Research Data Management: 2019 Research Data Management Symposium

DNP Quality Improvement Toolkit for Best Practice Use of Active Secondary Data

Courtney Omary, MSN, RN-BC                                                                                                                         

DNP, Emory University, Nell Hodgson Woodruff School of Nursing

Purpose

The purpose of this project is to describe a project for the development of an educational intervention focused on secondary EHR data in nursing quality improvement (QI) research.

Background

In Advancing Healthcare Transformation: A New Era for Academic Nursing, the American Association of Colleges of Nursing (AACN, 2016) make five recommendations based on their findings that academic nursing is not positioned as a partner in healthcare transformation. Their recommendation for academic nursing to invest in nursing research programs and better integrate research into clinical practice clearly intersects with the promise of big data (AACN, 2016). Big data and translation science were two emerging and priority areas identified by the Council for the Advancement of Nursing Science (Henly et al., 2015). Strengths of big data include availability from several sources to inform intervention research. Data sharing, faculty expertise and understanding of data science are barriers to using big data (Henly et al., 2015). Opportunities in translation science include collaboration between research and quality doctorates with clinical expertise, however a lack of commitment to dissemination, little opportunity for interdisciplinary collaboration and level of faculty expertise in methods of translation and implementation are threats (Henly et al., 2015).

Doctoral programs in nursing are designed to prepare nurse scientists and scholars, and Doctor of Nursing Practice (DNP) graduates must demonstrate the ability to translate knowledge to practice (AACN, 2006). Nursing participation, through collaboration, research and dissemination of data driven interventions, is important to representation and relevance of nursing knowledge and expertise in healthcare policy (Gephart, Davis, & Shea, 2018).

Recommendations from healthcare system leaders like Institute of Medicine (IOM), Centers for Medicare and Medicaid Services (CMS) and National Institutes of Health (NIH) all point to further development of continuous learning and translation of evidence into practice (Smith, 2012; CMS, 2018; NIH, 2017). Gaps in knowledge exist in best practice to achieve translation of data-driven evidence both to and from practice, and EHR data is a fundamental part of that translation (Frink, 2016). Based on an evaluation of health science students’ research skills, Kingsley and Kingsley (2009) make a strong suggestion to design and incorporate information literacy into graduate-level curricula. AACN’s Essentials of Doctoral Education for Advanced Nursing Practice (2006) established an essential competency for advance practice nursing to “utilize information systems to evaluate programs of care, outcomes of care and care systems.”

Methods

DNP students will receive an educational intervention in the form of a toolkit. The DNP QI Toolkit will include training on local policy for clinical research, data use agreements, best practice for accessing secondary data and foundational skills for clinical research. A toolkit is defined by the Agency for Healthcare Research and Quality (AHRQ, 2018) as a “collection of information, resources or tools to guide users to develop a plan or organize efforts to follow evidence-based recommendations or meet evidence-based specific practice standards.” In other words, a toolkit is an action-oriented technique of promoting the translation of research findings into policy and practice (AHRQ, 2018).

Non-invasive Biomarkers to Detect Acute Kidney Injury in Premature Infants

Terri Marin, PhD, NNP-BC, FAANP, Assistant Professor, College of Nursing, Augusta University

Bryan Williams, PhD, Professor, College of Nursing, Augusta University

Jatinda Bhatia, MD, FAAP, Section Chief, Department of Pediatrics: Neonatology, Augusta University

Ashok Sharma, PhD, Assistant Professor, Department of Population Health Science, Augusta University

Wenbo Zhi, PhD, Assistant Professor, Department of Obstetrics and Gynecology, Augusta University

Cynthia Mundy, DNP, NNP-BC, Assistant Professor, College of Nursing, Augusta University

Christy Cockfield, DNP, NNP-BC, Nurse Practitioner, Children’s Hospital of Georgia Neonatal Intensive Care Unit (NICU), Augusta University Health

Background: Acute kidney injury (AKI) occurs in approximately 40% of preterm infants born ≤ 34 weeks’ gestational age (GA), many of whom become hypotensive. This condition can injure the kidney and is both difficult to detect and treat. The immature kidney only receives 3-4% of total cardiac output under optimal conditions (vs. 20% in term infants); hence, mild reductions in perfusion secondary to hypotension may quickly result in renal ischemia. Once AKI develops, mortality rates increase to >50%. Current diagnostic criteria (serial serum creatinine measures) does not detect early, subclinical injury, as up to 50% of renal function is lost by the time this method detects AKI. 

Purpose:  We sought to determine the feasibility of non-invasive physiologic biomarkers to detect early, subclinical AKI in premature infants ≤ 34 weeks’ gestation during the first 14 days of life (DOL). We hypothesized that an inverse relationship exists between renal hypoxia and urinary biomarkers when ischemic renal injury is present.

Methods: We continuously observed regional renal oxygen saturation (rSO₂) using near infrared spectroscopy (NIRS) technology among 21 infants ≤ 34 weeks’ GA during the first 14 DOL. We also collected urinary AKI biomarker samples that indicate renal damage on DOL 3, 7, 10 and 14. Using non-parametric test statistics we estimated the extent to which rSO₂ slope changes were associated with that of each of the urinary biomarkers. 

Results: The rate of change in rSO₂ slope was positively associated with that observed in levels of osteopontin (r= -0.53) and Cystatin C (r= -0.53) urinary biomarkers (p ≤ 0.05).  There was no correlation between rSO₂ slope change and serum creatinine levels (r= -0.13; p=0.16).

Conclusions:  Renal oxygenation and urinary biomarker levels in infants experiencing hypoperfusion are inversely related and identify subclinical renal ischemia, prior to changes in serum creatinine.  Combining NIRS technology with urinary biomarker measurements may better predict and aid early diagnosis of subclinical AKI than serum creatinine measures in infants ≤ 34 weeks’ gestation during the first 14 DOL.

Opioid Crisis Trends in Georgia: Using Data Management Systems to Better Inform Public Policy

Authors: Nafiz Sheikh, Liniya Tauhidul

Introduction: The nationwide opioid epidemic is arguably the most consequential public health crisis of the new millennium. Unfortunately, a dearth of medical literature exists analyzing the scope of the epidemic in Georgia. This presentation will investigate trends in fatal opioid overdoses in Georgia using a robust healthcare data management system: Center for Disease Control’s (CDC) WONDER.

Methods: Using CDC WONDER, a cohort of all fatal opioid overdoses in Georgia from 1999 – 2017 was obtained (N=10,070). The group was then stratified by race, sex, age group, and overdosed opioid type. Time series analyses were used to determine trends, two-sided Chi-square tests with statistical significance set to p<0.05 used to compare opioid mortality proportions from different years and mortalities between age groups. Lastly findings were correlated to geography to ascertain if urbanization correlated to opioid mortality.

Results: Approximately 1056 fatal opioid overdoses occurred in 2017, up 192% from 550 deaths in 2010. Fatal overdoses from heroin and synthetic accounted for only 2% and 17% of total deaths in 2010 but magnified to 20% and 32% by 2017 (p<0.05). Beginning 2013, heroin and synthetic opioids such as fentanyl together drove Georgia opioid mortality sharply higher. Among different age groups, Georgians aged 25-34yrs experienced the highest mortalities compared to other females within the same age group (p<0.05) and to males and female in the 35-44yrs and 45-54yrs groups (p<0.05). Correlating fatalities to geography found urban areas in Atlanta, Augusta, and Columbus to have the highest mortality rates.

Conclusion: Georgians have experienced an unprecedented surge in mortality from opioid-related overdoses in recent years. With robust healthcare data management systems, however, new research endeavors are poised to generate more thorough epidemiological reports that will better inform local and state health policy.

SF12v2 Health Scores for African Americans in a Cluster-randomized Community Trial

Thomas V. Joshua¹; Jane T. Garvin²; Lucy Marion¹; Lovoria B. Williams³.

¹Augusta University, College of Nursing, ²University of St. Augustine for Health Sciences, ³University of Kentucky

Background: Self-administered health survey short form (SF12v2) is commonly used to assess the health related quality of life (HRQOL) among populations. However, there is lack of data regarding its effectiveness among African Americans (AAs).

Design: Secondary data analysis of a prospective cohort study.

Aim: To assess the quality of life among AAs enrolled in a faith-based diabetes prevention program, Fit Body and Soul (FBAS) compared to a health education (HE).

Methods: Data were collected at three time points; baseline, 12 weeks, and 52 weeks. SASv9.4 was used to score the data. The SF12v2 data calculate two summary component scores, Physical Component Summary Score (PCS) and Mental Health Component Summary Score (MCS) with eight sub-domains. Scores range from 0 to 100, where a zero score indicates the lowest level of health and 100 indicates the highest level of health. Both PCS and MCS combine the 12 items in such a way that they compare to a national norm with a mean score of 50.0 and a standard deviation of 10.

Results: Total of 604 people were enrolled. FBAS included 317(mean age=46.59±10.9) and 287 (mean age=46.39±10.9) were in the HE. General health was reported good or better for 85% of the sample in both groups at baseline. Overall PCS for FBAS was 49 at baseline, 51 at week 12, and 50 at week 52 and for HE was 48 at baseline, 49 at week 12 and 49 at week 52. Overall, MCS for FBAS was 51 at baseline, 53 at week 12, and 52 at week 52 and for HE was 51 at baseline, 52 at week 12 and 51 at week 52.

Conclusion: Quality of life among participants at week 12 was improved from baseline but not maintained at week 52. The SF-12v2 appears to be a valid survey tool for the assessment of HRQOL among AAs.

Authors:

V. Migdal MD, N. Haqqani MD, K. Harper MD, B. Janiak MD

Affiliations:

All Authors: Department of Emergency Medicine, Augusta University; Augusta University Medical Center

ABSTRACT

Purpose: To determine the time to obtain answers to five-preselected ED nursing triage/assessment questions and convert this time to the monetary cost to the emergency department.

Methods: A prospective observational study of ED Registered Nurses performing triage assessments on 200 adults presenting to the ED for medical care.  During the triage assessment, the nurse was timed by one of the study authors while the RN asked five pre-selected questions included in their current triage protocol.  The time cost of each question was determined by multiplying the time spent asking the question each year by the average hourly wage of our ED RNs. (T x V x S)/3600; where T= average time per question (in seconds), V=annual patient volume, S=average hourly RN wage

Results: A total of 200 triage assessments were observed. During these assessments, 130 patients were asked about pneumococcal vaccine status, 161 asked about tetanus vaccine status, 184 asked about medication allergies, 172 about influenza vaccine, and 73 about recent travel. The average time spent per question ranged from 4.4-6.3 seconds. The estimated annual time used to ask these 5 questions in the AUMC ED is 633.98 hrs, which equates to $22189.3 in nursing costs per year.

Conclusions: This is a cursory look at the potential monetary and time costs of standardized screening questions in the ED. These values directly affect time and cost efficiency in the ED process and could potentially be redirected to more pertinent patient care. The required screening questions are often unrelated to the patient’s chief complaint and have no impact on the medical management in the ED. Further studies are needed to determine cost effectiveness of required ED screenings.