclusions: The goal of this study was to extend the body of knowledge on how unpredictable patient flow influences a number of performance measures in a hospital environment. The hospitals prioritized quality and improved systems that influenced outcomes. The focus of the collaborative was to design systems that were efficient, improving their use of capital and human resources. Patient outcomes were improved, satisfaction increased, and financial performance was enhanced. The collaborative design of this initiative is a study in how large-scale improvements can be achieved across hospital systems. References 1. Institute of Medicine. Best Care at Lower Cost: The Path to Continuously Learning Health Care in America. Washington, DC: National Academies Press; 2012.

2. Litvak E, Laskowski-Jones L. Nurse staffing, hospital operations, care quality, and common sense. Nursing. 2011;41(8):6-7. 3. Needleman J, Buerhaus P, Pankratz VS, Leibson CL, Stevens SR, Harris M. Nurse staffing and inpatient hospital mortality. N Engl J Med. 2011;364:1037-1045. 4. Baker D, Pronovost P, Morlock L, Geocadin R, Holzmueller C. Patient flow variability and unplanned readmissions to an intensive care unit. Crit Care Med. 2009;37:2882-2887.

Abstract 3 Improved Prenatal and Postpartum Visit Compliance and Breastfeeding Rates Are Found in Pregnancy Texting Program Participants Mary Mason, Amy Poole-Yaeger, Marie McNeely Centene Corporation, Saint Louis, MO

Purpose: The Start Smart For Your Baby (SSFB) Texting Program, a component of our comprehensive SSFB Care Management program, aims to improve maternal and infant health by providing timely heath information, resources, and reminders to members during pregnancy through 6 months postpartum. Specifically, the program’s goals are to improve quality of care by increasing prenatal and postpartum visit compliance, as well as encourage breastfeeding since rates are substantially reduced in low-income mothers. Methodology: We evaluated 7882 Medicaid members across 13 health plans who participated in the SSFB Texting program and have delivered. HEDIS metrics for Frequency of Ongoing Prenatal Care and Timeliness of Prenatal and Postpartum Care were measured for the Texting members, compared to the overall HEDIS rates for all members in participating health plans. Additionally, a subset of 870 members participated in a survey to track breastfeeding. We evaluated breastfeeding and exclusivity in our texting survey members 21 days postdelivery, compared to a group of 1727 members with breastfeeding reported in a postpartum assessment 21 to 30 days postdelivery. Results: Compared to the overall health plan outcomes, HEDIS Frequency of Ongoing Prenatal Care (81% to 100%) rates were 6.9% higher (χ2 = 72.5, P < .001), Timeliness of Prenatal Care rates were 4.4% higher (χ2 = 37.9, P < .001), and Timeliness of Postpartum Care rates were 10.2% higher (χ2 = 123.0, P < .001) for those in the SSFB Texting Program at least 90 days during the prenatal and postpartum periods, respectively. Members with longer participation duration had higher rates for all metrics. Also, 23.0% more members in our program maintained some amount of breastfeeding (89.5% rate in survey participants) at 21 days postdelivery (χ2 = 158.3,


7S

The Proceedings of Medical Quality 2014 P < .001), compared to the postpartum assessment group (66.6%). Additionally, 14.7% more SSFB Texting Survey members still breastfed exclusively (49.7% rate in survey members) 21 days postdelivery (χ2 = 51.4, P < .001), compared to postpartum assessment members (35.0%). Conclusions: Members in the SSFB Texting Program had improved HEDIS rates, as well as higher breastfeeding maintenance and increased rates of exclusive breastfeeding. This program has been well received by our members, and 93% of respondents to a rating question reported the texts were somewhat helpful or very helpful. The SSFB Texting Program is an innovative approach that allows us to engage members and convey important and relevant information to influence health behaviors and improve quality of care.

Abstract 4 Health Outcome Disparities and Utilization Among US Children With Abdominal Pain at Major Emergency Departments Louise Wang1, Corinna Haberland1, Cary Thurm2, Jay Bhattacharya1, K. T. Park1

Stanford University, Stanford, CA; 2Children’s Hospital Association, Overland Park, KS

(69.2%) presented to the ED initially. Household median income, race, and hospital region differed significantly among those with functional or organic abdominal pain (P < .001-.011). Black, Hispanic, and children from households with low median incomes (≤$32 214/year) presenting to the ED with a primary complaint of abdominal pain had lower odds of being hospitalized (adjusted odds ratio [aOR] = 0.56, 0.75, 0.85, respectively; P < .001), with even lower odds for Black children of low income (aOR = 0.45, P < .001). Black children were less likely to receive computed tomography, X-ray, and ultrasound scans for any type of abdominal pain (P < .001) including appendicitis, but they had higher odds of appendicitis with peritonitis, including perforation (aOR = 1.42; P < .001) and higher ICU admissions (aOR = 1.92, P < .001). They had 9% lower likelihood of being discharged home for their appendicitis, which was compounded among Black patients of low income (adjusted hazard ratio = 0.85, P < .001). Conclusions: Disparities exist in the utilization of imaging and medical services in treating Black and Hispanic children and patients of lower socioeconomic status for abdominal pain in the ED. Future studies should characterize whether these disparities directly affect the increased risk of appendicitis with perforation in these groups.

1

Purpose: Each year in the United States, approximately 900 000 children ≤15 years of age present to the emergency department (ED) with abdominal pain, and appendicitis is the top surgical emergency among children. Evidence suggests that disparities of care involving inappropriate triage exist in ED settings. In addition, health care utilization and clinical care differences between abdominal pain patients of varying socioeconomic (SES) classes presenting to the ED are unknown. We would like to identify disparities in health care outcomes among children presenting to the ED with abdominal pain, based on race and SES, and to quantify the differences in utilization of emergency services and subsequent hospital admissions. Methodology: We developed multivariate regression models and Cox proportional hazard models to identify predictors for abdominal pain and to evaluate clinical care disparities for 4.2 million pediatric patient encounters and 2.7 million patients in the United States from 43 children’s hospitals from 2004 to 2011. Outcome data included hospital admissions, intensive care unit (ICU) admissions, appendicitis, appendicitis with perforation, surgical procedures, and utilization of imaging services. Explanatory data included patient demographic and socioeconomic information, hospital region, insurance status, comorbidities, and past medical history. Results: Of patient encounters with a diagnosis of nonspecific abdominal pain, approximately 2.9 million

Abstract 5 Perioperative Optimization of Senior Health (POSH): The Journey to a Collaborative Care Model Mamata Yanamadala, Shelley McDonald, Heather Whitson, Cornelia Poer, Julie Thacker, Heidi White, Mitchell Heflin Duke University, Durham, NC

Purpose: To develop a model of multidisciplinary and interprofessional collaborative perioperative care for high-risk older adults undergoing elective surgery. Methods: In 2011, a group of general surgeons partnered with geriatricians to develop a program to improve outcomes for high-risk older adults undergoing elective surgery. The Perioperative Optimization of Senior Health (POSH) group envisioned the ideal model of care, which streamlined the perioperative process from the existing complex, siloed system. They identified stakeholders and team members, defined best practices, and developed a data collection process. An iterative and incremental development process was utilized to implement the model through continual feedback and adaptation. This process involved integration of multiple providers across different professions, as well as integration of multiple units within the larger system. All aspects of the project underwent successive revisions, including model design,


8S


system requirements, scheduling requirements, desired data points, data collection templates, database design, process for reconciliation, and deployment. Results: Some of the major refinements and revisions made within a period of 2 years include the following: (a) streamlining process: preoperative anesthesia and geriatrics visits occur at the same location as do ancillary services (lab and EKG); (b) improving collaboration: face-to-face sign out between anesthesia and geriatric providers during patient evaluation; (c) enhancing patient-centeredness: engaging patient and family in discussion of goals and expectations for surgery; (d) optimizing communication: obtaining outside records for patients and communicating with primary care providers; and (e) enhancing postoperative care: postoperative in-patient geriatric consultation focusing on delirium prevention, medication, and pain management. Within 2 years, 77 patients have gone through this evolving structured partnership between physicians, nurses, nurse practitioners, PAs, and social workers allowing the patients to be followed from the ambulatory setting to the operating room to the inpatient wards while maintaining continuity of care. This process has resulted in the reduction of the median length of stay for referred patients to 4.20 days (±3.06) from 5.35 days (±7.62) for controls (P < .001). Also, 13% of referred patients were discharged to skilled nursing facilities versus 18.6% of controls. Conclusion: Implementation of a new model that focused on collaborative care for perioperative high-risk older adults required many revisions of existing processes sometimes occurring simultaneously before it could be implemented in a complex system. Constant feedback and subsequent modification of the process helped achieve meaningful patient care outcomes.

Poster Presentations

physical therapy. In an effort to reduce costs, health care payers, led by the Federal government, are encouraging patients to be discharged home rather than to an acute care rehabilitation facility, following uncomplicated lower extremity joint replacements. However, there is concern that discharge to home will result in an increased length of stay (LOS) and readmission rate compared to discharge to an inpatient rehabilitation facility. We analyzed all 1777 discharges following total hip replacement (THR) and total knee replacement (TKR) at our institution in fiscal year 2010 to determine their physician-specific discharge disposition. We then analyzed the effect of discharge disposition on physician-specific LOS and readmission rate. Discharge patterns following joint replacement varied widely between physicians at our institution. However, we could find only weak correlations between the cost of discharge and LOS or readmission rates. Patients discharged home did not have a greater LOS or readmission rate than those discharged to a rehabilitation center or skilled nursing facility. The large interphysician variance in discharge cost did not correlate to a difference in quality, as measured by the LOS and readmission rates, but does imply there is significant opportunity to modify physician discharge practices without affecting patient outcomes and the quality of care.

Abstract 7 The Effect of Discharge Disposition on Readmission Rates Following Total Joint Arthroplasty: An Analysis of 3533 Patients Nicholas Ramos, Raj Karia, Lorraine Hutzler, Aaron Brandt, James Slover, Joseph Bosco NYU Hospital for Joint Diseases, New York, NY

Friday, March 28, 2014 (8 am to 6:30 pm) (Abstracts 24-48 were submitted by students/residents/ fellows)

Abstract 6 Physician-Specific Correlation Between Discharge Disposition Cost, Readmission, and Length of Stay Following Total Hip and Knee Replacement: An Analysis of 1777 Cases Nicholas Ramos, Lorraine Hutzler, James Slover, Joseph Bosco NYU Hospital for Joint Diseases, New York, NY

Discharge to an acute rehabilitation facility is significantly more costly than discharge home with outpatient

Previous studies have demonstrated no significant difference in functional outcomes of patients discharged to a subacute setting versus home with health services after total joint replacement. These findings coupled with pressure to reduce health care costs and the implementation of a prospective payment system under Medicare have supported the use of home rehabilitation services and the trend toward earlier discharge after hospitalization. While the overall functional outcome of patients discharged to various settings has been studied, there is a relative dearth of investigation comparing postoperative complications and readmission rates between various discharge dispositions. Our study demonstrated total hip replacement patients discharged to an inpatient rehabilitation facility had a significantly higher rate of 30-day readmission as compared to those patients sent to a skilled nursing


9S

The Proceedings of Medical Quality 2014 facility or home with health services. This difference remained after statistically adjusting for age, sex, and comorbidity. Total knee replacement patients discharged to an inpatient rehabilitation facility trended toward a higher rate of 30-day readmissions, and multivariate regression analysis demonstrated a strong association between comorbidity and readmission. Patients discharged to rehabilitation facilities have a higher incidence of comorbidity, and this association could be responsible for their higher rate of readmission.

throughout the intervention with a trend toward overall decrease in mortality after the intervention (21.39% before vs 16.8% in year 3 of the intervention, P = .2155) although this did not achieve statistical significance. Conclusion: Improvement in emergency physician documentation rates of endotracheal tube position confirmation were associated with a trend toward lower patient mortality rates. This is consistent with a trend identified in a previous analysis of survival in a national registry.1

Abstract 8

Reference

Is Improvement In Emergency Physician Documentation Rates of Endotracheal Tube Position Confirmation Also Associated With Lower Patient Mortality Rates?

1. Phelan MP, Ornato JP, Peberdy MA, Hustey FM. Appropriate documentation of endotracheal tube placement confirmation and relationship to patient outcome from in-hospital cardiac arrest. Resuscitation. 2013;84:31-36.

Michael Phelan, Fredric Hustey, Stefanie Schrump, Stephen Meldon Cleveland Clinic, Emergency Services Institute, Cleveland, OH

Purpose: To assess whether improvement in emergency physician documentation rates of endotracheal tube position confirmation are associated with lower patient mortality rates. Methods: Secondary analysis of emergency department (ED) performance improvement (PI) database including ED airway registry. Setting: Urban, tertiary referral ED with annual census of 60 000 and affiliated Emergency Medicine (EM) residency program. Participants: ED physicians caring for patients undergoing prehospital or in-ED endotracheal tube placement. An airway registry was developed to monitor the care of all patients undergoing endotracheal intubation in study site ED or prehospital settings as part of a PI initiative. Initiatives included: Implementation of ED electronic health record (EHR) documentation templates for ETT placement/confirmation, emergency physician online educational module, and continuous performance feedback. After usual care phase, quality improvement interventions were initiated. Mortality was assessed via chart review, pre- and postintervention, by trained reviewers using standardized data collection forms. Proportions and risk ratios (RRs) with 95% confidence intervals (CIs) and P values are reported. Results: Between 2006 and 2010, a total of 961 patients were entered into the registry. Overall rates of documentation of appropriate methods of ETT position confirmation were higher after the intervention (557/758 [73.5%] vs 122/198 [61.6%], RR = 1.19; 95% CI = 1.06-1.34). Mortality rates steadily improved

Abstract 10 Intuitive and Interactive Workflow Toolkit to Strengthen Donabedian’s House of Medical Quality: Structure-Process-Outcome Ashok Singh1, Ranjit Singh2

1 Niagara Family Medicine Associates, Niagara Falls, NY, 2State University at Buffalo, Buffalo, NY

Context: Intervention with electronic health record (her), regarded as highly desirable, is complex. It has unexpected/unintended quality consequences. It requires process/workflow changes to be effective. Workflow is often cited as the No. 1 “pain point” by providers.1 Its redesign needs to be tailored to the practice context, current workflow, and the resources available, among other factors. The current methods, using esoteric charts and software,1,2 are alien to workers lacking the required expertise to benefit from them. Objective: Design and use an innovative, pictorial, and interactive workflow toolkit to strengthen Donabedian’s3 House of Medical Quality: StructureProcess-Outcome. The tool should aid assessment and redesign of process/workflow for achieving staff comfort, efficiency, communication, organization, productivity, effectiveness, safety, and satisfaction of patients and staff. Design: The toolkit,4 founded on “Improvement Science,” includes orientation presentation, 220 magnetic pictorial icons of 40 designs, dry-erase magnetic board, and animated visual aids for reducing cognitive and emotive biases. Setting: A Beacon Community of the Office of the National Coordinator of HIT. Participants: 20 Practice Enhancement Associates of the Beacon Community, employed as Health Extenders.


10S


Intervention: Motivational presentation reviewing the background was followed by explanations of What, Why, Who, When, Where of workflow. Use of the whole toolkit was illustrated with example scenarios with active input from participants. Outcome Measures: Qualitative feedback from all the participants and the Beacon officials. Results: Pilot tests were carried out in Western New York and Colorado. In addition, the toolkit was presented at the NAPCRG’11 meeting to solicit attendees’ responses as well as to the officers of the ONC. Qualitative feedback from all was extremely positive and enthusiastic. The respondents recommended that the toolkit be disseminated widely. It was used successfully by some for improvements in processes outside the EHR context and to help design organizational structures. Conclusions: The outstanding success with the kit is a compelling reason that this be disseminated at the MQ2014 meeting with full presentation of the whole workflow kit. References 1. Agency for Healthcare Research and Quality. Workflow assessment for health IT toolkit. http://healthit.ahrq. gov/portal/server.pt/community/health_it_tools_and_ resources/919/workflow_assessment_for_health_it_toolkit/27865. Accessed May 30, 2014. 2. Van der Aalst WMP, ter Hofstede AHM. YAWL: yet another workflow language. Information Systems. 2005;30(4):245275. 3. Donabedian A. Evaluating the quality of medical care. Milbank Q. 2005;83(4):691-729. 4. Singh R, Singh A, Singh DR, Singh G. Innovative workflow and process improvement toolkit to ease and enrich meaningful use of health information technology. Adv Med Educ Pract. 2013;4:231-236.

Abstract 11 Changes in Improvement in Oral Medication Management Associated With the National Home Health Quality Initiative Janet Lynch, YingHua Sun, Eve Esslinger, Jill Manna, Stephanie Snider West Virginia Medical Institute & Quality Insights, Richmond, VA

Purpose: To examine the impact of the national Home Health Quality Improvement (HHQI) Initiative on Improvement in Oral Medication Management in home care patients. Methodology: We used an observational pretest– posttest design with intervention and comparison

groups to examine the impact of HHQI on Medicarecertified home health agencies’ (HHAs’) change in improvement in oral medication management. Improvement in oral medication management was defined as the percentage of agency patients who were better able to manage their oral medications at the end of a home health episode than at the beginning. The data source was OASIS-C collected as a routine function of care. All Medicare-certified HHAs with 20 or more discharges during baseline and remeasurement were divided into intervention (enrollees) and comparison (nonenrollees) groups based on enrollment in HHQI prior to June 1, 2013. Baseline measures of improvement in oral medication management were taken from October 1, 2012, to March 31, 2013. Educational interventions were distributed as Web downloads and through other communications beginning in April 2013. Available remeasurement results were obtained for July 1, 2013, to Nov 30, 2013. We calculated baseline and remeasurement average improvement in oral medication management, weighted by the number of agency discharges, for both intervention and comparison groups. We used the General Linear Model (SAS 9.2 Proc GLM), weighted by the number of baseline agency discharges, to test the difference between enrollees and nonenrollees on change in improvement in oral medication management. In a second similar analysis, we compared only enrolled agencies that downloaded intervention materials to nonenrollees. Because age is related to improvement in oral medication management, we adjusted our analyses for change in the average age of agency patients. Results: A total of 4064 HHQI enrollees and 2790 nonenrollees met inclusion criteria. Of the enrollees, 294 downloaded educational materials prior to June 1, 2013. Subtracting baseline from remeasurement, the change in average improvement in oral medication management was 1.448% (53.230% − 51.782%) for enrollees, 1.615% (56.198% − 54.538%) for enrollees that downloaded educational materials, and −0.194% (44.654% − 44.848%) for nonenrollees. The average change in improvement in oral medication management was greater among HHQI enrollees compared to nonenrollees (P < .0001) and among enrolled downloaders compared to nonenrollees (P = .0058) and remained statistically significant when adjusted for changes in the average age of agency patients, P = .0004 and P = .0113, respectively. We applied the 1.448% improvement among enrollee agencies to the number of episodes of care in these agencies at remeasurement (1 002 729) to derive a point estimate of 14 520 episodes of care for which patients were safer through improved medication management.


11S

The Proceedings of Medical Quality 2014 Conclusions: HHQI enrollees, who tended to have higher scores on improvement in oral medication management, had a positive and statistically significant change in this measure when compared to nonenrollees.

Abstract 13

Abstract 12

Mamata Yanamadala, Lisa CriscioneSchreiber, Mitchell Heflin, Bimal Shah

Electronic Medical Record Documentation Guidance With Performance Feedback for Physicians Shawna Cain, Deborah Weidner Natchaug Hospital, Hartford HealthCare, Mansfield Center, CT

The Joint Commission Hospital Based Inpatient Psychiatric (HBIP) core measures include the tracking of patients discharged from a hospital-based inpatient psychiatric setting on 2 or more antipsychotic medications (HBIPS 4) and with appropriate justification (HBIPS 5). The Joint Commission provides the rationale that antipsychotic polypharmacy can lead to greater side effects, often without improving clinical outcomes; that APA practice guidelines recommend the use of a second antipsychotic only after multiple trials of a single antipsychotic have proven inadequate; and that randomly controlled trials do not support the use of multiple antipsychotic medications. They further state that because inpatient lengths of stay have decreased, some patients may continue on multiple antipsychotics on discharge, making communication with the provider in the next level of care paramount. This measure’s performance at our inpatient psychiatric units averaged 78.8% between 2009 and 2012. With the introduction of an electronic medical record (eMR), we sought to automate the documentation of the justification of multiple antipsychotics on discharge when applicable. Documentation screens were redesigned to require concise recording of 1 of the 3 acceptable justifications (history of a minimum of 3 failed multiple trials of monotherapy, recommended plan to taper to monotherapy due to previous use of multiple antipsychotic medications, or documentation of a crosstaper in progress at the time of discharge, and last, augmentation of Clozapine). In addition, management-level reports were designed to provide real-time feedback for process improvement. Monthly data reporting was incorporated into multiple quality committees, including Inpatient Medical Staff, Medical Executive Committee, Quality Council, and the Quality and Credentialing Subcommittee of the Board of Directors. Since the implementation of the eMR process, HBIP 5 rates improved to 100%.

Faculty Development Program in Health Care Quality Improvement

Duke University, Durham, NC

Purpose: To evaluate the effectiveness of a health care quality improvement (QI) course. Methodology: A curriculum was designed for interested faculty to provide a structured background in QI methods, relevant literature readings, pertinent case studies, and independent project generation. The didactic portions were delivered as monthly 1.5-hour sessions over 6 months. Each session included a 20-minute presentation covering key concepts in QI methodology, followed by small group activities to facilitate working on learners’ QI projects. Each faculty member was required to identify, lead, and implement a QI project within their clinical setting. Faculty met individually with mentors (authors MY, BS) twice during the 6 months, with additional advice and feedback offered through email. Faculty also presented their projects at the end of the course to their coparticipants, course directors, and department leadership. A pre–post survey measured changes in selfassessed competence course objectives course. QI knowledge was assessed using the validated Quality Improvement Knowledge Application Tool (QIKAT). A program evaluation survey designed to gather formative feedback from the faculty was analyzed. Results: Faculty had a wide range of ages, clinical settings, academic ranks, and experience with QI projects. Aggregate attendance at these sessions was 94%. A statistically significant difference in knowledge was seen between the pre- and post-QIKAT scores (maximum score = 15) (pre-mean = 7.75, standard deviation [SD] = 3.06; post-mean = 11.75, SD = 3.28, P = .02). A statistically significant difference in self-assessed confidence was also noted (scale of 1-5) (pre-mean = 3.08, SD = 0.65; post-mean = 4.5, SD = 0.68, P < .0001). All faculty implemented a QI project within the 6 months. Specifically, they were able to define local problem, write project charters, map existing processes, and identify areas for improvement. Five out of the 8 faculty were able to conduct small tests of change within the 6-month period. The remaining 3 participants had an opportunity to study the current process and plan to run tests of change after the 6 months. Conclusions: This 6-month course helped some participants run improvement cycles on their projects, while


12S


other participants had an opportunity to develop a QI proposal and evaluate current process for implementing improvement cycles. This course helped get buy-in from leadership to develop a leadership program in QI.

Abstract 14 Making the Connection Between Quality and Value: ACP’s New Approach to the PracticeBased Research Network Laura Lee Hall, Selam Wubu, Anne Marie Smith American College of Physicians, Washington, DC

Purpose: The American College of Physicians (ACP) is reconfiguring its practice-based research network (PBRN) so as to better meet the needs of clinical teams while measurably improving care processes and outcomes. The redesign of ACPNet as a quality improvement network incorporates evidence-based concepts of practice change.1,2 Key strategies include collaboration with state and regional health care system leaders, efficient linkage to professional and reimbursement opportunities and requirements (eg, self-evaluation of practice performance for maintenance of certification or MOC part IV; Physician Quality Reporting System), nurturance of peer-to-peer learning and collaboration, and a focus on patient engagement. Methodology: National advisory committees shaped the mission and approach for APCNet’s relaunch, along with the identification of performance measures and learning/quality improvement (QI) resources for the registry platform (MedConcert). Educational programs have launched in 4 states to date, with 4 more scheduled through February 2014. In addition, 2 virtual educational launch programs have been conducted. The live program agenda included discussions of strategies for better patient engagement, clinical updates in diabetes and adult immunizations, improving adherence, and care coordination. The evaluation includes a survey assessing attitudes, knowledge, and practice, developed by partners at the Center for Health Services and Outcomes Research at the Johns Hopkins University Bloomberg School of Public Health and NQF-endorsed performance measures. Monthly coaching calls are planned to assist participants. Results: More than 145 participants have registered for the initiative in the first 4 state and 2 virtual launch programs, ranging from residents to small practices to large health systems. An additional 200 participants are expected for upcoming state programs planned through February 2014. Most participants are physicians (80%) and utilize an electronic health record (74%). MOC part IV and peer-to-peer networking were identified as major

interests. Evaluations show strong support, with over 98% of participants recommending the program to their colleagues, 86% rating the program as “excellent” or “very good,” and 80% detailing practices changes they will seek to effect. Baseline performance measure data and survey results will be tallied in March 2014, to determine performance and other attributes of participating clinicians. Conclusions: ACP’s QI Network redesign has achieved enthusiastic early support from state chapters and health care providers. Programs that could upload performance measure data appear to facilitate physician participation. Gaps in understanding the QI process remain. However, making the connection with system data, incentive programs and professional requirements, and peers appear to nurture QI participation. References 1. Institute of Medicine. Best Care at Lower Cost: The Path to Continuously Learning Health Care in America. Washington, DC: National Academies Press; 2013. 2. PCORI. Dissemination and Implementation Roundtable; July 29, 2013; Meeting summary. http://www.pcori.org/ assets/2013/08/PCORI-Dissemination-ImplementationRoundtable-July-2013-Meeting-Summary-083013.pdf. Accessed November 8, 2013.

Abstract 15 Ovarian Cancer: Improving Diagnosis, Transforming Care Donald Monroe, Colleen Bimle, Karl Florence Vermillion, Inc, Austin, TX

Purpose: Ovarian cancer is the fifth leading cause of cancer death in women and the leading cause of death from gynecologic malignancy in the United States. Few cases of ovarian cancer are found at an early stage when treatment is more likely to be successful. Furthermore, approximately two thirds of women do not receive care per National Comprehensive Cancer Network (NCCN) guidelines. These factors define a significant opportunity to improve the quality of care for women with ovarian cancer. Methodology: Three differentiated research capabilities were brought together to create OVA1. First was the proteomics research capabilities of the Johns Hopkins Center for Biomarker Discovery and Translation. Second was the multivariate statistics and pattern recognition algorithm expertise of Johns Hopkins. Third was the clinical targeting of OVA1 to maximize patient benefit (sensitivity, negative predictive value [NPV], early stage


13S

The Proceedings of Medical Quality 2014 detection, and recognition of a broad range of ovarian malignancy subtypes) while mitigating risk (cancers missed and the possibility of unnecessary surgery). The final stages of design incorporated interactive dialog with the FDA. As a first of its kind diagnostic cleared by the FDA, OVA1 design targeted a strong risk–benefit proposition: (a) the test was only for triage to the gynecologic oncologist (GO) of patients already committed to surgery and (b) the test should not be used to overrule or counterselect patients deemed at high risk for malignancy, or under care of a GO. Two consecutive pivotal OVA1 trials were completed with leading academic gynecological oncologists as principal investigators: Dr Fred Ueland (University Kentucky) and Dr Robert Bristow (UC Irvine). Combined these studies represented over 1000 patients across 44 sites and identified >250 malignancies. Results were reported in 5 clinical research articles in preeminent peer-reviewed journals. Key Results: OVA1 sensitivity was 96% when combined with clinical impression (CI), reducing the number of cancers missed by CI from 25% to 4% across a broad range of ovarian malignancies. NPV ranged from 95% to 98%. OVA1 sensitivity was significantly greater than the benchmark methods: CI alone, CA125, or Dearkingmodified ACOG guidelines. Despite OVA1 test specificity ranging from 43 to 54%, referrals predicted by OVA1 were comparable to actual GO referral rate in the studies (56% vs 60%). OVA1 + CI detected 95% of all earlystage and 93% of stage I ovarian cancer, with significantly higher sensitivity than CA125, mod-ACOG, or CI alone. Conclusion: NCCN guidelines call for all ovarian cancers to be operated by board-certified GOs. Adherence to NCCN guidelines for ovarian cancer treatment correlates with 30% survival rate and may be a useful measure of quality cancer care. OVA1’s high sensitivity, NPV, and early-stage detection may offer an improvement in the standardized evaluation and GO referral of ovarian cancer patients. Future studies will examine how OVA1 use translates into actual clinical benefit and cost of care.

Abstract 17 Integration of Early Mobility in the ICU Setting Natasha Mehta, Karen Henz, Kevan Quinlan, Michelle Alexander, Rebekah Wang-Cheng Kettering Medical Center, Kettering, OH

Background: Hippocrates believed, “In every movement of the body, whenever one begins to endure pain, it will be relieved by rest.”1 Unfortunately, bed rest for various diseases does not improve outcomes.2 In the

first week of intensive care unit (ICU) stay, skeletal muscle wasting occurred most rapidly in critical illness survivors because of a net catabolic state that resulted despite adequate enteral nutrition and protein synthesis rates.3 Although a study of 155 ICUs4 found that 66% of 200 730 adults were discharged home, other literature demonstrates that over half of discharged ICU patients are unable to return to premorbid activity levels.5 Since cognitive impairment and functional disability can persist for at least 5 years, early mobilization of ICU patients is imperative.6,7 Methods: ICU nurses, rehab, respiratory, and physicians participated in a 5-series webinar8 featuring EMOB program development. Efforts addressed staff fear of mobilizing ICU patients via discussion and educational videos. Staff enthusiasm spread with the first successfully mobilized patients. At admission, a nurse screens in EPIC for contraindications, hemodynamic status, and wakefulness directly in EPIC. If patient is eligible, the physician orders therapy. A nurse screens every shift or if patient condition changes. Results: ICU data were collected pre/post EMOB for the following measures: ICU and total length of stay, discharge disposition, falls, restraints, decubitus ulcer, hospital-acquired infections, and unexpected death. Patients mobilized were also compared to those not. Conclusion: EMOB implementation in the ICUs improved outcomes. We attribute success to a committed interdisciplinary team, educational emphasis, and administrative support. Distinguishing mobilization from ambulation was an unexpected but crucial learning that led to a major cultural shift and expansion of EMOB outside the ICUs. References 1. Hippocrates. The Genuine Works of Hippocrates. London, England: Sydenham Society; 1849. 2. Allen C, Glasziou P, Del Mar C. Bed rest: A potentially harmful treatment needing more careful evaluation. Lancet. 1999;354:1229-1233. 3. Puthucheary ZA, Rawal J, McPhail M, et al. Acute skeletal muscle wasting in critical illness. JAMA. 2013;310: 1591-1600. 4. Wagner J, Gabler NB, Ratcliffe SJ, Brown SE, Strom BL, Halpern SD. Outcomes among patients discharged from busy intensive care units. Ann Intern Med. 2013;159: 447-455. 5. Engel HJ, Needham DM, Morris PE, Gropper MA. ICU early mobilization: from recommendation to implementation at three medical centers. Crit Care Med. 2013;41: S69-S80. 6. Herridge MS, Tansey CM, Matté A, et al. Canadian Critical Care Trials Group. Functional disability 5 years after acute respiratory distress syndrome. N Engl J Med. 2011;364:1293-1304.


14S


7. Iwashyna TJ, Ely EW, Smith DM, Langa KM. Long-term cognitive impairment and functional disability among survivors of severe sepsis. JAMA. 2010;304:1787-1794. 8. Institute for Healthcare Improvement Expedition. Mobility in the intensive care unit. November 2012-January 2013. http://www.ihi.org/offerings/virtualprograms/expeditions/ MobilityintheICU. Accessed May 31, 2014

also a significant association between comorbidities and the presence of AEs: AEs were reported in 39.4% of subjects with at least one comorbidity but only in 7.41% of subjects with no comorbidities (P < .0044). Conclusions: AEs in hypertensive long-term care residents appear to be associated with the use of multiple antihypertensive medications and the presence of comorbidities rather than with a “tight” control of their blood pressure.

Abstract 18 Is Tight Blood Pressure Control in Hypertensive Long-Term Care Patients Associated With Adverse Events (AEs)? Ahlam Alsomali1, Judith Beizer2, Lisa Rosen3, Christian Nouryan1, Gisele WolfKlein1

1 North Shore-LIJ Health System, Manhasset, NY; 2St John’s University, Queens, NY; 3Feinstein Institute for Medical Research, Manhasset, NY

Purpose: The association between blood pressure (BP), cardiovascular and cerebrovascular morbidity and mortality had been the focus of multiple studies in the older population, supporting the evidence for diagnosing and treating hypertension (HTN), regardless of age. The aim of this study is to explore potential adverse events (AEs) of “tight” BP control in residents of long-term care facilities. Methodology: The study utilized a 3-month retrospective electronic chart review of hypertensive subjects, 65 years and older, residing in a long-term care facility. The primary outcome variables were adverse events (AEs), namely, falls, syncopes, hypotensive episodes (systolic BP 25), noninstitutionalized older adults (65+) attending a geriatric faculty practice from September to November 2013. Acute “sick calls” were excluded. Descriptive statistics and associations between mean happiness score and potential factors were assessed. Results: Of the 88 subjects included, 35% were male, mean age was 83.4 years, 44% were married, 86% had family living within 25 miles, 51% were college graduates, 33% were members of religious groups, and 40% reported strong or very strong religious beliefs. Average happiness score was 4.2 out of 6 (range = 2.4-5.8). Factors reported as top “contributors to happiness” were health (71%), social relationships (69%), financial circumstances (41%), independence (36%), and psychological well-being (34%). Using Mann–Whitney U-tests and Kruskal–Wallis tests, results showed that education level, transportation method, strength of religious beliefs, and faith were significantly associated with mean happiness score (P = .03, P = .03, P = .004, and P = .02, respectively). Using Spearman correlation, the strongest correlate with happiness was religious belief (r = .43, P < .001). Furthermore, subjects who were active in faith-based groups had higher


15S

The Proceedings of Medical Quality 2014 happiness scores compared with those who were not (P = .007). Compared with subjects who drive, subjects who depend on others for transportation had lower happiness scores (P = .005). High school graduates had lower happiness scores compared with college graduates (P = .008), and subjects aged 85 to 89 had lower happiness scores compared with those aged 65 to 69 (P = .04). Conclusions: Increased awareness of factors that affect happiness and life satisfaction in geriatric patients may help health care practitioners to develop more relevant and realistic therapeutic plans of care for the overall functional, physical, and emotional health of older adults.

motivating since they have to attend orientation again if they cancel or no show. Staff see their work as more rewarding and less spinning their wheels on client intakes that do not materialize. Overall, the new client orientation process improves the quality of PHP services we provide. It has been so successful that we are now using it in our other adult PHPs.

Abstract 20

Jasdeep Sidhu1, Joanna Stein2, Liron Sinvani1, Suresh Basnet1, Nick Fitterman1, Corey Karlin-Zysman1, Daniel Bashari1, Roman Bernstein3, Mansi Mehta1, Christian Nouryan1, Martin Lesser2, Gisele Wolf-Klein1

Project Welcome Leah Russack-Baker, Mary Kinsella-Shaw, Deborah Weidner

Abstract 21 Embarking on a Hospitalist Career: Residents’ Views and Perspectives

Natchaug Hospital, Hartford HealthCare, Mansfield Center, CT

1

North Shore-LIJ Health System, Manhasset, NY; 2Feinstein Institute for Medical Research, Manhasset, NY; 3Lutheran Medical Center, Brooklyn, NY

Increasing the comfort level of newly referred clients is an important part of helping them become engaged and invested in their treatment in partial hospital programming (PHP). At the same time, we strive to help our staff make the most efficient use of their time. The Quinebaug Day Treatment Program (a Natchaug Hospital Adult PHP in northeastern Connecticut) has brought these 2 facets of treatment planning together in our Project Welcome orientation. Prior to implementation, potential clients were given an intake appointment with a clinician and a psychiatrist. The no show rate was approximately 60%, leaving the clinician and psychiatrist with about 90 and 60 minutes of unexpected time, respectively. Also, over one third of clients dropped out after completing the intake process before starting PHP. Reasons for high no show rates in both steps of the admissions process were often because of lack of understanding of PHP. For example, not realizing the program required daily attendance, included copays, is group therapy based, to name a few. With the Project Welcome orientation process, the client is able to learn about the program, meet other new clients, meet staff, practice being in a group setting, and complete much of the paperwork that was previously done when they arrived for their intake appointment. There are also 3 opportunities to reach out and help the client know we are invested in their success before they are admitted: calling the evening before orientation, the evening before intake, and the evening before their admission. After implementation of Project Welcome, our attendance rate for intakes increased by 12% to approximately 90%. Census has remained high (at or exceeding budget). Clients report they value the requirement of attending the orientation prior to scheduling an intake. Clients describe it as low-key and informative. They also say it is

Purpose: The number of hospitalists in the United States grew from 1000 in 1996 to 30 000 in 2010. In 2013, 70% of hospitals with over 200 beds are served by hospitalists. Driving forces behind hospital medicine’s increasing popularity remain uncertain, opening unique prospects for collaborative research at the intersection of geriatrics and hospital medicine. This study examines decision pathways of residents leading to hospitalist careers in our aging society. Methodology: An anonymous survey of Internal Medicine Residents was distributed at Grand Rounds in 5 New York academic hospitals from September to November 2013. Descriptive statistics were calculated and associations between potential factors and hospitalist career choices were assessed using χ2 or Fisher’s exact test. Results: There were 149 respondents: 49% male, 71% US citizens, 56% US trained. Half (57%) were less than 30 years old. Residency level was equally divided between PGY-1 (31%), PGY-2 (36%), and PGY-3 (32%). Almost all (92%) reported working over 50 hours/week. Half (48%) would consider a hospitalist career, while 8% would pursue a geriatric hospitalist future. When asked about most influential factors for career choice, 29% selected workload schedule, 16% professional environment, 16% reward incentives, and 15% personal autonomy. Significant factors associated with decision to pursue a hospitalist career included PGY level (P = .008), age (P = .02), enjoyment of interaction with older patients’ family members (P = .03), and timing of career decision during either medical school or residency (P < .001). Older respondents, those at higher residency training level, those who enjoy interacting with family members, and those who finalized career decisions during residency were more likely to embark on a hospitalist career.


16S


Conclusions: Although a hospitalist career has become a popular choice among internal medicine residents, few are open to becoming geriatric hospitalists. In view of aging demographics challenges in the United States, this study underlines pressing needs for educational initiatives, particularly during early residency, to better prepare young physicians to the realities of caring for the older population in hospital settings and to explore opportunities for geriatric innovations within hospital medicine.

Abstract 22 Use of a Personalized Medicine, Gene Expression Score Influenced Cardiology Referrals Among Patients Presenting With Symptoms Suggestive of Obstructive Coronary Artery Disease: Interim Results from the PRESET (A Registry to Evaluate Patterns of Care Associated With the Use of Corus® CAD in Real-World Clinical Care Settings) Registry Joseph Ladapo1, David Sharp2, Bruce Maniet3, Linda Ross4, John Blanchard4, Lon Castle4, Mark Monane4

1 NYU Langone Medical Center, New York, NY; 2Doctors for Health, Omaha, NE; 3Bells Medical Clinic, Bells, TX; 4CardioDx, Inc, Palo Alto, CA

Purpose: Approximately 3 million patients per year present to primary care clinicians with symptoms suggestive of obstructive coronary artery disease (CAD). Often, even after obtaining a detailed history and performing a physical examination and resting electrocardiogram, physicians are unable to confidently determine the primary etiology of these symptoms during an initial visit. This scenario leads to an overall $6.7 billion/year spend on the noninvasive and invasive cardiac testing in the United States for nondiabetic patients with no prior revascularization or myocardial infarction. The process may also expose patients to appreciable risk of radiation and contrast-dye-related side effects. Better methods are needed for the primary care provider to determine which patients can be safely managed in the primary care setting and which patients should be referred to the cardiologist. Methods: The gene expression score (GES) is a previously validated quantitative diagnostic test for nondiabetic patients, measuring expression levels of 23 genes from peripheral blood to determine the likelihood of a patient having >50% coronary artery stenosis. The GES has a 96% negative predictive value and can identify symptomatic patients at low risk for underlying obstructive CAD. We hypothesized that use of the GES would improve quality of care by safely excluding low-risk patients from unwarranted cardiac testing. We are currently enrolling into the prospective PRESET Registry

(NCT01677156), which will include 1000 nonacute, nondiabetic adult patients with no history of CAD from 21 US primary care practices. Clinicians provide the preand post-GES diagnosis and evaluation plan for each patient. Demographics, clinical factors, and GES results (predefined as low [GES ≤ 15] or elevated [GES > 15]) are collected, as well as diagnostic tests performed with results, and referrals to cardiology. Additional clinician and patient quality of care measures, such as satisfaction with care, are being assessed. Results: In a preliminary cohort of 393 patients, 199 (50.6%) are women, the median age was 55 years with 116 (29.5%) age ≥65, and the median body mass index was 29.8. The median GES was 17 (range = 1-40), and 177 patients (45.0%) had low scores. Referral rates to cardiology were only 10.7% (19/177) in the low GES group. At the 30-day follow-up post-GES, the major adverse cardiac event rate was 0% in the low GES group. Registry enrollment is ongoing, with final analysis planned on enrollment completion. Conclusions: In this community-based primary care registry, initial results show that a personalized medicine, gene-expression-based test could appropriately stratify patients presenting with symptoms suggestive of obstructive CAD. By removing these low-risk patients from further cardiac testing, unnecessary workups can be avoided. These results demonstrate that the GES could improve the quality of care without affecting patient safety.

Abstract 23 Designing and Evaluating a Resident Education Program on Patient-Centered Communication (PCC) Hanan Aboumatar1, Bickey Chang1, Adrian Alday1, Gail Berkenblit2, Sanjay Desai2, Carol Fleishman3, R. Samuel Mayer4, Patricia Thomas2, Zackary Berger2, Mark Hughes2, Paula Phurrough5, Charlene Rothkopf5, Peter Pronovost1

1 Johns Hopkins Medicine Armstrong Institute for Patient Safety and Quality, Baltimore, MD; 2Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD; 3Johns Hopkins Medicine Simulation Center, Baltimore, MD; 4Department of Physical Medicine and Rehabilitation, Johns Hopkins University School of Medicine, Baltimore, MD; 5Johns Hopkins Hospital Patient and Family Advisory Council, Baltimore, MD

Purpose: Communication is regularly cited by patients/ families as a key area for health care improvement. Health care professionals’ communication skills can be improved, and such improvement is associated with improved outcomes. We aimed to partner with patients/families to create and evaluate an education program on patient-centered communication (PCC).


17S

The Proceedings of Medical Quality 2014 Methodology: We assembled a curriculum development team of medical educators and patient representatives, reviewed PCC literature, developed the curriculum, and conducted focus groups with additional patients/family members for feedback and refinement. The final program included simulation-based training, followed by feedback from preceptors and patients in clinical settings. The simulation-based training was implemented in 2 residency programs. Half of the residents within each program were randomized to participate in the training (n = 24) with the other half serving as the control group. To evaluate the program, we (a) assessed participants’ satisfaction and confidence in handling select patient scenarios pre and post simulation and (b) compared participants’ self-reported use of 10 specific PCC skills in clinical practice to that within the control group. Results: Among participants, 83% reported the simulation was a positive learning experience, and 96% found the Standardized Patient feedback useful. Participants reported a significant increase in their confidence in conducting a patient-centered encounter. At 6-month followup, the total count of PCC skills reported by residents as “never”/“seldom” used averaged 0.57 in the control group versus 0.05 in the intervention group (P = .01). The total count of skills reported as “most/all of the time” used averaged 5.0 in the control group versus 6.3 in the intervention group (P = .16). Use of different PCC skills varied, with “listening to patient concerns” and “expressing empathy” most consistently used, and “verifying patient understanding” and “brainstorming and problem solving with patient” least consistently used. Conclusions: Patients/families could play an important role in developing programs to advance PCC. Simulation training in this area is effective and well received. Consideration should be given to advancing PCC skills that are less consistently used. Evaluation of the impact of similar training programs on patient experience of care and outcomes is needed. Grant Support: The Picker Institute Inc. and The Arnold P. Gold Foundation.

Abstract 24 The Relationship of Hospital Charges and Volume to Surgical Site Infection after Total Hip Replacement (THR) Rebecca Boas, Kelsey Ensor, Edward Qian, Lorraine Hutzler, James Slover, Joseph A. Bosco NYU Hospital for Joint Diseases, New York, NY

Introduction: The purpose of this study was to analyze the effect of hospital volume and charges on the rate of

surgical site infections (SSIs) for total hip replacements (THR) in New York State (NYS). Despite increasing costs for hip replacements, there remains little evidence that hospitals that charge more provide higher quality care. We examined if the hospital charges of high-volume hospitals were justified by lower SSI rates. Methods: The Statewide Planning and Research Cooperative System (SPARCS) database was used to identify individuals who underwent THR from 2008 to 2011. Hospitals were divided into volume groups based on annual procedural volume. Risk-adjusted SSI rates were obtained from the NYS Department of Health. Results: We analyzed 93 620 hip replacements performed. Hospital charges increased from $43 713 in 2008 to $50 652 in 2011 (P < .01), 5.43% higher than anticipated based on Medical Cost Inflation Rates. Compared with lower volume hospitals, patients who underwent THR at the highest volume hospitals (>600 procedures/ year) had significantly lower SSI rates (P = .003) and higher hospital charges (P < .0001). Using a hypothetical cohort of 1000 patients, we found that in the highest volume hospitals, preventing 1 SSI was associated with $1.6 million dollars in increased charges. Conclusion: Volume and charges for THR have increased dramatically from 2008 to 2011 in NYS. Hospitals with higher annual caseloads charged more for THR and had lower rates of SSIs. This examination suggests that higher charges correlate with a decreased risk of SSI at hospitals with the largest volumes. Further study is needed to examine elements of care that lead to increased charges and their impact on SSI rates, readmission, and revisions.

Abstract 25 Geographic Differences in Hospital Charges and Surgical Site Infections for Total Hip Replacement in New York State Rebecca Boas, Kelsey Ensor, Edward Qian, Lorraine Hutzler, James Slover, Joseph A. Bosco NYU Hospital for Joint Diseases, New York, NY

Introduction: Geographical variation in health care utilization and costs has been well documented. The purpose of this study is to explore regional effects on the quality and value of health care delivered. To do this, we analyzed overall hospital charges and surgical site infection (SSI) rates for total hip replacements (THR), in the different geographic regions of New York State (NYS). Methods: The Statewide Planning and Research Cooperative System (SPARCS) database was used to identify individuals who underwent hip replacement


18S


from 2008 to 2011. Hospitals were divided into regional groups defined by the NYS Department of Health. Riskadjusted SSI rates were obtained from the Department of Health. Results: There were 93 620 THR performed in NYS from 2008 to 2011: 9.6% in the Capital District, 11.3% in Central New York, 13.6% in Long Island, 8.5% in New Rochelle, 37.0% in New York City, 10.0% in Buffalo, and 9.6% in Rochester. Hospital charges had substantial variation across all regions. Charges were higher in New York City ($58 294), New Rochelle ($58 957), and Long Island ($59 126) when compared to the Capital District ($34 569), Central New York ($29 732), Buffalo ($29 643), and Rochester ($27 086) (P < .0001). Discussion: Regional analysis of NYS hospitals shows little variation in infection rates and more significant difference in charges. These significant regional differences in charges may be in part due to the presence of urban hospitals and academic medical centers. When analyzed by region these regional differences in charges do not correlate with any differences in quality as measured by SSI rates.

Abstract 26 Improving Patient Safety by Reducing Complications Related to Warfarin Monitoring and Highly Potentiating Antimicrobial Prescribing Processes Lisa Daniels, Beilei Zhang, John Kuth, Jason Barreto, Jeremy Anderson, Tosh Pritish Mayo Clinic, Rochester, MN

Warfarin accounts for 30 million prescriptions and over 40 000 adverse events annually. Warfarin has significant interaction with highly potentiating antimicrobials (eg, trimethoprim/sulfamethoxazole [TMP/SMX], fluconazole, metronidazole) and is therefore an important target for intervention. Retrospective cohort analysis evaluating international normalized ratio (INR) and adverse events among outpatients on chronic warfarin therapy who received a course of TMP/SMX between September 1, 2011, and December 31, 2011, compared to control patients on chronic warfarin for anticoagulation during the same time period who did not receive a course of antimicrobials. TMP/SMX patients had a significantly increased risk of INRs >4.5 with an absolute risk of 10%, relative risk of 4.9 (P = .0002) compared to controls, and number needed to harm (NNH) of 13. One associated adverse event of hematuria due to INR elevated to 6.6.

A process map for warfarin management and antimicrobial prescription (using TMP-SMX as a model) was conducted. The result from failure modes and effects analysis (FMEA) suggests that the current processes contains at least 30 failure modes, which cannot be detected automatically, could happen at least once a month, and may lead to patient death. FMEA identified the following interventions: (a) correct identification of patients who are on warfarin, (b) a message sent to the anticoagulation clinic managing a patient’s warfarin whenever a highly potentiating antimicrobial is prescribed elsewhere to facilitate close INR monitoring and adjustment of warfarin dose by the anticoagulation clinic, (c) a pop-up message at the point of prescription of a highly potentiating medication to patients on warfarin for chronic anticoagulation informing the provider of the risk and linking to the care process model to closely monitor INR and reduce warfarin dose.

Abstract 27 How Frequently Can Recommendations for Additional Imaging Be Avoided by Thoroughly Evaluating All Comparison Studies? A Retrospective Assessment in Abdominal and Pelvic Radiology Examinations Ankur Doshi, Andrew Rosenkrantz NYU Langone Medical Center, New York, NY

Purpose: Abdominal computed tomography (CT) and magnetic resonance imaging (MRI) reports frequently contain a recommendation for additional imaging (RAI) to evaluate an indeterminate finding, generating increased utilization and cost. Our aim was to determine the fraction of RAI that may be avoided via a thorough comparison with all available prior imaging and to identify the characteristics of such instances of “avoidable” RAI. Methods: A total of 305 abdominopelvic CT and MRI reports that contained an RAI for an indeterminate finding were included. The reports and original images from the prior imaging examinations were comprehensively reviewed. The RAI was considered avoidable if the prior imaging demonstrated either long-term stability or definitive proof of benignity of the finding. The fraction of RAI that was avoidable, as well as characteristics of identified avoidable RAI, were computed. Results: In 305 patients with RAI, there were 13 avoidable RAI in 11 patients (frequency of avoidable RAI of 4%). The 13 avoidable RAI were distributed as follows: 31% in the lung bases (stable nodules and interstitial lung disease), 23% in the kidneys (previously proven cysts), 23% in the adrenal glands (2 proven adenomas and 1


19S

The Proceedings of Medical Quality 2014 stable lesion), 8% in the pancreas (abnormal contour proven to be normal variant), 8% in the adnexa (stable ovarian cyst), and 8% in the lymph nodes (stable lymphadenopathy). The key prior was a different modality from the study with the RAI in 39% and was not mentioned as a comparison study in 92%. The key prior study included MRI (62%), CT (23%), ultrasound (8%), and PET-CT (8%), as well as studies of the abdomen (62%), spine (15%), chest (15%), and lower extremity (8%). The key finding was noted in the impression of the prior study in 31%, noted in only the report body in 15%, present on the prior images but not described in the report in 46%, and present in an outside study uploaded to our system in 8%. Conclusion: A small fraction of RAI (4%) can be avoided by a thorough evaluation of all prior imaging studies, including studies of other body parts and modalities. Prior images should be directly reviewed, as the key finding may not be mentioned in the prior reports, thus contributing to additional RAI.

Abstract 29 Case Report of a Quality Improvement Organization’s Statewide Activities to Reduce Preventable Hospitalizations: Results and Lessons Learned 1

2

Thomas Meehan Jr , Thomas Van Hoof , Thomas Meehan Sr1 1

Qualidigm, Wethersfield, CT; 2University of Connecticut School of Nursing, Storrs, CT

About 20% of Medicare beneficiaries are readmitted to a hospital within 30 days of their initial discharge. The Centers for Medicare & Medicaid Services has implemented a variety of strategies and programs to decrease preventable hospitalizations of Medicare patients, including training and technical assistance provided through quality improvement organizations (QIOs). Beginning in August 2011, Qualidigm, the Connecticut QIO, set out to achieve this goal through a variety of activities, including the following: (a) convening a statewide advisory panel of health care stakeholders to assist and coordinate activities; (b) promoting implementation of evidence-based quality improvement interventions in statewide and local venues; (c) providing assistance to geographically defined communities with formation, needs assessments, data collection and analysis, and intervention tools; (d) holding a series of 10 educational conferences and 6 webinars open to all; (e) hosting meetings for leaders of each community to provide feedback, share experiences, and learn from one another; and (f) posting a variety of links, intervention tools, and outcomes data reports on its Web site. Participation in the QIO’s statewide activities

was consistently high (mean conference attendance = 244, range = 90-456) and participants were positive in their evaluations. Ninety-four percent of satisfaction survey respondents indicated that they agreed or strongly agreed that the conferences were useful to their work. Web site traffic documented high utilization of patient education videos and outcomes reports. Local communities reported implementing multiple interventions to decrease preventable hospitalizations (October 2013: mean = 9.7, range = 5-20). The statewide hospital admission rate among fee for service (FFS) Medicare beneficiaries dropped from 157.4/1000 at baseline (10/2010 to 3/2011) to 146.4/1000 at follow-up (10/1/2012 to 3/31/2013), a decline of 4487 hospitalizations (7.0% relative improvement). Similarly, the statewide 30-day readmission rate among FFS Medicare beneficiaries dropped from 29.4/1000 at baseline to 26.0/1000 at follow-up, a decline of 1512 readmissions (relative improvement of 11.6%). Although statewide rates of hospital admission and 30-day readmission declined in CT concurrent with the QIO’s activities, it is impossible to determine the impact of these actions due to confounding from other national and local activities.

Abstract 32 Use of American College of Surgeons Trauma Quality Improvement Program (ACS-TQIP) to Analyze the Cause of Unplanned Extubation in Trauma Patients Alph Emmanuel1, Colleen Desai2, Ellen Boucher1, Sandy Gifford1, William Marshall1, David Shapiro1, Scott Ellner1

1 Saint Francis Hospital and Medical Center, Hartford, CT; 2Connecticut Children’s Medical Center, Hartford, CT

Introduction: The unplanned removal of an endotracheal tube is a life-threatening incident. A population of critically ill trauma patients at an ACS-designated urban level II trauma center was evaluated for variables leading to unplanned extubation (UE). Data were provided by the American College of Surgeons Trauma Quality Improvement Program (TQIP). Methods: A retrospective evaluation of trauma patients in a surgical critical care setting during a 12-month period was conducted. UE was identified as an unintended, premature removal of an endotracheal tube secondary to dislodgement due to activity, provision of patient care, or self-extubation by the patient. Results: A total of 20 cases of UE occurred among 17 patients, with 2 patients having more than one UE, among 144 ventilator days. This represents an incidence of 4.3 UEs per 100 ventilator days. The analysis


20S


of common factors for the UE population identified the following: 19 (95%) had no change to the ventilator settings within the 4 hours preceding UE; 15 (75%) were restrained at the time of extubation; 13 (65%) were perceived to have adequate analgesia; 12 (60%) had no sedation administered within 4 hours of UE; 11 (55%) had no analgesia administered within 4 hours of UE; time of day and ventilator settings did not appear to have any significant effect. Interestingly, 12 patients (60%) did not require reintubation. Conclusions: Though most patients were perceived to have adequate analgesia by providers, a majority had not had an analgesic or sedative medication administered within the 4 hours preceding the UE. We concluded that the currently utilized pain assessment tool may be less accurate in the intubated patient, with providers failing to identify behavioral indicators of anxiety and pain. Delirium and undertreatment of pain may have led to inadvertent self-extubation. The increased number of UE in the trauma population discovered by this study prompted a larger prospective study analyzing the factors leading to UE in the entire intensive care unit population. A new pain assessment scale was utilized based on the findings of this study which is currently in use, and to be reported. Also, most patients had no ventilator changes in the 4 hours prior to UE, which may indicate the need for a more aggressive initiative to assess readiness for extubation.

Abstract 33 Causality in Unplanned Extubations in a Medical-Surgical Critical Care Setting Alph Emmanuel1, Colleen Desai2, Laura Sanzari1, William Marshall1, Scott Ellner1, David Shapiro1 1

Saint Francis Hospital and Medical Center, Hartford, CT; 2Connecticut Children’s Medical Center, Hartford, CT

Introduction: Delirium, pain, anxiety, patient care activities, and other factors may contribute to the unplanned removal of an endotracheal tube. A medical-surgical critical care population at an ACS-designated urban level II trauma center was evaluated for variables leading to unplanned extubation (UE). Data provided by the American College of Surgeons Trauma Quality Improvement Program demonstrated an increase in UE in the trauma population, prompting this assessment among the entire ICU. Methods: A prospective evaluation of all patients who experienced UE in a mixed critical care setting during a 10-month period was conducted. UE was identified as an unintended, premature removal of an endotracheal tube

secondary to dislodgement due to activity, provision of patient care, or self-extubation by the patient. Results: A total of 58 cases of UE occurred among 3890 ventilator days, representing an incidence of 14.9 UEs per 1000 ventilator days. The analysis of common factors for the UE population identified the following: 53 (91%) were perceived to have adequate analgesia, 48 (83%) were restrained at the time of extubation, 45 (78%) occurred when no personnel were present in patient room, 36 (62%) had no analgesia administered within 4 hours of UE, 36 (62%) were perceived to have adequate sedation, 34 (59%) had no change to the ventilator settings within the 4 hours preceding UE, 34 cases (59%) had sedatives administered within 4 hours of UE, and 33 (57%) were intubated for more than 24 hours. Medical service patients represented 38 (66%) of the patients with UE, while the trauma population represented 8 (14%). Interestingly, 41 patients (71%) did not require reintubation. Conclusion: Though most patients were perceived to have adequate analgesia by providers, most had not had an analgesic or sedative administered within the 4 hours preceding the UE. It was the conclusion of the investigators that our currently utilized pain assessment tool may be less accurate in the intubated patient, with providers identifying behavioral indicators as anxiety, rather than pain, prompting nonanalgesic pharmacotherapy. Delirium and undertreatment of pain may have led to inadvertent self-extubation. A new pain assessment scale is currently in use, and it is to be reported.

Abstract 34 Using the American College of Surgeons Trauma Quality Improvement Program (ACSTQIP) to Determine 30-Day Complication and Hospital Readmission Rate Alph Emmanuel1, Laura Sanzari1, Colleen Desai2, Sandy Gifford1, William Marshall1, David Shapiro1, Scott Ellner1 1

Saint Francis Hospital and Medical Center, Hartford, CT; 2Connecticut Children’s Medical Center, Hartford, CT

Introduction: The incidence of complications and readmissions among trauma patients treated at Saint Francis Hospital and Medical Center (SFHMC) is lower than the national benchmark. Trauma patients who receive care at SFHMC may not return for follow-up care for a complication they may have suffered related to the trauma. They may go to a hospital that is more convenient. The 30-day complication and readmission incidence may be falsely low if we examine the rate of complications and readmissions in this institution alone because of the large number


21S

The Proceedings of Medical Quality 2014 of patients that are lost to follow-up. The true incidence can be obtained by sampling the entire trauma population for complications they may have addressed at another institution. This study applies the National Surgical Quality Improvement Program (NSQIP) 30-day followup standard to the Trauma Quality Improvement Program (TQIP) database to determine the real complication and readmission rate. Methods: Patients captured in the TQIP registry through admission to SFHMC for trauma care will have 30-day follow-up for 6 months. Through the outpatient and inpatient electronic medical records, patients will be identified for 30-day follow-up and associated complications and readmissions. Patients will be administered a predetermined questionnaire by phone interview inquiring about any complications or readmissions within 30 days of discharge from SFHMC. If we are unable to reach them by phone, a letter with the questionnaire will be mailed to them. Results: The study is ongoing. Conclusion: This article seeks to utilize the established Quality Assessment and Improvement tool utilized by the NSQIP database and apply it to the new and evolving TQIP database to assess the quality of care provided and identify areas of improvement. This is the first study of its kind using the TQIP database. If the study shows a significant disparity between the observed and expected rates of complications, it will lend validity to the incorporation of a quality assessment tool into the TQIP database. Studies have shown significant benefit to utilization of the NSQIP postdischarge feedback questionnaire. This could lead to significant improvement in the quality of care provided to trauma patients.

Abstract 35 Is Weight Gain in Long-Term Care Residents Associated With Lower Mortality Outcome? 1

1

Wei Zhou , Andrzej Kozikowski , James Lolis1, Barbara Tommasulo1, Yue Li1, Joanna Stein2, Renee Pekmezaris1, Martin Lesser2, Gisele Wolf-Klein1 1

2

North Shore-LIJ Health System, Manhasset, NY; Feinstein Institute for Medical Research, Manhasset, NY

Purpose: Despite the numerous health risks associated with being overweight, the effect of weight loss on health and longevity remains controversial, particularly in older adults. We explored the association between weight fluctuations in older residents in long-term care facilities and health outcomes. Methodology: A 6-year retrospective chart review of long-term care residents was conducted in one skilled

nursing facility, collecting monthly weights in addition to demographic data of all residents institutionalized for at least 1 year. Weight fluctuations were classified in increments of 5% from baseline month 1 through 12 and classified as stable, loss, or gain. Demographics, weight changes, body mass index (BMI), comorbidities, number of hospitalizations, and mortality were analyzed. Association between weight change (and the other demographic and clinical variables) and mortality outcome, as well as number of hospitalizations were assessed, respectively, using χ2 or Fisher’s exact test, as appropriate, and Poisson regression. Results: A total of 116 patients fit inclusion criteria; average age was 83 years (range = 66-98), with 71.6% female and 88.7% Caucasian. Median length of stay was 877.5 days (range = 383-2173). Average body weight at baseline was 141.6 lbs (range = 83-315 lbs) with BMI of 24.7 (range = 15.2-50.8). A third (36%) of residents had stable weight, 38% gained weight, and 26% lost weight. Neither weight change category, nor baseline BMI, were significantly associated with mortality (P = .06, P = .45, respectively). Furthermore, the association between weight change category and the number of intercurrent hospitalizations was also nonsignificant. However, significant associations with mortality were found among subjects with tube-feeding (P = .01), those taking a nutritional supplement (P = .03), or had a history of congestive heart failure (P = .046), depression (P = .02), or hypertension (P = .008). Finally, residents who lost greater than 5% of their body weight were more likely to die in the nursing home (P = .03). Conclusions: Whereas weight fluctuation in longterm care residents does not appear to be associated with health outcomes in general, weight loss is associated with increased probability of death in long-term care setting.

Abstract 36 Association Between Vitamin D Levels and Type 2 Diabetes in Elderly Patients in LongTerm Care Institutions Mansi Mehta1, Barbara Tommasulo1, Renee Pekmezaris1, Andrzej Kozikowski1, Meredith Akerman2, Nooshi Karim1, Judith Beizer1, Stuart Weinerman1, Gisele Wolf-Klein1

1 North Shore-LIJ Health System, Manhasset, NY; 2The Feinstein Institute for Medical Research, Manhasset, NY

Purpose: Older adults and particularly nursing home residents are known to be at risk for vitamin D deficiency. Vitamin D deficiency has been implicated as a contributing factor in a multiplicity of diseases including type 2


22S


diabetes mellitus (DM). We proposed to assess the prevalence of vitamin D deficiency in elderly patients with and without type 2 DM. Methodology: A retrospective chart review study was conducted in a nursing home, using electronic medical records of all patients over the age of 55 from January 1, 2007, to December 31, 2012. All diabetic residents met American Diabetes Association criteria for diagnosis of type 2 DM. Vitamin D levels, alkaline phosphatase levels, phosphorus, parathyroid hormone levels, and calcium and albumin levels were recorded as well as demographics, comorbidities, mobility, medications, and supplements. Vitamin D deficiency was defined as 80% and included “highly suspicious,” “consistent,” “likely,” “most likely,” “most consistent,” “representative of,” and “characteristic of.” These terms were associated with CV ranging from 4.6% (“highly suspicious”) to 18.7% (“characteristic of”) and IDR ranging from 0% (“highly suspicious”) to 27.5% (characteristic of”). Expressions of intermediate diagnostic confidence were defined as terms having a mean confidence between 40% and 80% and included “compatible with,” “suggestive of,” “probably,” “concerning for,” “worrisome for,” “suspicious for,” “presumably,” and “possibly.” These


24S


terms were associated with CV ranging from 19.2% (“compatible with”) to 37.8% (“possibly”) and IDR ranging from 25% (“suggestive of”) to 52.5% (“presumably”). Expressions of low diagnostic confidence were defined as terms having a mean confidence of

2014 SICPRE Proceedings: Meeting Proceedings.

Proceedings of the Stillbirth Summit 2014.

Introduction for the proceedings of Origins 2014.

Calcutta Medical Society, Proceedings of the.

Proceedings of the Gwalior District Medical Society.

Proceedings: On the medical use of isoniazid.

Proceedings of the Thirteenth Annual UT- KBRIN Bioinformatics Summit 2014.

A Multistakeholder Approach to Improving Quality in Rheumatoid Arthritis: Proceedings from the Rheumatoid Arthritis Quality Expert Medical Panel.

Proceedings of the 2014 National Toxicology Program Satellite Symposium.

Editorial comment: 2014 International Hip Society proceedings.

Editorial comment: 2014 Hip Society proceedings.

Proceedings of the 2014 annual meeting, April 30-May 4, 2014, Las Vegas, NV.

Proceedings of the Bengal Branch of the British Medical Association.

Proceedings of the Bengal Branch of the British Medical Association.

The quality of medical care.

Continuing medical education: september 2014.

Continuing medical education: June 2014.

Proceedings: General aspects of quality assurance.

Graduate medical education, 2013-2014.

Proceedings: A medical officer's view of the future.

[Medical University of Vienna researcher of the month, May 2014].

2014 Annual Report of the Canadian Veterinary Medical Association.

Preface. Biometals 2014--Proceedings of the 9th International Symposium Biometals 2014 at Duke University, Durham, NC, USA.

The equity of China's emergency medical services from 2010-2014.