Monica M. Horvath

The DEDUCE Guided Query tool

In many healthcare organizations, comparative effectiveness research and quality improvement (QI) investigations are hampered by a lack of access to data created as a byproduct of patient care. Data collection often hinges upon either manual chart review or ad hoc requests to technical experts who support legacy clinical systems. In order to facilitate this needed capacity for data exploration at our institution (Duke University Health System), we have designed and deployed a robust Web application for cohort identification and data extraction--the Duke Enterprise Data Unified Content Explorer (DEDUCE). DEDUCE is envisioned as a simple, web-based environment that allows investigators access to administrative, financial, and clinical information generated during patient care. By using business intelligence tools to create a view into Duke Medicines enterprise data warehouse, DEDUCE provides a Guided Query functionality using a wizard-like interface that lets users filter through millions of clinical records, explore aggregate reports, and, export extracts. Researchers and QI specialists can obtain detailed patient- and observation-level extracts without needing to understand structured query language or the underlying database model. Developers designing such tools must devote sufficient training and develop application safeguards to ensure that patient-centered clinical researchers understand when observation-level extracts should be used. This may mitigate the risk of data being misunderstood and consequently used in an improper fashion.

Reevaluating the Safety Profile of Pediatrics: A Comparison of Computerized Adverse Drug Event Surveillance and Voluntary Reporting in the Pediatric Environment

OBJECTIVES. Children are at exceptionally high risk for adverse drug events. At Duke University Hospital, computerized adverse drug event surveillance and voluntary safety reporting systems work synergistically to identify adverse drug events. Here we identify the most deleterious drug classes to pediatric inpatients and determine which detection methodology provides the greatest opportunity to reduce harm. PATIENTS AND METHODS. We evaluated all of the medication-related events detected by our computerized surveillance and safety reporting systems over a 1-year period for Duke University Hospital pediatric inpatients. Events from both systems were scored for severity and assigned a drug event category. Surveillance events were additionally scored for causality. RESULTS. A total of 849 medication-related reports were entered into the safety reporting system, and 93 caused patient harm, resulting in an adverse drug event rate of 1.8 events per 1000 pediatric patient-days. Seventy eight of the 1537 medication-related events detected by surveillance resulted in patient harm, giving a rate of 1.6 events per 1000 patient-days. The most common events identified by the safety reporting system were failures in the medication use process (26.9%), drug omissions (16.1%), and dose- or rate-related events (12.9%). The most frequent adverse drug event surveillance categories were nephrotoxins (20.7%), narcotics and benzodiazepines (19.3%), and hypoglycemia (11.5%). Most voluntarily reported events originated in ICUs (72.0%), whereas surveillance events were split evenly across intensive and general care. There was little overlap between methodologies. CONCLUSIONS. The epidemiology of pediatric adverse drug events is best addressed by using voluntary reporting in tandem with other strategies, such as computerized surveillance and targeted chart review. Although voluntary reporting excels at identifying administration errors, surveillance excels at detecting adverse drug events caused by high-risk medications and identifies evolving conditions that may provoke imminent patient harm. Surveillance underperformed in pediatrics when compared with adult detection rates, suggesting that tailored rules may be necessary for a robust pediatric adverse drug event surveillance system.

Impact of Health Portal Enrollment With Email Reminders on Adherence to Clinic Appointments: A Pilot Study

Background Internet portal technologies that provide access to portions of electronic health records have the potential to revolutionize patients’ involvement in their care. However, relatively few descriptions of the demographic characteristics of portal enrollees or of the effects of portal technology on quality outcomes exist. This study examined data from patients who attended one of seven Duke Medicine clinics and who were offered the option of enrolling in and using the Duke Medicine HealthView portal (HVP). The HVP allows patients to manage details of their appointment scheduling and provides automated email appointment reminders in addition to the telephone and mail reminders that all patients receive. Objective Our objective was to test whether portal enrollment with an email reminder functionality is significantly related to decreases in rates of appointment “no-shows,” which are known to impair clinic operational efficiency. Methods Appointment activity during a 1-year period was examined for all patients attending one of seven Duke Medicine clinics. Patients were categorized as portal enrollees or as nonusers either by their status at time of appointment or at the end of the 1-year period. Demographic characteristics and no-show rates among these groups were compared. A binomial logistic regression model was constructed to measure the adjusted impact of HVP enrollment on no-show rates, given confounding factors. To demonstrate the effect of HVP use over time, monthly no-show rates were calculated for patient appointment keeping and contrasted between preportal and postportal deployment periods. Results Across seven clinics, 58,942 patients, 15.7% (9239/58,942) of whom were portal enrollees, scheduled 198,199 appointments with an overall no-show rate of 9.9% (19,668/198,199). We found that HVP enrollees were significantly more likely to be female, white, and privately insured compared with nonusers. Bivariate no-show rate differences between portal enrollment groups varied widely according to patient- and appointment-level attributes. Large reductions in no-show rates were seen among historically disadvantaged groups: Medicaid holders (OR = 2.04 for nonuser/enrollee, 5.6% difference, P < .001), uninsured patients (OR = 2.60, 12.8% difference, P < .001), and black patients (OR = 2.13, 8.0% difference, P < .001). After fitting a binomial logistic regression model for the outcome of appointment arrival, the adjusted odds of arrival increased 39.0% for portal enrollees relative to nonusers (OR = 1.39, 95% CI 1.22 - 1.57, P < .001). Analysis of monthly no-show rates over 2 years demonstrated that patients who registered for portal access and received three reminders of upcoming appointments (email, phone, and mail) had a 2.0% no-show rate reduction (P < .001), whereas patients who did not enroll and only received traditional phone and mail reminders saw no such reduction (P < .09). Conclusions Monthly no-show rates across all seven Duke Medicine clinics were significantly reduced among patients who registered for portal use, suggesting that in combination with an email reminder feature, this technology may have an important and beneficial effect on clinic operations.

Using a computerized provider order entry system to meet the unique prescribing needs of children: description of an advanced dosing model

BackgroundIt is well known that the information requirements necessary to safely treat children with therapeutic medications cannot be met with the same approaches used in adults. Over a 1-year period, Duke University Hospital engaged in the challenging task of enhancing an established computerized provider order entry (CPOE) system to address the unique medication dosing needs of pediatric patients.MethodsAn advanced dosing model (ADM) was designed to interact with our existing CPOE application to provide decision support enabling complex pediatric dose calculations based on chronological age, gestational age, weight, care area in the hospital, indication, and level of renal impairment. Given that weight is a critical component of medication dosing that may change over time, alerting logic was added to guard against erroneous entry or outdated weight information.ResultsPediatric CPOE was deployed in a staggered fashion across 6 care areas over a 14-month period. Safeguards to prevent miskeyed values became important in allowing providers the flexibility to override the ADM logic if desired. Methods to guard against over- and under-dosing were added. The modular nature of our model allows us to easily add new dosing scenarios for specialized populations as the pediatric population and formulary change over time.ConclusionsThe medical needs of pediatric patients vary greatly from those of adults, and the information systems that support those needs require tailored approaches to design and implementation. When a single CPOE system is used for both adults and pediatrics, safeguards such as redirection and suppression must be used to protect children from inappropriate adult medication dosing content. Unlike other pediatric dosing systems, our model provides active dosing assistance and dosing process management, not just static dosing advice.

Clinical Research Informatics and Electronic Health Record Data

OBJECTIVES The goal of this survey is to discuss the impact of the growing availability of electronic health record (EHR) data on the evolving field of Clinical Research Informatics (CRI), which is the union of biomedical research and informatics. RESULTS Major challenges for the use of EHR-derived data for research include the lack of standard methods for ensuring that data quality, completeness, and provenance are sufficient to assess the appropriateness of its use for research. Areas that need continued emphasis include methods for integrating data from heterogeneous sources, guidelines (including explicit phenotype definitions) for using these data in both pragmatic clinical trials and observational investigations, strong data governance to better understand and control quality of enterprise data, and promotion of national standards for representing and using clinical data. CONCLUSIONS The use of EHR data has become a priority in CRI. Awareness of underlying clinical data collection processes will be essential in order to leverage these data for clinical research and patient care, and will require multi-disciplinary teams representing clinical research, informatics, and healthcare operations. Considerations for the use of EHR data provide a starting point for practical applications and a CRI research agenda, which will be facilitated by CRIs key role in the infrastructure of a learning healthcare system.

Computerized surveillance of opioid-related adverse drug events in perioperative care: a cross-sectional study

BackgroundGiven the complexity of surgical care, perioperative patients are at high risk of opioid-related adverse drug events. Existing methods of detection, such as trigger tools and manual chart review, are time-intensive which makes sustainability challenging. Using strategic rule design, computerized surveillance may be an efficient, pharmacist-driven model for event detection that leverages existing staff resources.MethodsComputerized adverse drug event surveillance uses a logic-based rules engine to identify potential adverse drug events or evolving unsafe clinical conditions. We extended an inpatient rule (administration of naloxone) to detect opioid-related oversedation and respiratory depression to perioperative care at a large academic medical center. Our primary endpoint was the adverse drug event rate. For all patients with a naloxone alert, manual chart review was performed by a perioperative clinical pharmacist to assess patient harm. In patients with confirmed oversedation, other patient safety event databases were queried to determine if they could detect duplicate, prior, or subsequent opioid-related events.ResultsWe identified 419 cases of perioperative naloxone administration. Of these, 101 were given postoperatively and 69 were confirmed as adverse drug events after chart review yielding a rate of 1.89 adverse drug events/1000 surgical encounters across both the inpatient and ambulatory settings. Our ability to detect inpatient opioid adverse drug events increased 22.7% by expanding surveillance into perioperative care. Analysis of historical surveillance data as well as a voluntary reporting database revealed that 11 of our perioperative patients had prior or subsequent harmful oversedation. Nine of these cases received intraoperative naloxone, and 2 had received naloxone in the post-anesthesia care unit. Pharmacist effort was approximately 3 hours per week to evaluate naloxone alerts and confirm adverse drug events.ConclusionA small investment of resources into a pharmacist-driven surveillance model gave great gains in organizational adverse drug event detection. The patients who experienced multiple events are particularly relevant to future studies seeking risk factors for opioid induced respiratory depression. Computerized surveillance is an efficient, impactful, and sustainable model for ongoing capture and analysis of these rare, but potentially serious events.

The design and implementation of an open-source, data-driven cohort recruitment system: the Duke Integrated Subject Cohort and Enrollment Research Network (DISCERN)

OBJECTIVE Failure to reach research subject recruitment goals is a significant impediment to the success of many clinical trials. Implementation of health-information technology has allowed retrospective analysis of data for cohort identification and recruitment, but few institutions have also leveraged real-time streams to support such activities. DESIGN Duke Medicine has deployed a hybrid solution, The Duke Integrated Subject Cohort and Enrollment Research Network (DISCERN), that combines both retrospective warehouse data and clinical events contained in prospective Health Level 7 (HL7) messages to immediately alert study personnel of potential recruits as they become eligible. RESULTS DISCERN analyzes more than 500000 messages daily in service of 12 projects. Users may receive results via email, text pages, or on-demand reports. Preliminary results suggest DISCERNs unique ability to reason over both retrospective and real-time data increases study enrollment rates while reducing the time required to complete recruitment-related tasks. The authors have introduced a preconfigured DISCERN function as a self-service feature for users. LIMITATIONS The DISCERN framework is adoptable primarily by organizations using both HL7 message streams and a data warehouse. More efficient recruitment may exacerbate competition for research subjects, and investigators uncomfortable with new technology may find themselves at a competitive disadvantage in recruitment. CONCLUSION DISCERNs hybrid framework for identifying real-time clinical events housed in HL7 messages complements the traditional approach of using retrospective warehoused data. DISCERN is helpful in instances when the required clinical data may not be loaded into the warehouse and thus must be captured contemporaneously during patient care. Use of an open-source tool supports generalizability to other institutions at minimal cost.

Sharing Adverse Drug Event Data Using Business Intelligence Technology

Introduction: Duke University Health System uses computerized adverse drug event surveillance as an integral part of medication safety at 2 community hospitals and an academic medical center. This information must be swiftly communicated to organizational patient safety stakeholders to find opportunities to improve patient care; however, this process is encumbered by highly manual methods of preparing the data. Description of Case: Following the examples of other industries, we deployed a business intelligence tool to provide dynamic safety reports on adverse drug events. Once data were migrated into the health system data warehouse, we developed census-adjusted reports with user-driven prompts. Drill down functionality enables navigation from aggregate trends to event details by clicking report graphics. Reports can be accessed by patient safety leadership either through an existing safety reporting portal or the health system performance improvement Web site. Discussion: Elaborate prompt screens allow many varieties of reports to be created quickly by patient safety personnel without consultation with the research analyst. The reduction in research analyst workload because of business intelligence implementation made this individual available to additional patient safety projects thereby leveraging their talents more effectively. Conclusions: Dedicated liaisons are essential to ensure clear communication between clinical and technical staff throughout the development life cycle. Design and development of the business intelligence model for adverse drug event data must reflect the eccentricities of the operational system, especially as new areas of emphasis evolve. Future usability studies examining the data presentation and access model are needed.

PCA safety data review after clinical decision support and smart pump technology implementation.

Introduction Medication errors account for 20% of medical errors in the United States with the largest risk at prescribing and administration. Analgesics or opioids are frequently used medications that can be associated with patient harm when prescribed or administered improperly. In an effort to decrease medication errors, Duke University Hospital implemented clinical decision support via computer provider order entry (CPOE) and “smart pump” technology, 2/2008, with the goal to decrease patient-controlled analgesia (PCA) adverse events. Methods This project evaluated PCA safety events, reviewing voluntary report system and adverse drug events via surveillance (ADE-S), on intermediate and step-down units preimplementation and postimplementation of clinical decision support via CPOE and PCA smart pumps for the prescribing and administration of opioids therapy in the adult patient requiring analgesia for acute pain. Discussion Voluntary report system and ADE-S PCA events decreased based upon 1000 PCA days; ADE-S PCA events per 1000 PCA days decreased 22%, from 5.3 (pre) to 4.2 (post) (P = 0.09). Voluntary report system events decreased 72%, from 2.4/1000 PCA days (pre) to 0.66/1000 PCA days (post) and was statistically significant (P < 0.001). There was a difference in the ADE-S data for causality (P < 0.0001) with sleep apnea and renal insufficiency approaching significance. Voluntary report system safety event were statistically significant for obesity [body mass index (BMI) ≥30] and weight. Conclusion This study demonstrated a decrease in PCA events between time periods in both the ADE-S and voluntary report system data, thus supporting the recommendation of clinical decision support via CPOE and PCA smart pump technology.

Characteristics of ambulatory anticoagulant adverse drug events: a descriptive study

BackgroundDespite the high frequency with which adverse drug events (ADEs) occur in outpatient settings, detailed information regarding these events remains limited. Anticoagulant drugs are associated with increased safety concerns and are commonly involved in outpatient ADEs. We therefore sought to evaluate ambulatory anticoagulation ADEs and the patient population in which they occurred within the Duke University Health System (Durham, NC, USA).MethodsA retrospective chart review of ambulatory warfarin-related ADEs was conducted. An automated trigger surveillance system identified eligible events in ambulatory patients admitted with an International Normalized Ratio (INR) >3 and administration of vitamin K. Event and patient characteristics were evaluated, and quality/process improvement strategies for ambulatory anticoagulation management are described.ResultsA total of 169 events in 167 patients were identified from December 1, 2006-June 30, 2008 and included in the study. A median supratherapeutic INR of 6.1 was noted, and roughly half of all events (52.1%) were associated with a bleed. Nearly 74% of events resulted in a need for fresh frozen plasma; 64.8% of bleeds were classified as major. A total of 59.2% of events were at least partially responsible for hospital admission. Median patient age was 68 y (range 36-95 y) with 24.9% initiating therapy within 3 months prior to the event. Of events with a prior documented patient visit (n = 157), 73.2% were seen at a Duke clinic or hospital within the previous month. Almost 80% of these patients had anticoagulation therapy addressed, but only 60.0% had a follow-up plan documented in the electronic note.ConclusionsAmbulatory warfarin-related ADEs have significant patient and healthcare utilization consequences in the form of bleeding events and associated hospital admissions. Recommendations for improvement in anticoagulation management include use of information technology to assist monitoring and follow-up documentation, avoid drug interactions, and engage patients in their care.