The Communicating Narrative Concerns Entered by Registered Nurses (CONCERN) Clinical Decision Support Early Warning System: Protocol for a Cluster Randomized Pragmatic Clinical Trial.

BACKGROUND: Every year, hundreds of thousands of inpatients die from cardiac arrest and sepsis, which could be avoided if those patients' risk for deterioration were detected and timely interventions were initiated. Thus, a system is needed to convert real-time, raw patient data into consumable information that clinicians can utilize to identify patients at risk of deterioration and thus prevent mortality and improve patient health outcomes. The overarching goal of the COmmunicating Narrative Concerns Entered by Registered Nurses (CONCERN) study is to implement and evaluate an early warning score system that provides clinical decision support (CDS) in electronic health record systems. With a combination of machine learning and natural language processing, the CONCERN CDS utilizes nursing documentation patterns as indicators of nurses' increased surveillance to predict when patients are at the risk of clinical deterioration. OBJECTIVE: The objective of this cluster randomized pragmatic clinical trial is to evaluate the effectiveness and usability of the CONCERN CDS system at 2 different study sites. The specific aim is to decrease hospitalized patients' negative health outcomes (in-hospital mortality, length of stay, cardiac arrest, unanticipated intensive care unit transfers, and 30-day hospital readmission rates). METHODS: A multiple time-series intervention consisting of 3 phases will be performed through a 1-year period during the cluster randomized pragmatic clinical trial. Phase 1 evaluates the adoption of our algorithm through pilot and trial testing, phase 2 activates optimized versions of the CONCERN CDS based on experience from phase 1, and phase 3 will be a silent release mode where no CDS is viewable to the end user. The intervention deals with a series of processes from system release to evaluation. The system release includes CONCERN CDS implementation and user training. Then, a mixed methods approach will be used with end users to assess the system and clinician perspectives. RESULTS: Data collection and analysis are expected to conclude by August 2022. Based on our previous work on CONCERN, we expect the system to have a positive impact on the mortality rate and length of stay. CONCLUSIONS: The CONCERN CDS will increase team-based situational awareness and shared understanding of patients predicted to be at risk for clinical deterioration in need of intervention to prevent mortality and associated harm. TRIAL REGISTRATION: ClinicalTrials.gov NCT03911687; https://clinicaltrials.gov/ct2/show/NCT03911687. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/30238.

Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers.

Deep learning has recently seen rapid development and received significant attention due to its state-of-the-art performance on previously-thought hard problems. However, because of the internal complexity and nonlinear structure of deep neural networks, the underlying decision making processes for why these models are achieving such performance are challenging and sometimes mystifying to interpret. As deep learning spreads across domains, it is of paramount importance that we equip users of deep learning with tools for understanding when a model works correctly, when it fails, and ultimately how to improve its performance. Standardized toolkits for building neural networks have helped democratize deep learning; visual analytics systems have now been developed to support model explanation, interpretation, debugging, and improvement. We present a survey of the role of visual analytics in deep learning research, which highlights its short yet impactful history and thoroughly summarizes the state-of-the-art using a human-centered interrogative framework, focusing on the Five W's and How (Why, Who, What, How, When, and Where). We conclude by highlighting research directions and open research problems. This survey helps researchers and practitioners in both visual analytics and deep learning to quickly learn key aspects of this young and rapidly growing body of research, whose impact spans a diverse range of domains.

Decreased Rates of Inpatient Hypoglycemia Following Implementation of an Automated Tool in the Electronic Medical Record for Identifying Root Causes.

OBJECTIVE: The objective was to identify root causes of hypoglycemia on medicine inpatient units using an automated tool. Data collected will guide educational interventions aimed at improving patient care and safety by decreasing rates of hypoglycemia. METHODS: A survey was conducted among RNs to identify risk factors for hypoglycemia. Survey data were used to create a hypoglycemia root cause survey tool in the EMR. RNs were prompted to utilize the tool when blood glucose (BG) < 70 mg/dL. Once the most common modifiable cause of hypoglycemia was identified, an educational intervention for safe and effective use of insulin was launched. This strategy was designed to empower the care team to reduce the insulin dose when appropriate to prevent future hypoglycemic episodes. RESULTS: BG data were compared from March and April in 2016 and 2017. Rates of hypoglycemia (BG < 70 mg/dL) decreased from 2.3% to 1.5%; BG values in target range (70-180 mg/dL) increased from 59.4% to 65.7%; hyperglycemia (BG > 180 mg/dL) decreased from 38.3% to 32.8% (all P values < .001). The number of patients with recurrent hypoglycemia (3 or more episodes) decreased from 5.7% to 2.2% ( P = .044). CONCLUSIONS: The two most frequent modifiable causes of hypoglycemia (insulin and nutrition) were identified by an RN survey and confirmed by chart review. A targeted educational intervention addressing safe and effective insulin dosing resulted in a significant decrease in both hypoglycemia and recurrent hypoglycemia. This was associated with an improvement in overall glycemic control. Ongoing clinician education regarding insulin and nutrition accompanied by discussions between RNs and prescribers to address hypoglycemic events in real-time could continue to lower the rate of occurrence.

Designing a clinical dashboard to fill information gaps in the emergency department.

Data fragmentation within electronic health records causes gaps in the information readily available to clinicians. We investigated the information needs of emergency medicine clinicians in order to design an electronic dashboard to fill information gaps in the emergency department. An online survey was distributed to all emergency medicine physicians at a large, urban academic medical center. The survey response rate was 48% (52/109). The clinical information items reported to be most helpful while caring for patients in the emergency department were vital signs, electrocardiogram (ECG) reports, previous discharge summaries, and previous lab results. Brief structured interviews were also conducted with 18 clinicians during their shifts in the emergency department. From the interviews, three themes emerged: 1) difficulty accessing vital signs, 2) difficulty accessing point-of-care tests, and 3) difficulty comparing the current ECG with the previous ECG. An emergency medicine clinical dashboard was developed to address these difficulties.

Enabling claims-based decision support through non-interruptive capture of admission diagnoses and provider billing codes.

The patient problem list, like administrative claims data, has become an important source of data for decision support, patient cohort identification, and alerting systems. A two-fold intervention to increase capture of problems on the problem list automatically - with minimal disruption to admitting and provider billing workflows - is described. For new patients with no prior data in the electronic health record, the intervention resulted in a statistically significant increase in the number of problems recorded to the problem list (3.8 vs 2.9 problems post-and pre-intervention respectively, p value 2×10(-16)). The majority of problems were recorded in the first 24 hours of admission. The proportion of patients with at least one problem coded to the problem list within the first 24 hours increased from 94% to 98% before and after intervention (chi square 344, p value 2×10(-16)). ICD9 "V codes" connoting circumstances beyond disease were captured at a higher rate post intervention than before. Deyo/Charlson comorbidities derived from problem list data were more similar to those derived from claims data after the intervention than before (Jaccard similarity 0.3 post- vs 0.21 pre-intervention, p value 2×10(-16)). A workflow-sensitive, non-interruptive means of capturing provider-entered codes early in admission can improve both the quantity and content of problems on the patient problem list.

Implementation of a computerized patient handoff application.

For hospitalized patients, handoffs between providers affect continuity of care and increase the risk of medical errors. Most commercial electronic health record (EHR) systems lack dedicated tools to support patient handoff activities. We developed a collaborative application supporting patient handoff that is fully integrated with our commercial EHR. The application creates user-customizable printed reports with automatic inclusion of a variety of EHR data, including: allergies, medications, 24-hour vital signs, recent common laboratory test results, isolation requirements, and code status. It has achieved widespread voluntary use at our institution (6,100 monthly users; 700 daily reports generated), and we have distributed the application to several other institutions using the same EHR. Though originally designed for resident physicians, today about 50% of the application users are nurses, 40% are physicians/physician assistants/nurse practitioners, and 10% are pharmacists, social workers, and other allied health providers.

Workarounds used by nurses to overcome design constraints of electronic health records.

Adoption of electronic health records (EHRs) has the potential to assist with clinical reasoning and streamline workflow; however, the data entry and review capabilities of most systems are suboptimal which may lead to workarounds. As an instance of a workaround, we examined nurses' use of optional free-text comments in EHR flowsheets to support clinical needs for data interpretation. This mixed-method study included: 1) Content analysis of comments, 2) Interviews with nurses. We performed a sub-analysis of flowsheet data for 201 patients that experienced a cardiac arrest and interviewed 5 acute care nurses. We found that nurses used workarounds in the EHR - despite the extra effort that they required - to convey clinically significant relationships and to communicate concerning events to physicians. EHRs should better support entry of clinical data that "belongs together" and enable messaging capabilities integrated with nurses' flowsheet documentation workflow.

Chemokine receptor CXCR4-mediated transformation of mammary epithelial cells by enhancing multiple RTKs expression and deregulation of the p53/MDM2 axis.

Recent studies have shown that CXCR4 is associated with tumor metastasis. Elevated levels of CXCR4 are also detected in a high percentage of DCIS cases. The high frequency of CXCR4 expression in DCIS suggests that many DCIS cases are "primed" for invasiveness. In this study, we demonstrated that expression of CXCR4 reveals morphological alterations in cells, from normal acinar morphological epithelial cells to a more invasive morphology in a 3D-culture system. Ectopic expression of CXCR4 induces invasion of MCF-10A cells. Interestingly, CXCR4 is capable of orchestrating a complex alteration in signaling networks, which include upregulation of multiple receptor tyrosine kinases (RTKs), deregulation of p53/MDM2 axis, upregulation of E-cadherin and c-myc, as well as modulation of cell cycle molecules to facilitate mammary epithelia cell transformation. These findings reveal that CXCR4 expression exerts a critical role in early stages of breast lesions, which may explain the high frequency of CXCR4 expression detected in DCIS. We believe that these studies will lead to new, biologically-based therapeutic strategies for clinical intervention, prevention and treatments of breast cancer.

Use of electronic clinical documentation: time spent and team interactions.

OBJECTIVE: To measure the time spent authoring and viewing documentation and to study patterns of usage in healthcare practice. DESIGN: Audit logs for an electronic health record were used to calculate rates, and social network analysis was applied to ascertain usage patterns. Subjects comprised all care providers at an urban academic medical center who authored or viewed electronic documentation. MEASUREMENT: Rate and time of authoring and viewing clinical documentation, and associations among users were measured. RESULTS: Users spent 20-103 min per day authoring notes and 7-56 min per day viewing notes, with physicians spending less than 90 min per day total. About 16% of attendings' notes, 8% of residents' notes, and 38% of nurses' notes went unread by other users, and, overall, 16% of notes were never read by anyone. Viewing of notes dropped quickly with the age of the note, but notes were read at a low but measurable rate, even after 2 years. Most healthcare teams (77%) included a nurse, an attending, and a resident, and those three users' groups were the first to write notes during an admission. Limitations The limitations were restriction to a single academic medical center and use of log files without direct observation. CONCLUSIONS: Care providers spend a significant amount of time viewing and authoring notes. Many notes are never read, and rates of usage vary significantly by author and viewer. While the rate of viewing a note drops quickly with its age, even after 2 years inpatient notes are still viewed.

Awareness of the Care Team in Electronic Health Records.

OBJECTIVE: To support collaboration and clinician-targeted decision support, electronic health records (EHRs) must contain accurate information about patients' care providers. The objective of this study was to evaluate two approaches for care provider identification employed within a commercial EHR at a large academic medical center. METHODS: We performed a retrospective review of EHR data for 121 patients in two cardiology wards during a four-week period. System audit logs of chart accesses were analyzed to identify the clinicians who were likely participating in the patients' hospital care. The audit log data were compared with two functions in the EHR for documenting care team membership: 1) a vendor-supplied module called "Care Providers", and 2) a custom "Designate Provider" order that was created primarily to improve accuracy of the attending physician of record documentation. RESULTS: For patients with a 3-5 day hospital stay, an average of 30.8 clinicians accessed the electronic chart, including 10.2 nurses, 1.4 attending physicians, 2.3 residents, and 5.4 physician assistants. The Care Providers module identified 2.7 clinicians/patient (1.8 attending physicians and 0.9 nurses). The Designate Provider order identified 2.1 clinicians/patient (1.1 attending physicians, 0.2 resident physicians, and 0.8 physician assistants). Information about other members of patients' care teams (social workers, dietitians, pharmacists, etc.) was absent. CONCLUSIONS: The two methods for specifying care team information failed to identify numerous individuals involved in patients' care, suggesting that commercial EHRs may not provide adequate tools for care team designation. Improvements to EHR tools could foster greater collaboration among care teams and reduce communication-related risks to patient safety.