Engaging Multidisciplinary Clinical Users in the Design of an Artificial Intelligence-Powered Graphical User Interface for Intensive Care Unit Instability Decision Support

BACKGROUND

Critical instability forecast and treatment can be optimized by artificial intelligence (AI)-enabled clinical decision support. It is important that the user-facing display of AI output facilitates clinical thinking and workflow for all disciplines involved in bedside care.

OBJECTIVES

Our objective is to engage multidisciplinary users (physicians, nurse practitioners, physician assistants) in the development of a graphical user interface (GUI) to present an AI-derived risk score.

METHODS

Intensive care unit (ICU) clinicians participated in focus groups seeking input on instability risk forecast presented in a prototype GUI. Two stratified rounds (three focus groups [only nurses, only providers, then combined]) were moderated by a focus group methodologist. After round 1, GUI design changes were made and presented in round 2. Focus groups were recorded, transcribed, and deidentified transcripts independently coded by three researchers. Codes were coalesced into emerging themes.

RESULTS

Twenty-three ICU clinicians participated (11 nurses, 12 medical providers [3 mid-level and 9 physicians]). Six themes emerged: (1) analytics transparency, (2) graphical interpretability, (3) impact on practice, (4) value of trend synthesis of dynamic patient data, (5) decisional weight (weighing AI output during decision-making), and (6) display location (usability, concerns for patient/family GUI view). Nurses emphasized having GUI objective information to support communication and optimal GUI location. While providers emphasized need for recommendation interpretability and concern for impairing trainee critical thinking. All disciplines valued synthesized views of vital signs, interventions, and risk trends but were skeptical of placing decisional weight on AI output until proven trustworthy.

CONCLUSION

Gaining input from all clinical users is important to consider when designing AI-derived GUIs. Results highlight that health care intelligent decisional support systems technologies need to be transparent on how they work, easy to read and interpret, cause little disruption to current workflow, as well as decisional support components need to be used as an adjunct to human decision-making.

Weakly Supervised Classification of Vital Sign Alerts as Real or Artifact

A significant proportion of clinical physiologic monitoring alarms are false. This often leads to alarm fatigue in clinical personnel, inevitably compromising patient safety. To combat this issue, researchers have attempted to build Machine Learning (ML) models capable of accurately adjudicating Vital Sign (VS) alerts raised at the bedside of hemodynamically monitored patients as real or artifact. Previous studies have utilized supervised ML techniques that require substantial amounts of hand-labeled data. However, manually harvesting such data can be costly, time-consuming, and mundane, and is a key factor limiting the widespread adoption of ML in healthcare (HC). Instead, we explore the use of multiple, individually imperfect heuristics to automatically assign probabilistic labels to unlabeled training data using weak supervision. Our weakly supervised models perform competitively with traditional supervised techniques and require less involvement from domain experts, demonstrating their use as efficient and practical alternatives to supervised learning in HC applications of ML.

Engaging clinicians early during the development of a graphical user display of an intelligent alerting system at the bedside

BACKGROUND

Artificial Intelligence (AI) is increasingly used to support bedside clinical decisions, but information must be presented in usable ways within workflow. Graphical User Interfaces (GUI) are front-facing presentations for communicating AI outputs, but clinicians are not routinely invited to participate in their design, hindering AI solution potential.

PURPOSE

To inform early user-engaged design of a GUI prototype aimed at predicting future Cardiorespiratory Insufficiency (CRI) by exploring clinician methods for identifying at-risk patients, previous experience with implementing new technologies into clinical workflow, and user perspectives on GUI screen changes.

METHODS

We conducted a qualitative focus group study to elicit iterative design feedback from clinical end-users on an early GUI prototype display. Five online focus group sessions were held, each moderated by an expert focus group methodologist. Iterative design changes were made sequentially, and the updated GUI display was presented to the next group of participants.

RESULTS

23 clinicians were recruited (14 nurses, 4 nurse practitioners, 5 physicians; median participant age ∼35 years; 60% female; median clinical experience 8 years). Five themes emerged from thematic content analysis: trend evolution, context (risk evolution relative to vital signs and interventions), evaluation/interpretation/explanation (sub theme: continuity of evaluation), clinician intuition, and clinical operations. Based on these themes, GUI display changes were made. For example, color and scale adjustments, integration of clinical information, and threshold personalization.

CONCLUSIONS

Early user-engaged design was useful in adjusting GUI presentation of AI output. Next steps involve clinical testing and further design modification of the AI output to optimally facilitate clinician surveillance and decisions. Clinicians should be involved early and often in clinical decision support design to optimize efficacy of AI tools.

Intensive Care Unit Scoring Systems

BACKGROUND

Illness severity scoring systems are commonly used in critical care. When applied to the populations for whom they were developed and validated, these tools can facilitate mortality prediction and risk stratification, optimize resource use, and improve patient outcomes.

OBJECTIVE

To describe the characteristics and applications of the scoring systems most frequently applied to critically ill patients.

METHODS

A literature search was performed using MEDLINE to identify original articles on intensive care unit scoring systems published in the English language from 1980 to 2020. Search terms associated with critical care scoring systems were used alone or in combination to find relevant publications.

RESULTS

Two types of scoring systems are most frequently applied to critically ill patients: those that predict risk of in-hospital mortality at the time of intensive care unit admission (Acute Physiology and Chronic Health Evaluation, Simplified Acute Physiology Score, and Mortality Probability Models) and those that assess and characterize current degree of organ dysfunction (Multiple Organ Dysfunction Score, Sequential Organ Failure Assessment, and Logistic Organ Dysfunction System). This article details these systems' differing features and timing of use, score calculation, patient populations, and comparative performance data.

CONCLUSION

Critical care nurses must be aware of the strengths, limitations, and specific characteristics of severity scoring systems commonly used in intensive care unit patients to effectively employ these tools in clinical practice and critically appraise research findings based on their use.

Accuracy of identifying hospital acquired venous thromboembolism by administrative coding: implications for big data and machine learning research

Big data analytics research using heterogeneous electronic health record (EHR) data requires accurate identification of disease phenotype cases and controls. Overreliance on ground truth determination based on administrative data can lead to biased and inaccurate findings. Hospital-acquired venous thromboembolism (HA-VTE) is challenging to identify due to its temporal evolution and variable EHR documentation. To establish ground truth for machine learning modeling, we compared accuracy of HA-VTE diagnoses made by administrative coding to manual review of gold standard diagnostic test results. We performed retrospective analysis of EHR data on 3680 adult stepdown unit patients identifying HA-VTE. International Classification of Diseases, Ninth Revision (ICD-9-CM) codes for VTE were identified. 4544 radiology reports associated with VTE diagnostic tests were screened using terminology extraction and then manually reviewed by a clinical expert to confirm diagnosis. Of 415 cases with ICD-9-CM codes for VTE, 219 were identified with acute onset type codes. Test report review identified 158 new-onset HA-VTE cases. Only 40% of ICD-9-CM coded cases (n = 87) were confirmed by a positive diagnostic test report, leaving the majority of administratively coded cases unsubstantiated by confirmatory diagnostic test. Additionally, 45% of diagnostic test confirmed HA-VTE cases lacked corresponding ICD codes. ICD-9-CM coding missed diagnostic test-confirmed HA-VTE cases and inaccurately assigned cases without confirmed VTE, suggesting dependence on administrative coding leads to inaccurate HA-VTE phenotyping. Alternative methods to develop more sensitive and specific VTE phenotype solutions portable across EHR vendor data are needed to support case-finding in big-data analytics.

Educating PhD students in research-intensive nursing doctorate programs regarding teaching competencies

In October 2019 an invitational summit was held addressing nursing PhD program competencies within research-intensive universities. One topic of discussion was related to whether or not teaching competencies should be included in the PhD program curricula of research-intensive universities, and where competencies should be learned. The discussion indicated a lack of uniform consensus. Rather, schools should be clear about their goals-to focus solely on developing nurse scientists, or a broader mission of preparing graduates to embrace the full scope of academic work inclusive of discovery, teaching, application and integration, or even possibly roles outside of academia. The discussion group coalesced around the notion that preparation in teaching be dependent upon mission clarification. Schools could then decide whether to incorporate teaching competencies or whether the best way to achieve their mission was to limit the acquisition of teaching competencies to elective experiential learning alone, or a combination of didactics, supervised practica and experiential learning. Based upon the summit conversation, this is something that each PhD program will have to decide based upon its own purpose and the environment it is preparing graduates to occupy. Nevertheless, preparing for a too narrow and specific career trajectory may not accommodate flexibility in the marketplace.

Prediction of hypotension events with physiologic vital sign signatures in the intensive care unit

BACKGROUND

Even brief hypotension is associated with increased morbidity and mortality. We developed a machine learning model to predict the initial hypotension event among intensive care unit (ICU) patients and designed an alert system for bedside implementation.

MATERIALS AND METHODS

From the Medical Information Mart for Intensive Care III (MIMIC-3) dataset, minute-by-minute vital signs were extracted. A hypotension event was defined as at least five measurements within a 10-min period of systolic blood pressure ≤ 90 mmHg and mean arterial pressure ≤ 60 mmHg. Using time series data from 30-min overlapping time windows, a random forest (RF) classifier was used to predict risk of hypotension every minute. Chronologically, the first half of extracted data was used to train the model, and the second half was used to validate the trained model. The model's performance was measured with area under the receiver operating characteristic curve (AUROC) and area under the precision recall curve (AUPRC). Hypotension alerts were generated using risk score time series, a stacked RF model. A lockout time were applied for real-life implementation.

RESULTS

We identified 1307 subjects (1580 ICU stays) as the hypotension group and 1619 subjects (2279 ICU stays) as the non-hypotension group. The RF model showed AUROC of 0.93 and 0.88 at 15 and 60 min, respectively, before hypotension, and AUPRC of 0.77 at 60 min before. Risk score trajectories revealed 80% and > 60% of hypotension predicted at 15 and 60 min before the hypotension, respectively. The stacked model with 15-min lockout produced on average 0.79 alerts/subject/hour (sensitivity 92.4%).

CONCLUSION

Clinically significant hypotension events in the ICU can be predicted at least 1 h before the initial hypotension episode. With a highly sensitive and reliable practical alert system, a vast majority of future hypotension could be captured, suggesting potential real-life utility.

Tele-Critical Care: An Update From the Society of Critical Care Medicine Tele-ICU Committee

OBJECTIVES

In 2014, the Tele-ICU Committee of the Society of Critical Care Medicine published an article regarding the state of ICU telemedicine, one better defined today as tele-critical care. Given the rapid evolution in the field, the authors now provide an updated review.

DATA SOURCES AND STUDY SELECTION

We searched PubMed and OVID for peer-reviewed literature published between 2010 and 2018 related to significant developments in tele-critical care, including its prevalence, function, activity, and technologies. Search terms included electronic ICU, tele-ICU, critical care telemedicine, and ICU telemedicine with appropriate descriptors relevant to each sub-section. Additionally, information from surveys done by the Society of Critical Care Medicine was included given the relevance to the discussion and was referenced accordingly.

DATA EXTRACTION AND DATA SYNTHESIS

Tele-critical care continues to evolve in multiple domains, including organizational structure, technologies, expanded-use case scenarios, and novel applications. Insights have been gained in economic impact and human and organizational factors affecting tele-critical care delivery. Legislation and credentialing continue to significantly influence the pace of tele-critical care growth and adoption.

CONCLUSIONS

Tele-critical care is an established mechanism to leverage critical care expertise to ICUs and beyond, but systematic research comparing different models, approaches, and technologies is still needed.

Clinical Distancing and Mitigation of Coronavirus Disease 2019

Social distancing as a technique to limit transmission of infectious disease has come into common parlance following the arrival and rapid spread of a novel coronavirus disease around the world in 2019 and 2020. But in the face of an emerging pandemic threat, it is crucial that we start to apply these principles to the clinic, the emergency department, and the hospital ward. We propose that this dynamic situation calls for a parallel "Clinical Distancing" in which we as a medical culture go against many of our fundamental instincts and, at least in the short term, begin to reduce unnecessary patient-care contacts for the benefit of our patients and our ability to continue to provide care to those who need it most. In this commentary, we provide specific recommendations for the rapid implementation of clinical distancing techniques.

Determinants of Intensive Care Unit Telemedicine Effectiveness. An Ethnographic Study

RATIONALE

Telemedicine is an increasingly common care delivery strategy in the ICU. However, ICU telemedicine programs vary widely in their clinical effectiveness, with some studies showing a large mortality benefit and others showing no benefit or even harm.

OBJECTIVES

To identify the organizational factors associated with ICU telemedicine effectiveness.

METHODS

We performed a focused ethnographic evaluation of 10 ICU telemedicine programs using site visits, interviews, and focus groups in both facilities providing remote care and the target ICUs. Programs were selected based on their change in risk-adjusted mortality after adoption (decreased mortality, no change in mortality, and increased mortality). We used a constant comparative approach to guide data collection and analysis.

MEASUREMENTS AND MAIN RESULTS

We conducted 460 hours of direct observation, 222 interviews, and 18 focus groups across six telemedicine facilities and 10 target ICUs. Data analysis revealed three domains that influence ICU telemedicine effectiveness: 1) leadership (i.e., the decisions related to the role of the telemedicine, conflict resolution, and relationship building), 2) perceived value (i.e., expectations of availability and impact, staff satisfaction, and understanding of operations), and 3) organizational characteristics (i.e., staffing models, allowed involvement of the telemedicine unit, and new hire orientation). In the most effective telemedicine programs these factors led to services that are viewed as appropriate, integrated, responsive, and consistent.

CONCLUSIONS

The effectiveness of ICU telemedicine programs may be influenced by several potentially modifiable factors within the domains of leadership, perceived value, and organizational structure.

Relationship between adherence to secondary prevention and health literacy, self-efficacy and disease knowledge among patients with coronary artery disease in China

Background:

Adherence to secondary prevention is an indispensable part of the management of patients with coronary artery disease. Finding patient factors affecting their adherence behaviours is important for improving the treatment effect and limiting further disease progression.

Aims:

To examine the association between health literacy, self-efficacy, disease knowledge and adherence to secondary coronary artery disease prevention in patients in China.

Methods:

In this cross-sectional study, 598 patients with coronary artery disease were enrolled in two tertiary hospitals in China during a hospitalisation for receiving percutaneous coronary intervention. Patient-reported data were collected on demographics, health literacy, self-efficacy, disease knowledge and adherence to secondary coronary artery disease prevention (medication-taking and heart-healthy lifestyle (exercise, reducing/eliminating alcohol intake and smoking, low salt and fat diet, stress reduction)). Chi-squared tests and regression analyses were performed.

Results:

The proportions of recalled self-report of adherence to medication-taking and a heart-healthy lifestyle immediately prior to the coronary artery disease hospitalisation were 84.7% and 53.2%, respectively. In logistic regression, health literacy, self-efficacy and disease knowledge was significantly associated with non-adherence to secondary coronary artery disease prevention. Limited health literacy demonstrated a 1.61-fold odds for non-adherence to a heart-healthy lifestyle. Each score increase of self-efficacy and disease knowledge had 0.98-fold odds and 1.05-fold odds of non-adherence to a heart-healthy lifestyle.

Conclusions:

Adherence to medication-taking was relatively good in Chinese patients prior to coronary artery disease hospitalisation, but adherence to heart-healthy lifestyle behaviours should be improved. Health literacy, self-efficacy and disease knowledge should be taken into account when intervening to improve secondary coronary artery disease prevention.

Predicting tachycardia as a surrogate for instability in the intensive care unit

Tachycardia is a strong though non-specific marker of cardiovascular stress that proceeds hemodynamic instability. We designed a predictive model of tachycardia using multi-granular intensive care unit (ICU) data by creating a risk score and dynamic trajectory. A subset of clinical and numerical signals were extracted from the Multiparameter Intelligent Monitoring in Intensive Care II database. A tachycardia episode was defined as heart rate ≥ 130/min lasting for ≥ 5 min, with ≥ 10% density. Regularized logistic regression (LR) and random forest (RF) classifiers were trained to create a risk score for upcoming tachycardia. Three different risk score models were compared for tachycardia and control (non-tachycardia) groups. Risk trajectory was generated from time windows moving away at 1 min increments from the tachycardia episode. Trajectories were computed over 3 hours leading up to the episode for three different models. From 2809 subjects, 787 tachycardia episodes and 707 control periods were identified. Patients with tachycardia had increased vasopressor support, longer ICU stay, and increased ICU mortality than controls. In model evaluation, RF was slightly superior to LR, which accuracy ranged from 0.847 to 0.782, with area under the curve from 0.921 to 0.842. Risk trajectory analysis showed average risks for tachycardia group evolved to 0.78 prior to the tachycardia episodes, while control group risks remained < 0.3. Among the three models, the internal control model demonstrated evolving trajectory approximately 75 min before tachycardia episode. Clinically relevant tachycardia episodes can be predicted from vital sign time series using machine learning algorithms.

A call to alarms: Current state and future directions in the battle against alarm fatigue

Research demonstrates that the majority of alarms derived from continuous bedside monitoring devices are non-actionable. This avalanche of unreliable alerts causes clinicians to experience sensory overload when attempting to sort real from false alarms, causing desensitization and alarm fatigue, which in turn leads to adverse events when true instability is neither recognized nor attended to despite the alarm. The scope of the problem of alarm fatigue is broad, and its contributing mechanisms are numerous. Current and future approaches to defining and reacting to actionable and non-actionable alarms are being developed and investigated, but challenges in impacting alarm modalities, sensitivity and specificity, and clinical activity in order to reduce alarm fatigue and adverse events remain. A multi-faceted approach involving clinicians, computer scientists, industry, and regulatory agencies is needed to battle alarm fatigue.

Clinical Presentation to the Emergency Department Predicts Subarachnoid Hemorrhage-Associated Myocardial Injury

INTRODUCTION

Aneurysmal subarachnoid hemorrhage (aSAH) is frequently seen in emergency departments. Secondary injury, such as subarachnoid hemorrhage-associated myocardial injury (SAHMI), affects one third of survivors and contributes to poor outcomes. SAHMI is not attributed to ischemia from myocardial disease but can result in hypotension and arrhythmias. It is important that emergency nurses recognize which clinical presentation characteristics are predictive of SAHMI to initiate proper interventions. The aim of this study was to determine whether patients who present to the emergency department with clinical aSAH are likely to develop SAHMI, as defined by troponin I ≥0.3 ng/mL.

METHODS

This was a prospective descriptive study. SAHMI was defined as troponin I ≥0.3 ng/mL. Predictors included demographics and clinical characteristics, severity of injury, admission 12-lead electrogardiogram (ECG), initial emergency department vital signs, and pre-hospital symptoms at time of aneurysm rupture.

RESULTS

Of 449 patients, 126 (28%) had SAHMI. Patients with SAHMI were more likely to report seizures and unresponsiveness with significantly lower Glasgow coma score and higher proportion of Hunt and Hess grades 3 to 5 and Fisher grades III and IV (all P < .05). Patients with SAHMI had higher atrial and ventricular rates and longer QTc intervals on initial ECG (P < .05). On multivariable logistic regression, poor Hunt and Hess grade, report of prehospital unresponsiveness, lower admission Glasgow coma score, and longer QTc interval were significantly and independently predictive of SAHMI (P < .05).

DISCUSSION

Components of the clinical presentation of subarachnoid hemorrhage to the emergency department predict SAHMI. Identifying patients with SAHMI in the emergency department can be helpful in determining surveillance and care needs and informing transfer unit care. Contribution to Emergency Nursing Practice.

The Relationships Between BNP and Neurocardiac Injury Severity, Noninvasive Cardiac Output, and Outcomes After Aneurysmal Subarachnoid Hemorrhage

INTRODUCTION

Neurocardiac injury, a type of myocardial dysfunction associated with neurological insult to the brain, occurs in 31-48% of aneurysmal subarachnoid hemorrhage (aSAH) patients. Cardiac troponin I (cTnI) is commonly used to diagnose neurocardiac injury. Brain natriuretic peptide (BNP), another cardiac marker, is more often used to evaluate degree of heart failure. The purpose of this study was to examine the relationships between BNP and (a) neurocardiac injury severity according to cTnI, (b) noninvasive continuous cardiac output (NCCO), and (c) outcomes in aSAH patients.

METHOD

This descriptive longitudinal study enrolled 30 adult aSAH patients. Data collected included BNP and cTnI levels and NCCO parameters for 14 days and outcomes (modified Rankin Scale [mRS] and mortality) at discharge and 3 months. Generalized estimating equations were used to evaluate associations between BNP and cTnI, NCCO, and outcomes.

RESULTS

BNP was significantly associated with cTnI. For every 1 unit increase in log BNP, cTnI increased by 0.05 ng/ml ( p = .001). Among NCCO parameters, BNP was significantly associated with thoracic fluid content ( p = .0003). On multivariable analyses, significant associations were found between BNP and poor mRS. For every 1 unit increase in log BNP, patients were 3.16 times more likely to have a poor mRS at discharge ( p = .021) and 5.40 times more likely at 3 months ( p < .0001).

CONCLUSION

There were significant relationships between BNP and cTnI and poor outcomes after aSAH. BNP may have utility as a marker of neurocardiac injury and outcomes after aSAH.

Risk for Cardiorespiratory Instability Following Transfer to a Monitored Step-Down Unit

BACKGROUND

Hospitalized patients who develop at least one instance of cardiorespiratory instability (CRI) have poorer outcomes. We sought to describe the admission characteristics, drivers, and time to onset of initial CRI events in monitored step-down unit (SDU) patients.

METHODS

Admission characteristics and continuous monitoring data (frequency 1/20 Hz) were recorded in 307 subjects. Vital sign deviations beyond local instability trigger threshold criteria, with a tolerance of 40 s and cumulative duration of 4 of 5 min, were classified as CRI events. The CRI driver was defined as the first vital sign to cross a threshold and meet persistence criteria. Time to onset of initial CRI was the number of days from SDU admission to initial CRI, and duration was length of the initial CRI epoch.

RESULTS

Subjects transferred to the SDU from units with higher monitoring capability were more likely to develop CRI (CRI n = 133 [44%] vs no CRI n = 174 [31%] P = .042). Time to onset varied according to the CRI driver. Subjects with at least one CRI event had a longer hospital stay (CRI 11.3 ± 10.2 d vs no CRI 7.8 ± 9.2 d, P < .001) and SDU stay (CRI 6.1 ± 4.9 d vs no CRI 3.5 ± 2.9 d, P < .001). First events were more often due to S_pO2 , whereas breathing frequency was the most common driver of all CRI.

CONCLUSIONS

Initial CRI most commonly occurred due to S_pO2 and was associated with prolonged SDU and hospital stay. Findings suggest the need for clinicians to more closely monitor SDU patients transferred from an ICU and parameters (S_pO2 , breathing frequency) that more commonly precede CRI events.

Impact of a Modified Early Warning Score on Rapid Response and Cardiopulmonary Arrest Calls in Telemetry and Medical-Surgical Units

To reduce the number of cardiac arrests in telemetry and medical- surgical units, a 70-bed community hospital integrated a weighted, aggregate, electronic modified early warning score into the elec- tronic medical record. Impact was evaluated via a quality improvement initiative.

Identifying Strategies for Effective Telemedicine Use in Intensive Care Units: The ConnECCT Study Protocol

Telemedicine, the use of audiovisual technology to provide health care from a remote location, is increasingly used in intensive care units (ICUs). However, studies evaluating the impact of ICU telemedicine show mixed results, with some studies demonstrating improved patient outcomes, while others show limited benefit or even harm. Little is known about the mechanisms that influence variation in ICU telemedicine effectiveness, leaving providers without guidance on how to best use this potentially transformative technology. The Contributors to Effective Critical Care Telemedicine (ConnECCT) study aims to fill this knowledge gap by identifying the clinical and organizational factors associated with variation in ICU telemedicine effectiveness, as well as exploring the clinical contexts and provider perceptions of ICU telemedicine use and its impact on patient outcomes, using a range of qualitative methods. In this report, we describe the study protocol, data collection methods, and planned future analyses of the ConnECCT study. Over the course of 1 year, the study team visited purposefully sampled health systems across the United States that have adopted telemedicine. Data collection methods included direct observations, interviews, focus groups, and artifact collection. Data were collected at the ICUs that provide in-person critical care as well as at the supporting telemedicine units. Iterative thematic content analysis will be used to identify and define key constructs related to telemedicine effectiveness and describe the relationship between them. Ultimately, the study results will provide a framework for more effective implementation of ICU telemedicine, leading to improved clinical outcomes for critically ill patients.

Erratum to: 36th International Symposium on Intensive Care and Emergency Medicine: Brussels, Belgium. 15-18 March 2016

[This corrects the article DOI: 10.1186/s13054-016-1208-6.].

Learning temporal rules to forecast instability in continuously monitored patients

Inductive machine learning, and in particular extraction of association rules from data, has been successfully used in multiple application domains, such as market basket analysis, disease prognosis, fraud detection, and protein sequencing. The appeal of rule extraction techniques stems from their ability to handle intricate problems yet produce models based on rules that can be comprehended by humans, and are therefore more transparent. Human comprehension is a factor that may improve adoption and use of data-driven decision support systems clinically via face validity. In this work, we explore whether we can reliably and informatively forecast cardiorespiratory instability (CRI) in step-down unit (SDU) patients utilizing data from continuous monitoring of physiologic vital sign (VS) measurements. We use a temporal association rule extraction technique in conjunction with a rule fusion protocol to learn how to forecast CRI in continuously monitored patients. We detail our approach and present and discuss encouraging empirical results obtained using continuous multivariate VS data from the bedside monitors of 297 SDU patients spanning 29 346 hours (3.35 patient-years) of observation. We present example rules that have been learned from data to illustrate potential benefits of comprehensibility of the extracted models, and we analyze the empirical utility of each VS as a potential leading indicator of an impending CRI event.

Predicting cardiorespiratory instability

This article is one of ten reviews selected from the Annual Update in Intensive Care and Emergency medicine 2016. Other selected articles can be found online at http://www.biomedcentral.com/collections/annualupdate2016. Further information about the Annual Update in Intensive Care and Emergency Medicine is available from http://www.springer.com/series/8901.

Monitoring cardiorespiratory instability: Current approaches and implications for nursing practice

Unrecognised in-hospital cardiorespiratory instability (CRI) risks adverse patient outcomes. Although step down unit (SDU) patients have continuous non-invasive physiologic monitoring of vital signs and a ratio of one nurse to four to six patients, detection of CRI is still suboptimal. Telemedicine provides additional surveillance but, due to high costs and unclear investment returns, is not routinely used in SDUs. Rapid response teams have been tested as possible approaches to support CRI patients outside the intensive care unit with mixed outcomes. Technology-enabled early warning scores, though rigorously studied, may not detect subtle instability. Efforts to utilise nursing intuition as a means to promote early identification of CRI have been explored, but the problem still persists. Monitoring systems hold promise, but nursing surveillance remains the key to reliable early detection and recognition. Research directed towards improving nursing surveillance and facilitating decision-making is needed to ensure safe patient outcomes and prevent CRI.

Modelling Risk of Cardio-Respiratory Instability as a Heterogeneous Process

Cardio-respiratory instability (CRI) occurs frequently in acutely ill. If not identified and treated early, it leads to significant morbidity and mortality. Current practice primarily relies on vigilance of the clinical personnel for early recognition of CRI. Given limited monitoring resources available in critical care environment, it can be suboptimal. Thus, an "Early Warning Scoring" mechanism is desirable to alert medical team when a patient is approaching instability. It is widely recognized that critically ill may show subtle changes prior to the onset of CRI, but it is not well known how their risk evolves before the onset. Using large amounts of physiological data routinely gathered from continuous noninvasive monitoring of Step-Down Unit patients, we demonstrate a data-driven approach that: (1) Characterizes patient's individual CRI risk process; (2) Identifies groups of patients that progress along similar risk evolution trajectories; (3) Utilizes grouping information to help forecast the emergence of CRI.

Thromboelastography: A Practice Summary for Nurse Practitioners Treating Hemorrhage

Nurse practitioners may manage patients with coagulopathic bleeding which can lead to life-threatening hemorrhage. Routine plasma-based tests such as prothrombin time and activated partial thromboplastin time are inadequate in diagnosing hemorrhagic coagulopathy. Indiscriminate administration of fresh frozen plasma, platelets or cryoprecipitate for coagulopathic states can be extremely dangerous. The qualitative analysis that thromboelastography provides can facilitate the administration of the right blood product, at the right time, thereby permitting the application of goal-directed therapy for coagulopathic intervention application and patient survival.

Patients in the radiology department may be at increased risk of developing critical instability

The purpose of this study was to calculate the event rate for in-patients in the Radiology Department (RD) developing instability leading to calls for Medical Emergency Team assistance (MET-RD) compared to general ward (MET-W) patients. A retrospective comparison was done of MET-RD and MET-W calls in 2009 in a U.S. tertiary hospital with a well-established MET system. MET-RD and MET-W event rates represented as MET calls/hour/1000 admissions, adjusted for length of stay (LOS); rates also calculated for RD modalities. There were 31,320 hospital ward admissions had 1,230 MET-W, and among 149,569 radiology admissions there were 56 MET-RD. When adjusted for LOS, the MET-RD event rate was 2 times higher than the MET-W rate (0.48 vs. 0.24 events/hour/1000 admissions). Event rates differed by procedure: computed tomography (CT) had 38% of MET-RDs (event rate 0.89); magnetic resonance imaging (MRI) accounted for 27% (event rate 1.56). Nuclear medicine had 1% of RD admissions but these patients accounted for 5% of MET-RD (event rate 1.53). Interventional radiology (IR) had 6% of RD admissions but 16% of MET-RD (event rate 0.61). While general x-ray comprised 63% of RD admissions, only 11% of MET-RD involved their care (event rate 0.09). In conclusion, the overall MET-RD event rate was twice the MET-W event rate; CT, MRI and IR rates were 3.7-6.5 times higher than on wards. RD patients are at increased risk for a MET call compared to ward patients when the time at risk is considered. Increased surveillance of RD patients is warranted.

The interface between monitoring and physiology at the bedside

Hemodynamic instability as a clinical state represents either a perfusion failure with clinical manifestations of circulatory shock or heart failure or 1 or more out-of-threshold hemodynamic monitoring values, which may not necessarily be pathologic. Different types of causes of circulatory shock require different types of treatment modalities, making these distinctions important. Diagnostic approaches or therapies based on data derived from hemodynamic monitoring assume that specific patterns of derangements reflect specific disease processes, which respond to appropriate interventions. Hemodynamic monitoring at the bedside improves patient outcomes when used to make treatment decisions at the right time for patients experiencing hemodynamic instability.

Discharge outcome in adults treated with extracorporeal membrane oxygenation

BACKGROUND

Extracorporeal membrane oxygenation (ECMO) is used for critically ill patients when conventional treatments for cardiac or respiratory failure are unsuccessful.

OBJECTIVES

To describe patient and treatment characteristics and discharge outcome for ECMO patients, determine which characteristics are associated with good (survival) versus poor (death before hospital discharge) outcomes, and compare characteristics of patients with cardiac versus respiratory failure indicating ECMO.

METHODS

Single-center, retrospective review of all adult patients treated with ECMO from 2005 through 2009.

RESULTS

A total of 212 patients received ECMO for cardiac (n = 126) or respiratory (n = 86) failure. Mean age was 51 (SD, 14.5) years; support duration was 135 (SD, 149) hours. Survival to discharge was 33% overall; 50% for respiratory indication and 21% for cardiac indication patients. Patients with poor outcomes were older (53 vs 47 years, P = .007), more likely to require cardiovascular support before ECMO (99% vs 91%; P = .02), and had more transfusions (48 vs 24 units, P = .005) and complications (99% vs 87%; P < .001) than did patients with good outcomes. For cardiac patients, older age was associated with poor outcome (poor, 55 vs good, 48 years; P = .01). For respiratory patients, poor outcome was associated with more ventilator days before ECMO (poor, 6 vs good, 3; P = .01), higher peak inspiratory pressure (poor, 39 vs good, 35 cm H2O; P = .02), and lower pulmonary compliance (poor, 19 vs good, 25 mL/cm H2O; P = .008).

CONCLUSIONS

Patients with respiratory indications for ECMO experienced better survival than did cardiac patients. Increasing age was associated with poor outcome. Complications, regardless of ECMO indication, were common and associated with poor outcome.

Development and implementation of an educational support process for electronic nursing admission assessment documentation

Educating nurses in use of the electronic health record nursing admission assessment using e-learning alone may not yield best results. Use of a hybrid instructional method of e-learning followed by a brief (20-minute) slide presentation with face-to-instruction significantly improved nursing documentation.

Cardiac abnormalities after aneurysmal subarachnoid hemorrhage: effects of β-blockers and angiotensin-converting enzyme inhibitors

BACKGROUND

Cardiac abnormalities attributed to adrenergic surge are common after aneurysmal subarachnoid hemorrhage. Prescribed medications that block adrenergic stimulation may suppress the onset of cardiopulmonary compromise in patients after aneurysmal subarachnoid hemorrhage.

OBJECTIVES

To compare the incidence of early cardiac complications between patients who reported prescribed use of β-blockers and/or angiotensin-converting enzyme inhibitors before aneurysmal subarachnoid hemorrhage and patients who did not.

METHODS

A retrospective review of 254 adult patients after acute aneurysmal subarachnoid hemorrhage who were enrolled in an existing R01 study. Demographic data and history were obtained from patients'/proxies' reports and charts. Cardiac enzyme levels, 12-lead electrocardiograms, and chest radiographs were obtained on admission. Holter monitoring and echocardiograms were completed as a part of the R01 study.

RESULTS

Patients reporting prescribed use of angiotensin-converting enzyme inhibitors or β-blockers before aneurysmal subarachnoid hemorrhage had more ventricular and supraventricular ectopy on a Holter report than did patients who did not (P < .05). When age, race, sex, and injury (Fisher grade) were controlled for, patients reporting use of β-blockers were 8 times more likely than others to have occasional to frequent ventricular ectopy (P = .02).

CONCLUSION

No concrete evidence was found that exposure to adrenergic blockade before aneurysmal subarachnoid hemorrhage provides protection from neurocardiac injury.

Characteristics of patients with cardiorespiratory instability in a step-down unit

BACKGROUND

Patients in step-down units are at higher risk for developing cardiorespiratory instability than are patients in general care areas. A triage tool is needed to identify at-risk patients who therefore require increased surveillance.

OBJECTIVES

To determine demographic (age, race, sex) and clinical (Charlson Comorbidity Index at admission, admitting diagnosis, care area of origin, admission service) differences between patients in step-down units who did and did not experience cardiorespiratory instability.

METHODS

In a prospective longitudinal pilot study, 326 surgical-trauma patients had continuous monitoring of heart rate, respirations, and oxygen saturation and intermittent noninvasive measurement of blood pressure. Cardiorespiratory instability was defined as heart rate less than 40/min or greater than 140/min, respirations less than 8/min or greater than 36/min, oxygen saturation less than 85%, or blood pressure less than 80 or greater than 200 mm Hg systolic or greater than 110 mm Hg diastolic. Patients' status was classified as unstable if their values crossed these thresholds even once during their stay.

RESULTS

Cardiorespiratory instability occurred in 34% of patients. The Charlson Comorbidity Index was the only variable associated with instability conditions. Compared with patients with no comorbid conditions (50%), more patients with at least 1 comorbid condition (66%) experienced instability (P = .006). Each 1-unit increase in the Charlson Index increased the odds for cardiorespiratory instability by 1.17 (P = .03).

CONCLUSION

Although the relationship between Charlson Comorbidity Index and cardiorespiratory instability was weak, adding it to current surveillance systems might improve detection of instability.

Medical emergency team calls in the radiology department: patient characteristics and outcomes

OBJECTIVE

We sought to identify the characteristics of patients who experience medical emergency team calls in the radiology department (MET-RD) and the relationship between these characteristics and patient outcomes.

DESIGN/PARTICIPANTS

Retrospective review of 111 inpatient MET-RD calls (May 2008-April 2010).

SETTING

Academic medical centre with a well established MET system.

MEASUREMENTS

The characteristics of patients before, during and after transport to radiology were extracted from medical records and administrative databases. These characteristics were compared between patients with good and poor outcomes.

MAIN RESULTS

The majority of patients who experience MET-RD calls had a Charlson Comorbidity Index ≥4 and were from non-intensive care units (60%). Almost half (43%) of MET-RD calls occurred during patients' first day in hospital. Patients commonly arrived with nasal cannula oxygen (38%), recent tachypnoea (28%) and tachycardia (34%). A minority (16%) fulfilled MET call criteria in the 12 h before the MET-RD. MET-RD etiologies were cardiac (41%), respiratory (29%) or neurological (25%), and occurred most frequently during CT (44%) and MRI (22%) testing. Post MET-RD, the majority of patients (70%) required a higher level of care. Death before discharge (25%) was associated with need for cardiovascular support prior to RD transport (p=0.02), need for RD monitoring (p=0.02) and need for heightened RD surveillance (p=0.04).

CONCLUSIONS

The majority of patients who experienced MET-RD calls came from non-intensive care units, with comorbidities and vital sign alterations prior to arrival at the RD. Risk appeared to be increased for those requiring CT and MRI. These findings suggest that prior identification of a subset of patients at risk of instability in the RD may be possible.

Psychosocial correlates of weight maintenance among black & white adults

OBJECTIVES

To investigate (1) weight maintenance among black and white participants and (2) psychosocial correlates (eg, healthy eating barriers, self-efficacy, stress) of weight maintenance 18 months after behavioral weight-loss treatment.

METHODS

Linear and logistic regression examined weight change and unsuccessful weight maintenance (>5% weight gain) among 107 black and white adults.

RESULTS

After controlling for socio-demographics, differences in weight maintenance between ethnicities were not generally noted. Healthy eating barriers and stressful life events were associated with weight gain, P<.04.

CONCLUSIONS

Strategies to cope with stressful events and overcome barriers to eating healthfully are needed for weight maintenance among both ethnicities.

Patients' instability, emergency response, and outcomes in the radiology department

BACKGROUND

Inpatients may be at risk of cardiopulmonary instability during radiologic testing. Calling the medical emergency team is one rescue intervention that brings a team of critical care providers to the unstable patient. Little is known, however, about patients' instability and activations of the medical emergency team in the radiology department (RD-MET).

OBJECTIVES

To describe the cause of activation of the RD-MET for hospitalized patients, temporal attributes of RD-MET involvement, characteristics of RD-MET patients, and characteristics associated with good and poor outcomes after RD-MET activation.

METHODS

Retrospective pilot study of RD-MET calls for 64 inpatients in a tertiary care hospital during 2009.

RESULTS

Reasons for RD-MET activation were 39% neurological, 38% cardiac, and 22% respiratory, and nearly half (42%) occurred during a computed tomography scan. Most RD-MET calls were made between 10 am and noon. RD-MET patients had a mean age of 61 (SD, 19) years; 52% were female, and 89% were white. Admitting diagnoses were most commonly neurological (20%), cardiovascular (16%), and abdominal (16%). The most common comorbid conditions were chronic obstructive pulmonary disease (23%) and diabetes (20%). Half of RD-MET inpatients were from a general care unit, and 56% required preexisting oxygen support. After RD-MET involvement, 61% of patients required a higher level of care; 3% died during the MET intervention, and 19% died later in hospitalization. Patients with preexisting comorbid conditions were more likely to have poor outcomes after the RD-MET intervention (P = .001).

CONCLUSIONS

RD-MET patients with comorbid conditions, from a general care unit, and at risk for neurological deterioration arrive in the radiology department with potentially underestimated support needs. Greater support in specific time frames and locations may be warranted to improve outcomes.

Intrahospital Transport to the Radiology Department: Risk for Adverse Events, Nursing Surveillance, Utilization of a MET and Practice Implications

Nurses providing care in the Radiology Department (RD) are challenged by the broad scope of conditions and varied acuity of patients served by this unit. Nurses must facilitate the required diagnostic testing and simultaneously provide the surveillance necessary to detect physiologic changes signaling the need for rescue interventions. When instability occurs, one method of rescue involves activation of a Medical Emergency Team (MET) to bring an experienced cadre of critical care providers to the unstable patient. Despite recognition that the RD can be a high risk area, there is little in the literature specific to the surveillance of RD patients, risk for and prevention of adverse events, MET activation or the management of patient instability specific to the RD. The purpose of this paper is to examine what is known regarding risk for adverse events during intrahospital transport, utilization of a MET as a rescue intervention, and practice implications.

Predictors of delayed cerebral ischemia after aneurysmal subarachnoid hemorrhage: a cardiac focus

BACKGROUND

Myocardial injury after aneurysmal subarachnoid hemorrhage (aSAH) is associated with poor outcomes. Delayed cerebral ischemia (DCI) is also a complication of aSAH. We sought to determine whether (1) DCI could be predicted by demographics, aSAH severity/aneurysm location, or aSAH-associated myocardial injury (SAHMI), and (2) DCI is associated with increased mortality after aSAH.

METHODS

Prospective longitudinal study of 149 aSAH subjects with definitive DCI evaluation, age 18-75 years, Hunt and Hess (HH) ≥ 3, and/or Fisher ≥ 2, and admitted to the Neurovascular ICU. DCI was defined by the presence of neurological deterioration accompanied by evidence of abnormal cerebral blood flow.

RESULTS

Subjects were 48% DCI(+) and 52% DCI(-). DCI(+) subjects had more severe aSAH [HH (P = 0.002), Fisher (P = 0.004), admission Glasgow Coma Scale (P = 0.018)]. More DCI(+) subjects had pulmonary congestion than DCI(-) subjects (63 vs. 39%, P = 0.003). On echocardiogram, cardiac output (CO, liters per minute [LPM]) was significantly higher in DCI(+) than in DCI(-) subjects (6 ± 2 vs. 5 ± 1 LPM; P = 0.015). Multivariate analysis identified CO and Fisher grade as independent predictors of DCI (P = 0.02, 0.019). For each 1 LPM increase in CO, the odds of DCI increased by 1.5 (95% CI: 1.1-2.1). Fisher grade 4 increased the odds of DCI by 6.5 compared to Fisher grade 2 (95% CI: 1.6-25.8). After controlling for Fisher grade, CO remained an independent predictor of DCI (P = 0.02). Three-month mortality rate was not significantly different between DCI groups, P = 0.786.

CONCLUSION

Elevated CO and Fisher grade are predictors of DCI after aSAH. However, prevention of DCI may not decrease mortality.

Elevated cardiac troponin I and functional recovery and disability in patients after aneurysmal subarachnoid hemorrhage

BACKGROUND

Patients with aneurysmal subarachnoid hemorrhage experience myocardial injury at the time of rupture, but its effect on functional recovery and disability is unclear.

OBJECTIVE

To describe the prevalence of myocardial injury, as indicated by high serum levels of cardiac troponin I (≥0.3 ng/mL), within the first 5 days after aneurysmal subarachnoid hemorrhage and the effect of the injury on 3-month functional recovery and disability.

METHODS

In a prospective longitudinal study, 239 patients with Hunt/Hess grade 3 or greater and/or Fisher grade 2 or greater at admission had serum level of troponin I measured on days 0 to 5. Patients were interviewed at 3 months to evaluate functional recovery (Glasgow Outcome Scale) and functional disability (Modified Rankin Scale). Statistics included χ² analysis, t tests, and binary logistic regression.

RESULTS

Troponin values were elevated in 33.5% of the patients, and few patients in either group had a history of coronary artery disease (7.4% with troponin levels ≥0.3 ng/mL vs 2.7% with levels <0.3 ng/mL, P = .12). Higher troponin levels were significantly related to age and Hunt/Hess and Fisher grades, but not race, and were significantly associated with poorer functional recovery (P < .001) and more functional disability (P < .001). Even after controls for age, race, and more severe Hunt/Hess grades, higher levels remained a significant predictor of poorer functional recovery (P = .04) and disability (P = .01). CONCLUSION Elevated levels of cardiac troponin I after aneurysmal subarachnoid hemorrhage are common in patients with no cardiac history, are associated with severity of the hemorrhage, and are independently predictive of poorer functional recovery and increased disability.

Neuroglobin genetic polymorphisms and their relationship to functional outcomes after traumatic brain injury

Neuroglobin has shown rich neuroprotective effects against cerebral hypoxia, and therefore has the potential to impact outcomes after traumatic brain injury (TBI). However, to date an association between genetic variation within the human neuroglobin (NGB) gene and recovery post-TBI has not been reported. The purpose of this study was to explore the relationship between NGB genotypes and outcomes (as assessed by the Glasgow Outcome Scale [GOS], the Disability Rating Scale [DRS], and the Neurobehavioral Rating Scale-Revised [NRS-R]) after severe TBI. Genotyping using TaqMan allele discrimination for two tagging single nucleotide polymorphisms (tSNPs) that represent the two haplotype blocks for NGB (rs3783988 and rs10133981) was completed on DNA obtained from 196 Caucasian patients recovering from severe TBI. Patients were dichotomized based on the presence of the variant allele for each tSNP. Chi-square and Fisher's exact tests were used to compare characteristics between groups. Multivariate linear regression was used to examine NGB tSNPs and recovery from severe TBI. Subjects with the TT genotype (wild-type) for rs3783988 were more likely to have better GOS and DRS scores at 3, 6, 12, and 24 months, while rs10133981 genotype was not significantly related to functional outcome. After controlling for age, gender, and Glasgow Coma Scale (GCS) score, those subjects with the rs3783988 TT genotype had more than a 2.65-times greater likelihood of better functional outcomes than individuals with genotypes harboring a variant allele. Data suggest that the haplotype block represented by rs3783988 in NGB appears to influence recovery after severe TBI. Represented within this haplotype block of NGB is the region that codes for the oxygen-binding portion of NGB.